Measuring the Burden—Alcohol’s Evolving Impact on Individuals, Families, and Society

117 Editors’ Note 120 In This Issue 122 Measuring the Burden: Alcohol’s Evolving Impact Ralph hingson, Sc.D., and Jürgen Rehm, Ph.D. 128 Using Surveys to Calculate Disability-Adjusted Life- Years Wolfgang Wiedermann, Ph.D., and Ulrich Frick, Ph.D. 135 Community Indicators: Assessing the Impact of Alcohol Use on Communities Andrea Flynn, Ph.D., and Samantha Wells, Ph.D. 150 Focus on: The Burden of Alcohol Use—Trauma and Emergency Outcomes Cheryl J. Cherpitel, Dr.P.H. 155 Chronic Diseases and Conditions Related to Alcohol Use Kevin D. Shield, M.H.Sc.; Charles Parry, Ph.D.; and Jürgen Rehm, Ph.D. 174 Alcohol and Mortality: Global Alcohol-Attributable Deaths From Cancer, Liver Cirrhosis, and Injury in 2010 Jürgen Rehm, Ph.D., and Kevin D. Shield, M.H.Sc. 184 APIS: The NIAAA Alcohol Policy Information System Michael Hilton, Ph.D. 186 The Burden of Alcohol Use: Focus on Children and Preadolescents John E. Donovan, Ph.D. 193 Prevalence and Predictors of Adolescent Alcohol Use and Binge Drinking in the United States Megan E. Patrick, Ph.D., and John E. Schulenberg, Ph.D. 201 The Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students Aaron White, Ph.D., and Ralph Hingson, Sc.D. 219 Focus On: Women and the Costs of Alcohol Use Sharon C. Wilsnack, Ph.D.; Richard W. Wilsnack, Ph.D.; and Lori Wolfgang Kantor, M.A. 229 Focus On: Ethnicity and the Social and Health Harms From Drinking Karen G. Chartier, Ph.D.; Patrice A.C. Vaeth, Dr.P.H.; and Raul Caetano, M.D., Ph.D. 238 Gaps in Clinical Prevention and Treatment for Alcohol Use Disorders: Costs, Consequences, and Strategies Mark L. Willenbring, M.D. 244 The World Health Organization’s Global Monitoring System on Alcohol and Health Vladimir Poznyak, M.D., Ph.D.; Alexandra Fleischmann, Ph.D.; Dag Rekve, MS.c.; Margaret Rylett, M.A.; Jürgen Rehm, Ph.D., and Gerhard Gmel, Ph.D. 250 Measuring the Burden—Current and Future Research Trends: Results From the NIAAA Expert Panel on Alcohol and Chronic Disease Epidemiology Rosalind A. Breslow, Ph.D., M.P.H., R.D., and Kenneth J. Mukamal, M.D. EDitoRS’ NotE

Measuring the Burden Alcohol’s Evolving Impact on Individuals, Families, and Society

Jürgen Rehm, Ph.D., and Ralph Hingson, Sc.D.

lcohol use is associated with tremendous costs to the drinker, Jürgen Rehm, Ph.D., is director those around him or her, and society as a whole. These costs of the Social and Epidemiological A result from the increased health risks (both physical and mental) Research Department at the associated with alcohol consumption as well as from the social harms Centre for Addiction and Mental caused by alcohol. This issue of Alcohol Research: Current Reviews examines Health, chair and professor in the public health impact of alcohol consumption, looking at the full burden of disease that can be attributed to drinking. the Dalla Lana School of Public The attempt to measure the impact of alcohol use on various disease Health, University of Toronto, categories is relatively new to the alcohol research field. In fact, much of Canada, and section head at the our understanding of how alcohol affects health and disease in society is Institute for Clinical Psychology rooted in work from the 1980s and 1990s. This research reflects a truly and Psychotherapy, Technische international perspective. A group of Australian authors, led by Dr. Universität Dresden, Dresden, Dallas English, were some of the first to look at the issues involved in Germany. attributing mortality and morbidity to substance abuse (English et al. 1995). Dr. James Shultz and his colleagues conducted other seminal Ralph Hingson, Sc.D., is director research for the Centers for Disease Control and Prevention. They of the Division of Epidemiology developed the Alcohol-Related Disease Impact (ARDI) software to and Prevention Research at the allow States to calculate mortality, years of potential life lost (YPLL), National Institute on Alcohol direct health care costs, indirect morbidity and mortality costs, and Abuse and Alcoholism, National non–health sector costs associated with alcohol use (Shultz et al. 1991). Institutes of Health, Rockville, Canadian researchers, under the direction of Dr. Eric Single, developed the Canadian version of the alcohol-attributable fractions in the mid- Maryland. 1990s (Single 1999). Also at the forefront of research in this field are Dr. Robin Room and this issue’s Scientific Review Editor, Dr. Jürgen Rehm, both of whom have made significant contributions to our over- all understanding of the field. The study of the burden of disease on a truly global scale began with the World Health Organization’s (WHO’s) Global Burden of Disease Study (Murray and Lopez 1996). Thanks to the WHO efforts, we now have a worldwide view of the far-reaching consequences of alcohol use and misuse. Although the field has made much progress and in a short time, there are research gaps that still remain. For example, more research is needed on the relationship between diseases and detailed drinking patterns; more needs to be known about the burden of alcohol-related mental disorders (e.g., depression); future research needs to disentangle the effect of comorbid conditions when assessing the burden of disease attributable to alcohol; the estimates of how alcohol contributes to infectious diseases like HIV need to be further refined; and the alcohol-attributable fractions need to be updated more frequently in response to new developments in science and as the population’s health status and behaviors change.

Editors’ Note 117 EDitoRS’ NotE ContInuEd

Finally, the impact of alcohol on social harm, including harm to people other than the drinker, still is terra incognita in many areas (Gmel and Rehm 2003). Over the last two decades, the United States has made substantial progress in improving public health. Still, alcohol remains an important risk factor for disease burden and social harm, not only in the United States but also globally (Murray et al. 2013). Additional research in this area will increase our understanding of alcohol’s role in creating disease burden and social harm and aid in the development of stronger, more effective measures to prevent these devastating effects. ■

References

English, D.R.; holman, C.D.J.; milnE, E.; Et al. The Quantification of Drug Caused Morbidity and Mortality in Australia 1995. Canberra, australia: Commonwealth Department of human services and health, 1995.

gmEl, g., anD REhm, J. harmful alcohol use. Alcohol Research & Health 27(1):52–62, 2003. PmiD: 15301400

muRRay, C., anD loPEz, a. The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability From Diseases, Injuries and Risk Factors in 1990 and Projected to 2020. Cambridge, ma: harvard university and World health organization, 1996.

muRRay, C.J.; abRaham, J.; ali, m.K.; Et al. the state of us health, 1990-2010: burden of diseases, injuries, and risk factors. JAMA: Journal of the American Medical Association 310(6):591–608, 2013. PmiD: 23842577

singlE, E.; Robson, l.; REhm, J.; anD XiE, X. morbidity and mortality attributable to alcohol, tobacco, and illicit drug use in Canada. American Journal of Public Health 89(3):385–390, 1999. PmiD: 10076491

shultz, J.m.; RiCE, D.P.; PaRKER, D.l.; Et al. Quantifying the disease impact of alcohol with aRDi software. Public Health Report 106(4):443–450, 1991. PmiD: 1652146

118 Alcohol Research: Current Reviews iN tHiS iSSUE

MEASURiNG tHE BURDEN: their strengths and limitations, and help further refine our knowledge of ALCoHoL’S EvoLviNG iMPACt discusses indicators used in reference which chronic diseases and conditions lcohol’s impact stretches far beyond to four main topics relating to alcohol are causally linked to alcohol consump- A the drinker, affecting his or her use and problems at the community tion. (pp. 155–173) family as well as the community at level: alcohol use, patterns, and problems; large. This widespread impact makes alcohol availability; alcohol-related FoCUS oN: ALCoHoL AND alcohol a major public health concern. health outcomes/trauma; and alcohol- MoRtALitY This introductory article by Drs. Ralph related crime and enforcement. (pp. lcohol consumption has long been 135–149) Hingson and Jürgen Rehm provides Arecognized as a risk factor for mor- an overview of the history of alcohol’s FoCUS oN: tRAUMA AND tality. By combining data on alcohol burden of disease. (pp. 122–127) EMERGENCY oUtCoMES per capita consumption, alcohol- drink- ing status and alcohol-drinking patterns, USiNG SURvEYS to CALCULAtE ospital emergency departments risk relationships, and mortality, the DiSABiLitY-ADJUStED LiFE-YEARS H(EDs) often see patients arriving Comparative Risk Assessment Study with alcohol-attributable injuries. estimated alcohol-attributable mortality o quantify the total burden of mor- For this reason, researchers frequently tbidity attributable to alcohol use, for 1990 and 2010. In this article, use EDs to examine the relationship Drs. Jürgen Rehm and Kevin D. Shield so-called disability weights (DWs) between alcohol consumption and must be generated. General-population discuss alcohol’s role in the global bur- injury risk using different study designs. den of mortality. (pp. 174–183) surveys can be used to derive DWs As reported by Dr. Cheryl J. Cherpitel, from health valuation tasks. This article these analyses found elevated injury by Drs. Wolfgang Weidermann and FoCUS oN CHiLDREN AND risk after alcohol consumption, with PREADoLESCENtS Ulrich Frick discusses the application of the extent of risk increase depending three psychometric methods—pairwise on factors such as drinking patterns, ecause there are few surveillance comparisons, ranking tasks, and visual concurrent use of other drugs, or even Bstudies of alcohol use and alcohol- related problems among children and analog scales—in general-population study design. Additionally, Dr. Cherpitel surveys and outlines their strengths preadolescents, estimating the alcohol explores numerous other aspects of burden in this population is especially and weaknesses. In addition, the article the relationship between alcohol use discusses a recently proposed health difficult. This article by Dr. John E. and injury risk assessed by ED studies. Donovan summarizes information from valuation framework, which highlights (pp. 150–154) the underlying cognitive processes U.S. national and Statewide surveys on the prevalence of alcohol use among from a social judgment perspective. FoCUS oN: CHRoNiC DiSEASES Furthermore, it presents a structured children in grades 6 and lower. Although AND CoNDitioNS RELAtED to the rates of alcohol use are relatively data-collection procedure that seems ALCoHoL USE low in this population, substantial promising in deriving DWs from general- umerous chronic diseases and con- numbers of children do in fact have population surveys. (pp. 128–134) Nditions are entirely or partially experience with alcohol. Limited attributable to heavy alcohol use; for available data highlight the need for ASSESSiNG tHE iMPACt oF better ongoing surveillance of this ALCoHoL USE oN CoMMUNitiES other conditions, however, alcohol can have a beneficial effect. Various factors, population. Although alcohol burden ommunity indicators are used to such as the average amount and pattern in children appears relatively low, it is Cassess the impact of alcohol on of alcohol consumption, have been increased through the alcohol use and communities. For communities, indi- found to influence alcohol’s impact on abuse of their parents, and through cator data can be used to inform priority- the mortality and morbidity related to the increased likelihood among early setting agendas by identifying specific chronic diseases and conditions. However, drinkers of alcohol problems and other concerns within a community, guide as Mr. Kevin D. Shield and Drs. Charles negative outcomes in adolescence and policy and education initiatives, moni- Parry and Jürgen Rehm report, although young adulthood. (pp. 186–192) tor community status on a particular alcohol consumption indisputably measure over time or in comparison contributes to the burden of chronic PREvALENCE AND PREDiCtoRS with other communities, and evaluate diseases and conditions, the methods oF ADoLESCENt ALCoHoL USE programs or policies. This article by currently used to calculate the relative AND BiNGE DRiNkiNG iN tHE Drs. Andrea Flynn and Samantha risks and alcohol-attributable fractions UNitED StAtES Wells reviews the main data sources have several limitations. Moreover, he Monitoring the Future study for community indicators, discusses new studies and confounders may tis an annual survey among 8th-,

120 Alcohol Research: Current Reviews ALCoHoL RESEARCH: Current Reviews

10th-, and 12th-grade students assess- moderate drinking, such levels of ment strategies on public health is ing their alcohol use, thereby allowing drinking also are associated with unclear. For example, existing screening researchers to better understand ado- increased risks of breast cancer and and brief interventions for nondepen- lescents’ consumption patterns during liver problems, and heavy drinking dent drinkers may potentially have a this vulnerable developmental period. increases risks of hypertension and large public health impact, whereas These surveys have found high preva- bone fractures and injuries. In addition, more effective ways are needed to reduce lence of drinking and binge drinking women's heavy-drinking patterns and the public health burden associated by the time students leave high alcohol use disorders are associated with functional dependence and alcohol school, report Drs. Megan E. Patrick with increased likelihood of many use disorders. (pp. 238–243) and John E. Schulenberg. The psychiatric problems, including authors also discuss the factors that depression, posttraumatic stress disorder, tHE WoRLD HEALtH influence adolescent alcohol use, such eating disorders, and suicidality, as oRGANizAtioN’S GLoBAL as parent and peer relationships, school well as increased risks of intimate MoNitoRiNG SYStEM oN and work, behavioral and drug-use partner violence and sexual assault. In ALCoHoL AND HEALtH problems, or personality characteristics, this article, Drs. Sharon C. Wilsnack as well as review potential long-term and Richard W. Wilsnack and Ms. ith growing awareness of the effects of adolescent alcohol con- Lori Wolfgang Kantor discuss drinking Wimpact of alcohol consumption sumption. (pp. 193–200) patterns among women during midlife on global health, the demand for and the risks of heavy drinking in. global information on alcohol con- ExCESSivE ALCoHoL this population. (pp. 219–228) sumption and alcohol-attributable CoNSUMPtioN AND RELAtED and alcohol-related harm as well as CoNSEqUENCES AMoNG FoCUS oN: EtHNiCitY AND tHE related policy responses has increased CoLLEGE StUDENtS SoCiAL AND HEALtH HARMS significantly. This article by Drs. esearch shows that multiple factors FRoM DRiNkiNG Vladimir Poznyak and Alexandra Rinfluence college drinking, from ative Americans, Hispanics, and Fleischmann, Mr. Dag Rekve, Ms. an individual's genetic susceptibility NBlacks are at higher risk of experi- Margaret Rylett, and Drs. Jürgen to the positive and negative effects of encing alcohol-attributable harms Rehm and Gerhard Gmel examines alcohol, alcohol use during high school, than Whites or Asians. This article by the increasing demand from World campus norms related to drinking, Drs. Karen G. Chartier, Patrice A.C. Health Organization (WHO) expectations regarding the benefits Vaeth, and Raul Caetano examines Member States and the latest policies and detrimental effects of drinking, this health disparity with regard to set forth by the WHO in response to penalties for underage drinking, alcohol and unintentional injuries, alcohol’s growing burden of disease parental attitudes about drinking intentional injuries, fetal alcohol syn- worldwide. (pp. 244–249) while at college, whether one is member drome, gastrointestinal diseases, car- of a Greek organization or involved in diovascular diseases, cancers, diabetes, RESULtS FRoM tHE NiAAA athletics, and conditions within the and infectious diseases. (pp. 229–237) ExPERt PANEL oN ALCoHoL larger community that determine how AND CHRoNiC DiSEASE accessible and affordable alcohol is. GAPS iN CLiNiCAL PREvENtioN EPiDEMioLoGY This article by Drs. Aaron White and AND tREAtMENt FoR ALCoHoL Ralph Hingson examines recent findings USE DiSoRDERS IAAA Expert Panel on Alcohol and Chronic Disease Epidemiology about the causes and consequences umerous prevention and treatment N of excessive drinking among college Napproaches have been developed workshop provided an excellent forum students relative to their noncollege to ameliorate the morbidity, premature for summarizing the current state of the peers and many of the strategies used mortality, and other social and eco- field of alcohol research and for identify- to collect and analyze relevant data, nomic burdens on society caused by ing future research opportunities. This as well as the inherent hurdles and heavy drinking. Those measures can article by Drs. Rosalind A. Breslow and limitations of such strategies. (pp. be targeted at different groups of Kenneth J. Mukamal reviews the out- 201–218) drinkers, including nondependent comes of the workshop and the ideas heavy drinkers, those with functional that highlight areas in need of additional FoCUS oN: WoMEN AND tHE dependence, and those with alcohol study and offer a roadmap for moving CoStS oF ALCoHoL USE use disorders. However, as Dr. Mark forward across a variety of methodologi- lthough there are proven beneficial L. Willenbring explains, the impact of cal approaches and content areas. (pp. Aeffects associated with light-to- current selective prevention and treat- 250–259)

in this issue 121 ALCoHoL RESEARCH: Current Reviews

MMeeaassuurriinngg tthhee BBuurrddeenn:: AAllccoohhooll’’ss EEvvoollvviinngg iimmppaacctt

Ralph Hingson, Sc.D., and Jürgen Rehm, Ph.D.

measuring the impact of alcohol consumption on morbidity and mortality depends on Jürgen Rehm, Ph.D., is director the accurate measurement of alcohol exposure, risk relationships, and outcomes. a of the Social and Epidemiological variety of complicating factors make it difficult to measure these elements. this article Research Department at the reviews these factors and provides an overview of the articles that make up this special Centre for Addiction and Mental issue on current research examining alcohol’s role in the burden of disease. these Health, chair and professor in topics include estimating alcohol consumption as well as alcohol-related morbidity the Dalla Lana School of Public and mortality in various demographic groups, and the burden of alcohol use disorders. kEY WoRDS: Alcohol use consumption; alcohol use frequency; alcohol use pattern; Health, University of Toronto, alcohol burden; public health impact; burden of disease; morbidity; mortality; Canada, and section head at the disease; measurement; outcomes; specificity of measurement; sensitivity of Institute for Clinical Psychology measurement and Psychotherapy, Technische Universität Dresden, Dresden, Germany.

Ralph Hingson, Sc.D., is director of the Division of Epidemiology and Prevention Research at the National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Rockville, Maryland.

his issue of Alcohol Research: Current • The measurement of outcomes. Acute Mortality and Morbidity Reviews examines the public health (Injuries and Poisonings) timpact of alcohol consumption These different measurements are Postmortem alcohol test data are not central to this issue of Alcohol Research: beyond the role of alcohol use disorders consistently available for many types of Current Reviews. alone (Room et al. 2005)—that is, it looks acute injury or poisoning deaths. The at the burden of disease. Determining best available U.S. estimates indicate impact hinges on accurate and consis- Measurement Challenges alcohol-attributable acute deaths out- tent “measurements.” As demonstrated number chronic disease deaths 44,000 in the articles in this issue, impact typi- Numerous challenges exist when mea- to 35,000 (Centers for Disease Control cally is estimated based on three elements suring the extent and predictors of and Prevention [CDC] 2013). Traffic (Rehm et al. 2010b; Walter 1976): alcohol-related mortality and morbidity. crashes have been the leading category Those challenges also affect our ability of alcohol-attributable injury or death • The measurement of exposure (i.e., to evaluate the effectiveness of inter- over the past 30 years. During that the relevant dimension of alcohol ventions to reduce alcohol-related time period, the majority of drivers in use) causing the burden (Kehoe morbidity and mortality. Different fatal crashes (both fatally injured and, et al. 2012; Rehm et al. 2010b); challenges exist for acute alcohol-related to a lesser extent, surviving drivers) mortality and chronic disease mortality have been tested for alcohol. This permits • The measurement of the risk rela- and morbidity, although some of the researchers to make accurate estimates tions (i.e., what level/pattern of same challenges confront measurement of the number of drivers, passengers, and consumption is linked to what in both areas. The following list of others who die in fatal crashes in which outcome) (Rehm et al. 2010a); and challenges is illustrative but not exhaustive. a driver was known to have been drinking.

122 Alcohol Research: Current Reviews In addition, by examining the Using this approach, researchers over time. In addition, researchers are characteristics of crashes and drivers in have been able to determine annual able to use quasi-experimental and fatal crashes where alcohol is present, State and national estimates of alcohol other research designs to evaluate the National Highway Traffic Safety involvement in fatal traffic crashes whether State-level traffic safety legisla- Administration (NHTSA) (Klein since 1982. Furthermore, having those tion and community-level education 1986; NHTSA 2002) has developed accurate direct-test results and imputed and law enforcement and treatment an imputational approach to estimate results has permitted researchers to programs are effective in reducing the proportion of fatal crashes involv- make epidemiologic estimates of the alcohol-related traffic deaths (Ferguson ing alcohol even when the driver is not increased odds of fatal crashes and other 2012; Hingson and White 2013). tested. This approach was verified using crash involvement at various BALs Such studies have guided policy- data from States with high percentages (Voas et al. 2012). Drivers who were makers to select and implement effective of drivers involved in fatal crashes that stopped at random in roadside surveys programs and policies. Since the early were tested for alcohol use. Researchers were compared with drivers who were 1980s, alcohol-related traffic death used the NHTSA model to predict fatally injured in single-vehicle crashes rates per 100,000 have been reduced the percentage of fatal crashes that and who were driving in the same States more than 50 percent versus the decline involved alcohol and then compared on the same types of roads on the same in traffic crash death rates where alco- those findings with the actual alcohol days of the week and times of day. hol is not involved (figure 1). It has test results. NHTSA found they could Because these data are available been estimated that as many as 300,000 estimate with great accuracy not only monthly on a national, State, and deaths have been prevented as a result the proportion of fatal crashes involv- community level, researchers also have of reduced incidences of drinking and ing alcohol-positive drivers but also the been able to monitor trends in fatal driving, which is greater than those blood alcohol level (BAL) of the driver crashes involving alcohol relative to attributed to increased use of airbags, at the time of the crash. fatal crashes where alcohol is not involved seat belts, and motorcycle and bicycle

Figure 1 alcohol-related versus non–alcohol-related traffic fatalities, rate per 100,000, all ages, united states, 1982–2010

souRCEs: national highway traffic safety administration, 2012; u.s. Census bureau, 2012.

Measuring the Burden: Alcohol’s Evolving impact 123 helmets combined (Cummings et al. overdose deaths (White et al. 2011). driving. Half of the deaths in crashes 2006; Fell and Voas 2006). People may be involved in traffic crashes involving drinking drivers under the Unfortunately, unlike traffic deaths, or poisoning deaths at lower BALs if age of 25 are those other than the driver. postmortem alcohol testing is not other drugs are present, and this may This has incited citizen activists and nearly as complete for other types of modify BAL levels used in establishing policymakers to pass more than 2,000 unintentional or intentional poisoning attributable fractions for motor-vehicle laws at the State and Federal levels to and injury deaths. In 18 States (figure and poisoning deaths. Most national reduce alcohol-impaired driving 2), a violent-death registry is in place surveys and many research projects (Hingson et al. 2003). where 80 percent or more of all homi- inquire about alcohol and drug consump- Fourth, many prevention activities cides and suicides are tested. However, tion separately not simultaneously. If are implemented at the community testing levels of alcohol for other types alcohol and drugs pharmacologically level, and community-level data are of injuries or deaths is not routine. As interact, simultaneous-use questions needed to stimulate the planning and a consequence, imputations for alcohol should be considered. evaluation of those interventions involvement in other types of injuries Third, it is important to calculate (Hingson and White 2012). Yet most or deaths have not been developed, the secondhand harm alcohol misuse surveillance data-monitoring systems and studies of laws and programs to poses. Just as awareness of the second- measure behavior and consequences at reduce those types of injuries and deaths hand negative consequences of passive the State and Federal levels. Strategies do not have the same precision as smoke inhalation has heightened the are needed to either facilitate more evaluations of efforts to prevent alcohol- public health resolve to curb smoking, community-level data collection or related traffic deaths. learning about the secondhand effects to offer technical assistance to con- Second, research is emerging indi- of alcohol misuse may heighten the cerned communities and researchers cating that alcohol may interact with resolve to study and implement effective so that they can collect their local and pharmacologically potentiate the interventions to reduce alcohol misuse. data using standardized questions and effects of other drugs, thereby increasing For example, 40 percent of people who sampling procedures for comparison risks of motor-vehicle crashes (Asbridge die in traffic crashes involving drinking with other communities, their State, et al. 2012; Li et al. 2012) and poisoning/ drivers in the United States are not and the Nation.

Figure 2 states participating in national Violent Death Registry (18 states)

souRCE: Centers for Disease Control and Prevention, national Violent Death Reporting system, 2013.

124 Alcohol Research: Current Reviews Chronic Conditions mortality consequences almost imme- overview of Measuring the When examining either acute-disease diately. This can be seen in the imme- Burden: Alcohol’s Evolving and chronic-disease mortality and diate gains in mortality and life impact morbidity, a variety of measurement expectancy in Russia following the challenges may produce underreporting. Gorbachev reforms that led to a reduc- This issue of ARCR examines the First, drinking levels reported in surveys tion of drinking (Leon et al. 1997). methodology involved in measuring account for only 40 to 60 percent of However, these immediate gains could the burden of alcohol use in greater alcohol sales (Midanik 1982; World be found for some chronic diseases, detail. Drs. Flynn and Wells (2013) Health Organization [WHO] 2011). but not for others, such as cancer. provide an overview on consumption Underreporting may lead to underesti- In cohort studies, drinking patterns indicators; environmental background mates of alcohol’s contribution to can vary in the same individual over indicators such as availability informa- chronic disease (Meier et al. 2013). time, and etiology studies vary in how tion; alcohol-attributable problems; Second, survey respondents often often someone drank over time and/or indicators for alcohol-attributable underestimate alcohol serving sizes, the intervals over time when drinking health outcomes (both chronic and particularly when consumed in contain- was measured. acute); and, last but not least, law ers that vary from accepted standard Third, maintaining high response enforcement indicators. They also drink sizes. Memory may become an rates in surveys and longitudinal studies make a case for triangulating different issue after respondents have consumed has become increasingly difficult over data sources in order to come to valid so many drinks so rapidly that they time, particularly using telephone conclusions as well as outline methods incur partial memory lapses or total methods, as the percentage of the and statistical techniques to technically blackouts. Also, the duration of time population who uses mobile phones integrate these data. that respondents are asked to recall increases. If nonresponse becomes high Dr. Cheryl Cherpitel focuses more consumption can vary in different and disproportionately involves people in depth on one of these data sources studies. In general, shorter time periods with characteristics and behaviors for the community in her examination of recall (e.g., days and weeks) produce (involving but not limited to alcohol of hospital emergency departments higher consumption, estimates than use that influence disease and injury (Cherpitel 2013). She describes not requests for monthly, yearly, or lifetime etiology), that may cloud our under- only the methodologies to make use of consumption. On the other hand, standing of alcohol’s role in the devel- these data, such as case-control or case- drinking patterns may vary over time opment and progression of disease. It crossover designs and their potential and even in the same year, prompting also can limit the ability of researchers biases, but also the use of such data to recommendations to use yearly recall to monitor disease and death-rate derive alcohol-attributable fractions, periods. Method or mode of data col- trends over time. which is a research topic in its own lection (i.e., face-to-face, telephone, Fourth, both in estimates of acute right (Shield et al. 2012a). mail, or Web based) also can influence and chronic conditions, attributable The next two chapters deal with drinking reports as well as response fractions from meta-analyses of epi- two other outcomes of alcohol use. rates and biases in survey samples. demiologic studies are used to estimate Dr. Shield and colleagues (2013b) Household-based surveys may not alcohol’s contribution to mortality and summarize findings on the impact include groups with high levels of alcohol disability. Yet, these attributable fractions of alcohol and chronic disease (i.e., consumption, such as students, the may change over time. For example, cancer, neuropsychiatric conditions, homeless, or people in institutions or the percentage of factual traffic-crash cardiovascular, and digestive diseases). in inpatient alcohol treatment facilities deaths that involve alcohol have dropped Methodologically, they discuss limita- (Meier et al. 2013; Stockwell et al. from 60 percent to just under 40 percent tions of current techniques used to 2004). Also, unrecorded alcohol, use in the past 30 years (NHTSA 2012). If derive risk relations and consequently, of alcohol in food, spillage, waste, and the most current epidemiologic studies attributable fractions. consumption by children and tourists are not used in alcohol-attributable Drs. Rehm and Shield (2013) may not be considered in surveys fraction estimates, the proportion of focus on mortality, more specifically on (Meier et al. 2013). acute and chronic disease mortality global estimates of alcohol-attributable Second, especially with chronic and morbidity attributed to alcohol mortality for the year 2010. They disease, in etiology studies the time may be inaccurate. report the causes of death with com- proximity of drinking data collection Fifth, when chronic disease mor- prise the overwhelming majority of to disease outcome must be carefully bidity and mortality attributions are all alcohol-attributable deaths: cancer, evaluated. Although some chronic dis- made, the range of diseases considered liver cirrhosis, and injury. Clearly, eases may take years to develop, cessa- may vary. Current U.S. estimates may cancer reflects the mortality-related tion or reduction of drinking may stop not fully consider alcohol’s role in alcohol use 15 to 20 years ago, liver the process and reduce morbidity or chronic diseases such as HIV. cirrhosis mainly current drinking but

Measuring the Burden: Alcohol’s Evolving impact 125 also a bit of the history, and injury been identified. Some of them may be ventions in the hospital setting were with the exception suicide mainly the further in the future, as there are links associated with a reduction of mortality level of current acute consumption between age of onset of alcohol use after one year of 40 percent, or Rehm (Holmes et al. 2012). and alcohol dependence or other con- et al. 2013, who found that increases No overview on alcohol use and sequences in later years in the United of the treatment rate to 40 percent in consequences would be complete without States (Donovan 2013; Grant and Europe could help avoid more than 10 mentioning the efforts to enumerate Dawson 1997). The most important percent of alcohol-attributable mortal- alcohol-attributable economic costs. consequence is alcohol-attributable ity). Again, more research is needed to The sidebar focuses on the last attempt death. Although this outcome is rela- understand the long-term consequences to estimate such costs for the United tively infrequent, a comparison to all- of interventions. States, focusing on heavy drinking cause deaths during this stage of life The issue concludes with contri- (Bouchery et al. 2010). shows that alcohol is the most important butions on the new Diagnostic and The first part of the volume is risk factor for mortality (and serious Statistical Manual of Mental Disorders–IV complimented by three sidebars. Dr. illness) (Rehm et al. 2006). (Grant 2013). Notes from a special Poznyak and colleagues (2013) give Other groups of concern high- NIAAA expert panel on alcohol and insight into the WHO system to col- lighted in this issue include women chronic diseases (Breslow and Mukamal lect data on alcohol consumption, (Wilsnack et al. 2013) and ethnic alcohol-attributable harm, and alcohol groups (Chartier et al. 2013). Although 2013) also are included and outline policy. Drs. Wiedermann and Frick women in all countries drink less, have future research opportunities for the (2013) describe the use of surveys to less heavy-drinking occasions, and field. ■ derive disability weights to calculate experience less alcohol-attributable harm the disability-adjusted life-years. Dr. than men (WHO 2011), this gap Hilton (2013) introduces an important seems to be closing in several countries Financial Disclosure national data bank—the NIAAA including the United States (Shield et The authors declare that they have no Alcohol Policy Information System— al. 2012b; Wilsnack et al. 2013). As for competing financial interests. for alcohol research. ethnicity and race, Native Americans, In the second part of the volume, Hispanics, and Blacks experience the focus is on alcohol use and its con- higher rates of alcohol-attributable References sequences over the lifespan. It starts harm than Whites in the United States. chronologically with use by children This is of course in part linked to dif- asbRiDgE, m.; hayDEn, J.a.; anD CaRtWRight, J.l. acute and adolescents (Donovan 2013; ferent drinking patterns (Chartier et al. cannabis consumption and motor vehicle collision risk: Patrick 2013). Another group, in part 2013; Shield et al. 2013a); but it also systematic review of observational studies and meta- defined by age, is college students may be worsened by an interaction analysis. British Medical Journal 344:E536, 2012. PmiD: (White 2013). All three groups again of socioeconomic status and alcohol 22323502 are characterized by specific method- (Schmidt et al. 2010). Drs. Chartier bouChERy, E.; simon, C.; anD haRWooD, h. Economic Costs ological problems. For example, regu- and colleagues show that more detailed of Excessive Alcohol Consumption in the United States, lar quantity–frequency measures do studies are needed, specifically on the 2006. Fallsview, Va: the lewin group, 2010. not capture alcohol use best, as con- mechanisms of alcohol’s impacts on bREsloW, R.a., anD muKamal, K.J. measuring the burden: sumption tends to vary. Also, especially health consequences in different eth- Current and future research trends: Results from the for high-school and younger students, nicities and races. niaaa expert panel on alcohol and chronic disease epidemiology. Alcohol Research: Current Reviews a lot of research is based on cross- Alcohol use disorders (AUDs) are 35(2):250–259, 2013. sectional studies, which makes causal one of the most important consequences conclusions impossible. More longitu- of alcohol use. Dr. Willenbring (2013) Centers for Disease Control and Prevention (CDC). alcohol-Related Disease impact (aRDi), 2013. available dinal studies are needed to start disen- examines some of the issues related to at: http://apps.nccd.Cdc.gov/ardi/homepage.aspx. tangling the web of the impact of measuring the public health impact of alcohol starting from earliest consump- AUDs and treatment. Although heavy ChaRtiER, K.g.; VaEth, P.a.C.; anD CaEtano, R. Ethnicity and the social and health harms from drinking. Alcohol tion. Such studies may append the drinking is responsible for the majority Research: Current Reviews 35(2):229–237, 2013. usual self-report measures with objec- of the alcohol-attributable burden of tive measures such as repeated BAL disease and mortality (for estimates, see ChERPitEl, C.J. Focus on: trauma and emergency out- comes. Alcohol Research: Current Reviews 35(2): for college students (for example, see Rehm et al. 2013), the public health 150–154, 2013. Thombs et al. 2009). impact of interventions—from screening For children, adolescents, and and brief interventions to treatment Cummings, P.; RiVaRa, F.P.; olson, C.m.; anD smith, K.m. Changes in traffic crash mortality rates attributed to use college students, despite the problems of alcohol dependence—is not fully of alcohol, or lack of a seat belt, air bag, motorcycle with establishing causality, a number of understood (for exceptions, see McQueen helmet, or bicycle helmet, united states, 1982-2001. alcohol-attributable consequences have et al. 2011, who found that brief inter- Injury Prevention 12:148–154, 2006. PmiD: 16751443

126 Alcohol Research: Current Reviews DonoVan, J.E. the burden of alcohol use: Focus on chil- geal cancers in great britain. Alcohol and Alcoholism shiElD, K.D.; gmEl, g., JR.; PatRa, J., Et al. global burden dren and preadolescents. Alcohol Research: Current 48(2):241–249, 2013. PmiD: 23345391 of injuries attributable to alcohol consumption in 2004: Reviews 35(2):186–192, 2013. a novel way of calculating the burden of injuries miDaniK, l. the validity of self-reported alcohol consump- attributable to alcohol consumption. Population Health FEll, J.C., anD Voas, R.b. mothers against drunk driving tion and alcohol problems: a literature review. British Metrics 10:9, 2012a. PmiD: 22607112 (maDD): the first 25 years. Traffic Injury Prevention 7: Journal of Addiction 77(4):357–382, 1982. PmiD: 195–212, 2006. PmiD: 16990233 6762224 shiElD, K.D.; KEhoE, t.; gmEl, g., JR.; Et al. societal bur- den of alcohol. in: anDERson, P.; mollER, l.; anD galEa, FERguson, s.a. alcohol-impaired driving in the united national highway traffic safety administration (nhtsa). g., EDs. Alcohol in the European Union: Consumption, Transitioning to Multiple Imputation: A New Method to states: Contributors to the problem and effective coun- Harm and Policy Approaches. Copenhagen, Denmark: Estimate Missing Blood Alcohol Concentration Values termeasures. Traffic Injury Prevention 13(5):427–441, World health organization Regional office For Europe, in FARS. Washington, DC: u.s. Department of 2012. PmiD: 22931172 2012b. transportation, 2002 (tech. Rep. no. Dot hs 809-403). Flynn, a., anD WElls, s. Current indicators: assessing shiElD, K.D.; PaRRy, C.; anD REhm, J. Chronic diseases and nhtsa. 2011 Motor Vehicle Crashes: Overview. the impact of alcohol use on communities. Alcohol conditions related to alcohol use. Alcohol Research: Washington, DC: u.s. Department of transportation, Research: Current Reviews 35(2):135–149, 2013. Current Reviews 35(2):155–173, 2013b. 2012 (tech. Rep. no. Dot hs 811 701). gRant, b.F., anD DaWson, D.a. age at onset of alcohol use shiElD, K.; gmEl, g.; KEhoE, t.; Et al. mortality and poten- nhtsa. Traffic Safety Facts 2010. Washington, DC: u.s. and its association with Dsm-iV alcohol abuse and tial years of life lost attributable to alcohol consumption Department of transportation, 2012 (tech. Rep. no. dependence: Results from the national longitudinal by race and sex in the united states in 2005. PLOS ONE Dot hs 811 659). alcohol Epidemiologic survey. Journal of Substance 8:E51923, 2013a. PmiD: 23300957 Abuse 9:103–110, 1997. PmiD: 9494942 PatRiCK, m.E., anD sChulEnbERg, J.E. Prevalence and pre- dictors of adolescent alcohol and binge drinking in the stoCKWEll, t.; Donath, s.; CooPER-stanbuRy, m.; Et al. hilton, m. aPis: the niaaa alcohol Policy information united states. Alcohol Research: Current Reviews under-reporting of alcohol consumption in household system. Alcohol Research: Current Reviews 35(2): 35(2):193–200, 2013. surveys: a comparison of quantity-frequency, graduated- 134–135, 2013. frequency and recent recall. Addiction 99(8):1024– PoznyaK, V.; FlEishman, a.; REKVE, D.; Et al. the World 1033, 2004. PmiD: 15265099 hingson, R., anD WhitE, a. Costs and consequences organization’s global monitoring system on alcohol and (morbidity and mortality) associated with adolescent health. Alcohol Research: Current Reviews 35(2): 244– thombs, D.l.; olDs, R.s.; bonDy, s.J., Et al. undergraduate and college drinking and related problems. in: Principles 249, 2013. drinking and academic performance: a prospective of Addiction: Comprehensive Addictive Behaviors and investigation with objective measures. Journal of Disorders. san Diego, Ca: academic Press, pp. REhm, J., anD shiElD, K.D. alcohol-attributable deaths. Studies on Alcohol and Drugs 70:776–785, 2009. 605–616, 2013. Alcohol Research: Current Reviews 35(2):174–183, 2013. PmiD: 19737503 holmEs, J.; mEiER, P.s.; booth, a.; Et al. the temporal rela- REhm, J.; baliunas, D.; boRgEs, g.l.g., Et al. the relation Voas, R.b.; toRREs, P.; Romano, E., Et al. alcohol-related tionship between per capita alcohol consumption and between different dimensions of alcohol consumption risk of driver fatalities: an update using 2007 data. harm: a systematic review of time lag specifications in and burden of disease: an overview. Addiction Journal of Studies on Alcohol and Drugs 73(3):341– aggregate time series analyses. Drug and Alcohol 105:817–843, 2010a. PmiD: 20331573 350, 2012. Dependence 123:7–14, 2012. PmiD: 22197480 REhm, J.; KEhoE, t.; gmEl, g., Et al. statistical modeling of WaltER, s.D. the estimation and interpretation of volume of alcohol exposure for epidemiological studies KEhoE, t.; gmEl, g., JR.; shiElD, K., Et al. Determining the attributable risk in health research. Biometrics of population health: the example of the u.s. popula- best population-level alcohol consumption model and 32:829–849, 1976. PmiD: 1009228 tion health. Metrics 8:3, 2010b. PmiD: 20202213 its impact on estimates of alcohol-attributable harms. WhitE, a., anD hingson, R. the burden of alcohol use: Population Health Metrics 10:6, 2012. PmiD: 22490226 REhm, J.; KlotsChE, J.; anD PatRa, J. Comparative quantifi- Excessive alcohol use and binge drinking in the united cation of alcohol exposure as risk factor for global bur- KlEin, t. Methods for Estimating Posterior BAC states. Alcohol Research: Current Reviews den of disease. International Journal of Methods in Distribution for Persons Involved in Fatal Crashes. 35(2):201–218, 2013. Psychiatric Research 16:66–176, 2007. PmiD: Washington, D.C.: national highway safety administration, 17623386 WhitE, R.; hingson, R; Pan, i.-J.; Et al. hospitalization for 1986 (tech. Rep. no. Dot hs 807– 904). alcohol and drug overdoses in young adults aged REhm, J.; shiElD, K. D.; REhm, m.X.; Et al. modelling the lEon, D.a.; ChEnEt, l.; shKolniKoV, V.m.; Et al. huge varia- 18–24 in the united states, 1999–2008: Results from impact of alcohol dependence on mortality burden and tion in Russian mortality rates 1984-94: artifact, alcohol, the nationwide inpatient sample. Journal of Studies on the effect of available treatment interventions in the Alcohol and Drugs 72(5):774-786, 2011. or what? Lancet 350(9075): 383–988, 1997. PmiD: European union. European Neuropsychopharmacology 9259651 23:89–97, 2013. PmiD: 22920734 WiEDERmann, W., anD FRiCK, u. using surveys to calculate disability-adjusted life-years. Alcohol Research: Current li, m.C.; bRaDy, J.E.; Dimaggio, C.J.; Et al. marijuana use REhm, J.; tayloR, b.; anD Room, R. global burden of dis- Reviews 35(2):128–134, 2013. and motor vehicle crashes. Epidemiologic Reviews ease from alcohol, illicit drugs and tobacco. Drug And 34(1):65–72, 2012. PmiD: 21976636 Alcohol Review 25:503–513, 2006. PmiD: 17132570 WillEnbRing, m. l. gaps in clinical prevention and treat- ment for alcohol use disorders: Costs, consequences, mCQuEEn, J.; hoWE, t. E.; allan, l.; Et al. brief interven- Room, R.; baboR, t.; anD REhm, J. alcohol and public and strategies. Alcohol Research: Current Reviews tions for heavy alcohol users admitted to general hospi- health: a review. Lancet 365:519–530, 2005. PmiD: 35(2):283–243, 2013. tal wards. Cochrane Database of Systematic Reviews 15705462 8:CD005191, 2011. PmiD: 21833953 WilsnaCK, s.C.; WilsnaCK, R.W.; anD WolFgang KantoR, l. sChmiDt, l.a.; mäKElä, P.; REhm, J., Et al. alcohol: Equity Focus on: Women and the costs of alcohol use. Alcohol mEiER, P.s.; mEng, y.; holmEs, J.; Et al. adjusting for and social determinants. in: blas, E., anD KuRuP, a.s., Research: Current Reviews 35(2):219–228, 2013. unrecorded consumption in survey and per capita sales EDs. Equity, Social Determinants and Public Health data: Quantification of impact on gender- and age-spe- Programmes. geneva, switzerland: World health World health organization (Who). Global Status Report cific alcohol-attributable fractions for oral and pharyn- organization, 2010. On Alcohol And Health. geneva, switzerland: Who, 2011.

Measuring the Burden: Alcohol’s Evolving impact 127 Using Surveys to Calculate Disability-Adjusted Life-Years

Wolfgang Wiedermann, Ph.D., and Ulrich Frick, Ph.D.

Wolfgang Wiedermann, Ph.D., mapping a certain disease into a Mapping a certain disease into a is a junior researcher at the system of disabling attributes allows system of disabling attributes (e.g., Carinthia University of Applied researchers to compare diseases physical functioning, pain, memory Sciences, Feldkirchen, Austria. within a common framework. to and thinking, etc.) enables health quantify the total burden of morbidity researchers to compare qualitatively Ulrich Frick, Ph.D., is a professor (e.g., morbidity attributable to alcohol different diseases within a common framework. To quantify the total bur- of public health at the Carinthia use), so-called disability weights (DWs) must be generated. general- den of alcohol-attributable morbidity, University of Applied Sciences, it is necessary to provide so-called Feldkirchen, Austria, and senior population surveys can be used to derive DWs from health valuation disability weights (DWs) for each of scientist at the Psychiatric these health states, which are bounded tasks. this article describes the University Hospital, University by the DWs of 0 (for complete health) application of three psychometric and 1 (for death). It should be noted of Regensburg, Regensburg, methods (i.e., pairwise comparisons, Germany. that health states are considered, rather ranking tasks, and visual analog than diseases with labels (and their scales) in general-population surveys psychological and/or medical impli- and outlines their strengths and cations), when DWs are determined. weaknesses. a recently proposed How DWs can validly be mea- health valuation framework also is sured, defined, or (more neutrally presented, which highlights the speaking) elicited is of equal impor- underlying cognitive processes from a tance for the results as the question, social-judgment perspective and “Who is asked to provide the DWs?” presents a structured data-collection Although elicitation methods will be procedure that seems promising in discussed below, this article does not deriving DWs from general-population focus on the question of which sources surveys. (e.g., patients, clinical experts, etc.) should be consulted to quantify DWs. Rather, this article considers only o quantify the burden of a disease within a population, a health- general-population surveys (i.e., tele- tgap measure is more useful than phone, face to face, or mailed) as measures of health expectancy or sources of information on the disabilities quality-adjusted life-years (see Etches associated with different health states. et al. 2006). Disability-adjusted life- years (DALYs), the most prominent of the health-gap measures, combine How Are DWs Elicited? the burden attributable to early death and to morbidity into one single Three popular methods to construct number. Alcohol affects a long list of DWs stem from econometric utility diseases and disabilities in varying theory: standard gamble (SG), time intensities, each of which can be tradeoff (TTO), and person tradeoff described by a number of health-state (PTO). They all share the central attributes. Common measures of health idea that a respondent’s point of outcomes include the EuroQol5D indifference, at which he or she cannot (EQ5D) (Brooks and EuroQol Group unequivocally decide on a certain 1996), the Health Utilities Index III (HUI III) (Feeny et al. 2002), the judgmental task, enables researchers Short-Form 36 Health Survey (SF36) to measure utility differences via the (Ware and Sherbourne 1992), and traded “goods.” For example, in SG, the CLAssification and MEasurement respondents are given a choice System of Functional Health between an outcome that is certain (CLAMES) (McIntosh et al. 2007). (i.e., remaining in ill health) and a

128 Alcohol Research: Current Reviews gamble with one better and one and de Pouvourville 2011; Dolan As alternatives to the methods worse outcome (e.g., full health or 1997; Greiner et al. 2005; Jelsma et described above, psychometric theory death). Respondents are asked what al. 2003; Jo et al. 2008; Lamers et al. provides paired comparisons, ranking probability of the better outcome 2006; Lee et al. 2009; Shaw et al. tasks, and visual analogue scales as would make them indifferent to 2010; Tsuchiya et al. 2002; Wittrup- tools to elicit health-state preferences. remaining in the described state (ill Jensen et al. 2009; Zarate et al. 2008), These tools are discussed below. health) for certain or choosing the in mail surveys only two studies risky option. Therefore, if they are (Burström et al. 2006; Lundberg et indifferent to the ill-health state and al. 1999) used TTO to quantify Paired Comparisons gamble with a 0.8 probability of the respondents’ own health states. SG better outcome (but 0.2 probability and PTO have rarely been used for In the context of health-state valua- of the worse outcome), 0.8 repre- eliciting health-state preferences in tion, a paired comparison (PC) task sents the utility of the ill health. mailed surveys (i.e., they are usually simply means that respondents must In a TTO task, respondents are used in face-to-face or phone inter- choose which of two given states is asked to consider the relative amounts views), and they have only been used more disabling, worse, or dominant of time (e.g., number of life-years) among former patients (i.e., not the in some way. Because measuring via they would be willing to sacrifice to general public) (Hammerschmidt et PC seems quite simple and feasible avoid a certain poorer health state al. 2004). (because it is only necessary to pre- (e.g., frequent headaches). Assuming Readers are referred to Rehm sent all health states in a consistent a scenario of 10 years with frequent and Frick (2010) for an overview on descriptive system), it has been headaches, the respondent may be the methodological problems associ- applied in various surveys in the gen- indifferent to this state and a shorter ated with econometric elicitation eral population (Bijlenga et al. 2009; lifetime of 7 years, resulting in an methods in this context. Recently, Kind 1982, 2005; Prieto and Alonso estimated utility for the frequent- Wittenberg and Prosser (2011) 2000; Ratcliffe et al. 2009; Stolk et al. headaches health state of 0.7 (7 years described two additional sources of 2010). For a recent application of divided by 10 years). bias or mistaken responses in prefer- PCs among an expert panel see Rehm A typical PTO elicitation asks ence measurement in surveys: order- and Frick (2013). Deriving DWs respondents to choose between two ing errors (i.e., illogical responses, from the resulting pattern of domi- equally expensive health care treat- which violate a naturally given order, nance relations, by contrast, constitutes ment programs that improve quality whereas inconsistent responses con- a complex statistical task for which of life or save lives for two groups of tradict each other within a person), solutions have been formulated from patients. The decisionmaker must and objections/invariance (i.e., the theory of Thurstone scaling choose to fund one of the two mutu- respondents may refuse to participate (Thurstone 1927), conditional logis- ally exclusive programs, one of which because of an unwillingness to trade tic regression (Hosmer and Lemeshow has a fixed number of patients. time [in the TTO task] or risk [in 2000), and loglinear modeling Respondents are asked how many the SG task]). Furthermore, the mean- (Critchlow and Fligner 1991). patients would need to be treated to ing of SG results has been criticized Methodological challenges asso- make them indifferent to the two as rather a measurement of risk attitude ciated with PC stem from logically programs. For example, program A than a representation of subjective inconsistent judgments (e.g., A > B might extend the life of 100 healthy utility (Lenert and Kaplan 2000). TTO and B > C, but C > A) and from individuals for 1 year, whereas pro- results as a metric for utility have rapidly increasing burden of task gram B might cure 100 individuals been shown to vary with respon- when comparing larger numbers of a chronic health condition. dents’ age, education, and current of health states (i.e., combinatorial All three methods are time con- health state (Ayalon and King- explosion). Intransitive judgments suming, require highly motivated Kallimanis 2010; Meropol et al. 2008; (e.g., in comparing 10, 7, and 5, 10 respondents, and are hardly feasible Stiggelbout et al. 1996; Voogt et al. is preferred to 7 and 7 is preferred to without a trained interviewer or 2005). Feasibility of PTO frequently 5, but 5 is preferred to 10) may orig- computer program. Whereas TTO is hampered because people tend to inate in unintended framing effects has been used in face-to-face interviews refuse such tasks because of their desire as well as in imperfect judgment (von in the general population quite often to avoid prejudice and discrimina- Winterfeldt and Edwards 1986). (e.g., Badia et al. 2001; Chevalier tion (Damschroder et al. 2005). Recently published experimental stud-

Using Surveys to Calculate Disability-Adjusted Life-Years 129 ies favor the position that excluding From a more theoretical view- visual Analog Scale inconsistent ratings cannot improve point, articles by Flynn and col- the description of true preferences leagues (2010) and Flynn (2010) To use a visual analog scale (VAS), and therefore might to some degree have raised serious statistical concerns respondents are asked to specify their be an inevitable consequence of the about the use of ranks as a substitu- level of agreement to a statement by decisionmaking process itself (Linares tion for econometric valuation tasks. indicating a position along a contin- 2009). To keep the number of judg- Their critique focuses on modeling uous line between two endpoints. ments at manageable dimensions, assumptions and thus seems beyond Numerous studies have used VAS several studies have used incomplete the scope of this article. Nevertheless, responses to derive health-state values factorial designs (Bijlenga et al. their argument suggests that it can in the general population (Björk and 2009; Prieto and Alonso 2000; be important to restrict the number Norinder 1999; Cleemput 2010; Ratcliffe et al. 2009). of alternatives to be ranked and to pay Devlin et al. 2003; Dolan and Kind Asking subjects to rank order special attention to how a respondent 1996; Essink- Bot et al. 1993; several health states, but statistically generates rankings. In addition, Lenert Greiner et al. 2003; Johnson and analyzing rankings as PCs, was used and colleagues (1998) have demon- Pickard 2000; Johnson et al. 1998; as an alternative in several studies strated that reported utilities are Leidl and Reitmeir 2011). Krabbe (Krabbe 2008; Ip et al. 2004). Rankings heavily influenced by the search pro- and colleagues (2007) proposed a can be transformed into a series of cess used to form a certain judgment. methodology based on differences in VAS values, where the ranks of PCs (Francis et al. 2002), which at first This matches the notion that prefer- pairwise VAS differences are used in glance avoids inconsistent judgments. ences often are constructed (instead a multidimensional scaling analysis of merely obtained) in the elicitation to estimate cardinal health-state val- process (Slovic 1995). Ranking tasks Ranking tasks ues. However, other researchers have within self-administered question- questioned the validity of VAS data Health-state rankings (i.e., putting naires might be hampered by limited as cardinal values (Bleichrodt and several health states into an ordinal control of the mechanism respon- Johannesson 1997; Devlin et al. sequence of disability), which also dents use to generate the rank order. 2004; van Osch and Stiggelbout provide comparative information, This introduces at least two issues: 2005) for various reasons. First, VAS require less cognitive effort for survey First, it remains unclear which refer- tasks in which the top and the bottom respondents. Furthermore, simulta- ence attributes the respondent uses endpoints are precisely defined (e.g., neous comparisons of multiple health to generate the rank order, which death versus perfect health) allow states might be less sensitive to biases constricts intersubjective comparability direct comparison between individuals, (e.g., those provoked by arbitrarily and provokes primacy biases (i.e., the whereas vague labels such as “worst labeled endpoints of rating scales) tendency to give more attention to imaginable” and “best imaginable” (Maydeu-Olivares and Böckenholt items listed first) (Bowling 2005). hamper an interindividual compari- 2008). Although ranking exercises Second, from a more technical per- son (Torrance et al. 2001). Second, had been included in numerous val- spective, statistical ranking models VAS responses might be affected by uation studies as an external compar- (such as the rank-ordered logit a so-called end-aversion bias, the ison measure for TTO and SG, model) assume that rankings were phenomenon of respondents tending researchers had not used the resulting obtained using a particular psycho- to be reluctant to mark positions ordinal data (McCabe et al. 2006) logical mechanism (Flynn 2010). near the endpoints of the scale for construction of DWs before the For free rankings, however, it remains (Bleichrodt and Johannesson 1997; seminal article by Salomon (2003). unclear which statistical model is Robinson et al. 2001; Torrance et al. Cardinal utilities derived from most appropriate to describe the 2001). Third, a VAS score for a cer- health-state rankings displayed high ranking mechanism. Furthermore, it tain health state may depend on other agreement to utilities from TTO or cannot be ensured that respondents states presented at the same time SG methods (Craig et al. 2009a, b; using a self-administered question- (i.e., context bias) (Torrance et al. Kind 2005) and were more stable naire judge along repeated best/worst 2001). Fourth, the accuracy of VAS in a cross-cultural comparison than choices, a “ping-pong” method that responses may be influenced by hand weights derived from SG (Ferreira et was shown to produce reliable data preferences and which hand was al. 2011). (Louviere at el. 2008). used (McKechnie and Brodie 2008).

130 Alcohol Research: Current Reviews Finally, the orientation of the VAS of a certain disease), and contrast 4. A deliberate editing of the scale (vertical versus horizontal) itself effects (i.e., severely ill patients may response. In this last step, for might affect the shape of the result- underestimate the impact of lenient example, a respondent’s attempt to ing score distribution (e.g., Lundqvist be compatible with perceived norms diseases, while healthy people may et al. 2009). Taken together, VAS (e.g., perceived fairness, political responses therefore should be inter- overestimate this impact). Conducting correctness, or ethical considerations) preted on an ordinal scale level only. a survey in the general public will further biases a subjective valuation result in a weighted mixture of affected (Rehm and Frick 2010). Health valuation: A Social and healthy valuation perspective. Judgment Perspective 2. Interpretation/primary appraisal. Conclusion Stiggelbout and de Vogel-Voogt (2008) The interpretation of a health state Econometric elicitation methods presented a four-step framework depends on a subject’s values, goals, were not originally developed for describing respondents’ cognitive and beliefs, as well as on the cognitive self-administered questionnaires. processes while valuing health states: Given the many methodological framing (Kahneman and Tversky perspective/perception of the stimulus, risks of using this data collection interpretation, judgment, and forma- 1984) and/or context (Schwarz 1999) mode, TTO, PTO, or SG elicitation tion of a manifest response (see also of the health-state description. methods are not recommended for Rehm and Frick 2010). For each step, paper-and-pencil surveys. Under- several mechanisms have been identi- standing the introductory scenarios fied, which may affect the final response. 3. Judgments on health states. and autonomously and successively Like human judgments in general, approaching the point of indiffer- 1. Perspective/perception of the these are not formed to fulfill the ence seems too complicated a task stimulus. In a meta-analysis, criteria of an exhaustive information for lay respondents. Though VAS Dolders and colleagues (2006) scales were developed specifically for reported no significant differences processing. By contrast, they serve self-administered questionnaires, in preferences when patient surveys as decision rules to govern behavior their validity and reliability are too were compared with those of the (e.g., giving an answer in a questionnaire) weak to measure the utilities of com- general public, whereas a more and follow the principles of parsimony plex health states on the interval recent and more extensive meta- and functional pragmatism rather level. Choosing between rankings analysis by Peeters and Stiggelbout and PC tasks would mean a tradeoff (2010) suggests that patients differ than coherence and rationality. between economy and validity of the from the general public in their Stiggelbout and de Vogel-Voogt measurement procedure. valuations. Frick and colleagues (2008) identified various sources of Among PCs presented to (2012) reported on the importance biases that might be relevant in the respondents from the general public, of social relationships as determinants those with the following characteris- of health valuation, especially for context of health valuation, such as tics seem to be most promising: (1) health professionals. Health states focusing illusion (see step 1), status The number of pairs of health states hampering social relationships are quo bias (i.e., respondents are more should be limited (to a number judged as more disabling. Ubel and sensitive to changes in their own determined by pre-analysis) so that colleagues (2003) described several health state compared with imagined annoyance effects or reactance can factors that may contribute to these mostly be precluded. (2) Cognitive discrepancies: adaptation effects health states), loss aversion (see complexity of the health state (i.e., affected patients often adapt Tversky and Kahneman 1992), or descriptions should not exceed seven physically and emotionally to their failure to anticipate negative events (plus or minus two) judgmental health state, resulting in a more (i.e., poor hedonic forecasting). In attributes (Miller 1956). However, positive valuation of the respective this does not necessarily mean that state), focusing illusion (i.e., healthy addition, affects and mood are health-state descriptions should be people focus on impaired attributes, known to be highly influential limited to seven dimensions or largely ignoring unchanged attributes during judgmental processing. attributes, as respondents tend to

Using Surveys to Calculate Disability-Adjusted Life-Years 131 organize redundant information into boWling, a. mode of questionnaire administration can EtChEs, V.; FRanK, J.; Di RuggiERo, E.; anD manuEl, D. broader superconcepts. That being have serious effects on data quality. Journal of Public measuring population health: a review of indicators. Health 27:281–291, 2005. PmiD: 15870099 Annual Review of Public Health 27:29–55, 2006. said, this ability should also be evalu- PmiD: 16533108 ated prior to the survey. Applying bRooKs, R. Euroqol: the current state of play. Health Policy 37:53–72, 1996. PmiD: 10158943 FEEny, D.; FuRlong, W.; toRRanCE, g.W.; Et al. these principles would allow surveys multiattribute and single-attribute utility functions for buRstRöm, K.; JohannEsson, m.; anD DiDERiChsEn, F. a to pose complex vignettes to respon- the health utilities index mark 3 system. Medical Care comparison of individual and social time trade-off dents. (3) To avoid biases due to the 40:113–128, 2002. PmiD: 11802084 values for health-states in the general population. direction of a comparison (e.g., A Health Policy 76:359–370, 2006. PmiD: 16214258 FERREiRa, l.n.; FERREiRa, P.l.; RoWEn, D.; anD bRaziER, J.E. versus B is not the same as B versus Do Portuguese and uK health state values differ ChEValiER, J., anD DE PouVouRVillE, g. Valuing EQ-5D A) (Wänke 1996), presentation of using time trade-off in France. European Journal of across valuation methods? Quality of Life Research health states within one comparison Health Economics 14(1):57–66, 2013. PmiD: 21935715 20:609–619, 2011. PmiD: 21061071 should be randomly balanced. To Flynn, t.n. using conjoint analysis and choice experi- ClEEmPut, i. a social preference valuations set for EQ- avoid order effects or carryover effects, 5D health states in Flanders, belgium. European Journal ments to estimate Qaly values: issues to consider. factorial design techniques that also of Health Economics 11:205–213, 2010. PmiD: 19582490 Pharmacoeconomics 28:711–722, 2010. PmiD: preclude repetitive presentations of 20568837 CRaig, b.m.; bussChbaCh, J.J.; anD salomon, J.a. Keep it certain health states (A versus B fol- simple: Ranking health states yields similar to cardi- Flynn, t.n.; louViERE, J.J.; PEtERs, t.J.; anD Coast, J. lowed by C versus D and not by A nal measurement approaches. Journal of Clinical using discrete choice experiments to understand versus C, for instance) should be Epidemiology 62:296–305, 2009a. PmiD: 18945585 preferences for quality of life: Variance-scale hetero- geneity matters. Social Science & Medicine 70:1957– used in the assignment of comparison CRaig, b.m.; bussChbaCh, J.J.; anD salomon, J.a. 1965, 2010. PmiD: 20382460 tasks to respondents. Complex survey modeling ranking, time trade-off, and visual analogue designs like the one proposed here scale values for EQ-5D health states: a review and FRanCis, b.; DittRiCh, R.; hatzingER, R.; anD PEnn, R. comparison of methods. Medical Care 47:634–641, analysing partial ranks by using smoothed paired require adequate techniques for sta- 2009b. PmiD: 19433996 comparison methods: an investigation of value orien- tistical analysis (Hox et al. 1991). ■ tation in Europe. Applied Statistics 51:319–336, 2002. CRitChloW, D.E., anD FlignER, m.a. Paired comparison, triple comparison, and ranking experiments as gener- FRiCK, u.; iRVing, h.; anD REhm, J. social relationships alized linear models, and their implementation in as a major determinant in the valuation of health Financial Disclosure glim. Psychometrika 56:517–533, 1991. states. Quality of Life Research 21:209–213, 2012. PmiD: 21633877 The authors declare that they have DamsChRoDER, l.J.; RobERts, t.R.; golDstEin, C.C.; Et al. trading people versus trading time: What is the differ- gREinER, W.; WEiJnEn, t.; niEuWEnhuizEn, m.; Et al. a sin- no competing financial interests. ence? Population Health Metrics 3:10, 2005. PmiD: gle European currency for EQ-5D health states: 16281982 Results from a six-country study. European Journal of Health Economics 4:222–231, 2003. PmiD: 15609189 DEVlin, n.J.; hansEn, P.; KinD, P.; anD Williams, a. References logical inconsistencies in survey respondents’ health gREinER, W.; ClaEs, C.; bussChbaCh, J.J.; anD Von DER state valuations: a methodological challenge for esti- sChulEnbuRg, J.m. Validating the EQ-5D with time trade ayalon, l., anD King-Kallimanis, b.l. trading years for mating social tariffs. Health Economics 12:529–544, off for the german population. European Journal of perfect health: Results from the health and Retirement 2003. PmiD: 12825206 Health Economics 6:124–130, 2005. PmiD: 19787848 study. Journal of Aging and Health 22:1184–1197, 2010. PmiD: 20660638 DEVlin, n.J.; hansEn, P.; anD sElai, C. understanding hammERsChmiDt, t.; zEitlER, h.P.; guliCh, m.; anD lEiDl, R. health state valuations: a qualitative analysis of a comparison of different strategies to collect stan- baDia, X.; RosEt, m.; hERDman, m.; anD KinD, P. a compari- respondents’ comments. Quality of Life Research dard gamble utilities. Medical Decision Making son of united Kingdom and spanish general popula- 13:1265–1277, 2004. PmiD: 15473505 24:493–503, 2004. PmiD: 15358998 tion time trade-off values for EQ-5D health states. Medical Decision Making 21:7–16, 2001. PmiD: Dolan, P. modeling valuations for EuroQol health states. hosmER, D.W., anD lEmEshoW, s. Applied Logistic 11206949 Medical Care 35:1095–1108, 1997. PmiD: 9366889 Regression (2nd ed.). new york: Wiley and sons, 2000.

biJlEnga, D.; biRniE, E.; anD bonsEl, g.J. Feasibility, Dolan, P., anD KinD, P. inconsistency and health state hoX, J.J.; KREFt, i.g.g.; anD hERmKEns, P.l.J. the analy- reliability and validity of three health-state valuation valuations. Social Science & Medicine 42:609–615, sis of factorial surveys. Sociological Methods & methods using multiple-outcome vignettes on moder- 1996. PmiD: 8643985 Research 19:493–510, 1991. ate-risk pregnancy at term. Value in Health 12:821– DolDERs, m.g.; zEEgERs, m.P.; gRoot, W.; anD amEnt, a. 827, 2009. PmiD: 19508667 iP, W.C.; Chiu, l.l.; anD KWan, y.K. Construction of a meta-analysis demonstrates no significant differ- health indices using paired comparisons. Social bJöRK, s., anD noRinDER, a. a weighting exercise for the ences between patient and population preferences. Indicators Research 67:353–373, 2004. swedish version of the EuroQol. Health Economics Journal of Clinical Epidemiology 59:653–664, 2006. 8:117–126, 1999. PmiD: 10342725 PmiD: 16765267 JElsma, J.; hansEn, K.; DE WEERDt, W.; Et al. how do zimbabweans value health states? Population Health blEiChRoDt, h., anD JohannEsson, m. an experimental EssinK-bot, m.l.; stouthaRD, m.E.; anD bonsEl, g.J. Metrics 1:11, 2003. PmiD: 14678566 test of a theoretical foundation for rating-scale valua- generalizability of valuations on health states collect- tions. Medical Decision Making 17:208–216, 1997. ed with the EuroQol questionnaire. Health Economics Jo, m.W.; yun, s.C.; anD lEE, s.i. Estimating quality PmiD: 9107617 2:237–246, 1993. PmiD: 8275169 weights for EQ-5D health states with the time trade-

132 Alcohol Research: Current Reviews off method in south Korea. Value in Health 11:1186– lunDbERg, l.; JohannEsson, m.; isaCson, D.g.; anD Robinson a.; loomEs g.; anD JonEs-lEE, m. Visual ana- 1189, 2008. PmiD: 18489498 boRgQuist, l. the relationship between health-state log scales, standard gambles, and relative risk aver- utilities and sF-12 in a general population. Medical sion. Medical Decision Making 21:17–27, 2001. Johnson, J.a., anD PiCKaRD, a.s. Comparison of the EQ- Decision Making 19:128–140, 1999. PmiD: 10231075 PmiD: 11206943 5D and sF-12 health surveys in a general population survey in alberta, Canada. Medical Care 38:115– lunDQVist, C.; bEnth, J.s.; gRanDE, R.b.; Et al. a vertical salomon, J.a. Reconsidering the use of rankings in 121, 2000. PmiD: 10630726 Vas is a valid instrument for monitoring headache the valuation of health states: a model estimating pain intensity. Cephalalgia 29:1034–1041, 2009. cardinal values from ordinal data. Population Health Johnson, J.a.; Coons, s.J.; ERgo, a.; anD szaVa-KoVats, Metrics 1:12, 2003. PmiD: 14687419 g. Valuation of EuroQol (EQ-5D) health states in an PmiD: 19735531 adult us sample. Pharmacoeconomics 13:421–433, mayDEu-oliVaREs, a., anD böCKEnholt, u. modeling sub- sChWaRz, n. self-reports: how the question shape the 1998. PmiD: 10178666 jective health outcomes: top 10 reasons to use answers. American Psychologist 54:93–105, 1999. thurstone’s method. Medical Care 46:346–348, KahnEman, D., anD tVERsKy, a. Choices, values, and shaW, J.W.; PiCKaRD, a.s.; yu, s.; Et al. a median frames. American Psychologist 39:341–350, 1984. 2008. PmiD: 18362812 model for predicting united states population-based EQ-5D health state preferences. Value in Health KinD, P. a comparison of two models for scaling mCCabE C.; bRaziER, J.; gilKs, P.; Et al. using rank data health indicators. International Journal of Epidemiology to estimate health state utility models. Journal of Health 13:278–288, 2010. PmiD: 19961566

11:271–275, 1982. PmiD: 7129741 Economics 25:418–431, 2006. PmiD: 16499981 sloViC, P. the construction of preference. American Psychologist 50:364–371, 1995. KinD, P. applying paired comparisons models to EQ- mCintosh , C.n.; goRbER, s.C.; bERniER, J.; anD bERthElot,

5D valuations: Deriving tto utilities from ordinal pref- J.m. Eliciting Canadian population preferences for stiggElbout, a.m.; DE haEs, J.C.; KiEbERt, g.m.; Et al. erence data. in: Kind, P., brooks, R., and Rabin R., health states using Classification and measurement tradeoffs between quality and quantity of life: Eds. EQ-5D Concepts and Methods: A Developmental system of Functional health (ClamEs). Chronic Development of the QQ questionnaire for cancer History. Dordrecht, the netherlands: springer, 2005, Diseases in Canada 28:29–41, 2007. PmiD: 17953796 patient attitudes. Medical Decision Making pp. 201–220. 16:184–192, 1996. PmiD: 8778537 mCKEChniE, J.g., anD bRoDiE, E.E. hand and hand pref- KRabbE, P.F. thurstone scaling as a measurement erences in use of a visual analogue scale. Perceptual stiggElbout, a.m., anD DE VogEl-Voogt, E. health state method to quantify subjective health outcomes. and Motor Skills 107:643–650, 2008. PmiD: 19235396 utilities: a framework for studying the gap between Medical Care 46:357–365, 2008. PmiD: 18362814 the imagined and the real. Value in Health mERoPol, n.J.; EglEston, b.l.; buzaglo, J.s.; Et al. KRabbE, P.F.; salomon, J.a., anD muRRay, C.J. 11:76–87, 2008. PmiD: 18237362 Cancer patient preferences for quality and length of Quantification of health states with rank-based non- life. Cancer 113:3459–3466, 2008. PmiD: 18988231 stolK, E.a.; oPPE, m.; sCalonE, l.; anD KRabbE, P.F. metric multidimensional scaling. Medical Decision Discrete choice modeling for the quantification of Making 27:395–405, 2007. PmiD: 17761959 millER, g.a. the magical number seven, plus or health states: the case of EQ-5D. Value in Health minus two: some limits on our capacity for process- lamERs, l.m.; mCDonnEll, J.; stalmEiER, P.F.; Et al. the 13:1005–1013, 2010. PmiD: 20825618 ing information. Psychological Review 63:81–97, Dutch tariff: Results and arguments for an effective 1956. PmiD: 13310704 thuRstonE, l.l. a law of comparative judgment. design for national EQ-5D valuation studies. Health Psychological Review 34:273–286, 1927. Economics 15:1121–1132, 2006. PmiD: 16786549 PEEtERs, y., anD stiggElbout, a.m. health state valua- tions of patients and the general public analytically toRRanCE, g.W.; FEEny, D.; anD FuRlong, W. Visual ana- lEE, y.K.; nam, h.s.; Chuang, l.h.; Et al. south Korean log scales: Do they have a role in the measurement time trade-off values for EQ-5D health states: modeling compared: a meta-analytical comparison of patient with observed values for 101 health states. Value in and population health state utilities. Value in Health of preferences for health states? Medical Decision Health 12:1187–1193, 2009. PmiD: 19659703 13:306–309, 2010. PmiD: 19744288 Making 21:329–334, 2001. PmiD: 11475389 tsuChiya, a.; iKEDa, s.; iKEgami, n.; Et al. Estimating an lEiDl, R., anD REitmEiR, P. a value set for the EQ-5D PRiEto, l., anD alonso, J. Exploring health preferences based on experienced health states: Development in sociodemographic and health related groups EQ-5D population value set: the case of Japan. Health and testing for the german population. Pharma - through the paired comparison of the items of the Economics 11:341–353, 2002. PmiD: 12007165 nottingham health Profile. Journal of Epidemiology coeconomics 29:521–534, 2011. PmiD: 21247225 tVERsKy, a., anD KahnEman, D. advances in prospect and Community Health 54:537–543, 2000. PmiD: lEnERt, l., anD KaPlan, R.m. Validity and interpretation theory: Cumulative representation of uncertainty. 10846197 of preference-based measures of health-related qual- Journal of Risk and Uncertainty 5:297–323, 1992. ity of life. Medical Care 38(9 suppl.):ii 138–ii 150, RatCliFFE, J.; bRaziER, J.; tsuChiya, a.; Et al. using DCE ubEl, P.a.; loEWEnstEin, g.; anD JEPson, C. Whose qual- 2000. PmiD: 10982099 and ranking data to estimate cardinal values for ity of life? a commentary exploring discrepancies health states for deriving a preference-based single lEnERt, l.a.; ChER, D.J.; golDstEin, m.K.; Et al. the effect between health state evaluations of patients and the of search procedures on utility elicitations. Medical index from the sexual quality of life questionnaire. Health general public. Quality of Life Research 12:599–607, Decision Making 18:76–83, 1998. PmiD: 9456212 Economics 18:1261–1276, 2009. PmiD: 19142985 2003. PmiD: 14516169 linaREs, P. are inconsistent decisions better? an REhm, J., anD FRiCK, u. Valuation of health states in the Van osCh, s.m.C., anD stiggElbout, a.m. understanding experiment with pairwise comparisons. European us study to establish disability weights: lessons from Vas valuations: Qualitative data on the cognitive pro- Journal of Operational Research 193:492–498, 2009. the literature. International Journal of Methods in cess. Quality of Life Research 14:2171–2175, 2005. Psychiatric Research 19:18–33, 2010. PmiD: 20191661 PmiD: 16328897 louViERE, J.J.; stREEt, D.; buRgEss, l.; Et al. modeling the choices of individual decision-makers by combin- REhm, J., anD FRiCK, u. Establishing disability weights Voogt, E.; Van DER hEiDE, a.; RiEtJEns, J.a.; Et al. attitudes ing efficient choice experiment designs with extra from pairwise comparisons for a us burden of dis- of patients with incurable cancer toward medical treat- preference information. Journal of Choice Modelling ease study. International Journal of Methods in ment in the last phase of life. Journal of Clinical 1:128–163, 2008. Psychiatric Research 22(2):144–154, 2013. Oncology 23:2012–2019, 2005. PmiD: 15774792

Using Surveys to Calculate Disability-Adjusted Life-Years 133 nomic, health, and social costs to individuals, communities, Community indicators and societies (Rehm et al. 2009). Alcohol intoxication is linked to injury, violence, and traffic crashes (Edwards et al. 1994) and chronic alcohol use increases the risk of liver damage and various cancers, among other health harms Assessing the Impact of Alcohol (Edwards et al. 1994; Rehm et al. 2003; Room et al. 2005). National surveys have revealed a great deal of variability use on Communities across different communities in the extent of alcohol use and related harms (Gruenewald et al. 1997). Thus, it may not be practical or fiscally responsible to base local prevention and intervention initiatives on national data that do not reflect patterns or problems within a particular community. Moreover, Andrea Flynn, Ph.D., and Samantha Wells, Ph.D. prevention, treatment, and enforcement activities are com- monly enacted at the local level (Gruenewald et al. 1997). Therefore, community-level data on the impact of alcohol Community indicators are used to assess the impact of alcohol use that take into consideration the local economic, social, on communities. this article reviews the main data sources for and policy context are key to guiding local decisionmaking community indicators, discusses their strengths and limitations, and maximizing the effectiveness of prevention and interven- and discusses indicators used in reference to four main topics tion approaches. relating to alcohol use and problems at the community level: Community indicators have been used extensively for alcohol use, patterns, and problems; alcohol availability; a variety of purposes by both researchers and community alcohol-related health outcomes/trauma; and alcohol-related stakeholders. For communities, indicator data can be used crime and enforcement. it also reviews the challenges to inform priority-setting agendas by identifying specific associated with collecting community indicator data, along with concerns within a community, guide policy and education important innovations in the field that have contributed to better initiatives, monitor community status on a particular mea- knowledge of how to collect and analyze community-level data sure over time or in comparison with other communities, on the impact of alcohol. kEY WoRDS: Alcohol use, abuse, and and evaluate programs or policies (Besleme and Mullin dependence; alcohol burden; problematic alcohol use; 1997; Gabriel 1997; Gruenewald et al. 1997; Mansfield harmful drinking; alcohol-related harm; alcohol use patterns; and Wilson 2008; Metzler et al. 2008). Local-level data also alcohol effects and consequences; alcohol availability; risk are critical for justifying requests for funding and provide a factors; environmental impact; crime; community indicators; powerful tool for resource allocation within communities community monitoring; community epidemiology; data (Mansfield and Wilson 2008). For researchers, community collection; public policy on alcohol indicators are central for improving knowledge of factors influencing community well-being, advancing innovative n the United States and other countries around the world, theoretical models and analytical approaches for use in researchers have long been interested in community-level research and prevention planning (for example, see Holder imeasurement of population health in the form of com- 1998a), and monitoring and evaluating community prevention/ munity indicators. Community indicators are measures that intervention initiatives (Metzler et al. 2008). communicate information about a given dimension of a This article provides an overview of community indicators community’s well-being (Besleme and Mullin 1997). In the of alcohol use and related harms, outlining common sources United States, the current popularity of community indicators can be traced back to the social-indicators movement of the 1960s and 1970s (see Gross and Straussman 1974; Land and Spilerman 1975; MacRae 1985), which saw growing Andrea Flynn, Ph.D., is a project scientist at the Centre for research attention paid to the measurement of social prob- Addiction and Mental Health, Social and Epidemiological lems and issues such as divorce, crime, education, and social Research Department, London, Ontario, Canada. mobility. Although the social-indicators movement initially focused on issues at the national level, recognition of consid- Samantha Wells, Ph.D., is a scientist at the Centre for erable regional and local variation in the prevalence and Addiction and Mental Health, Social and Epidemiological causes of social problems led to increased interest in mea- Research Department, London, Ontario, Canada; adjunct surement at the local level and, as such, the development assistant professor at the Western University, Department of “community indicators.” of Epidemiology and Biostatistics, London, Ontario, Community indicators that assess alcohol use and related Canada; and assistant professor at the University of harm are of great interest to community stakeholders and Toronto, Dalla Lana School of Public Health, Toronto, researchers. Alcohol use has been identified as a major risk Ontario, Canada. factor for acute and chronic health harms and imparts eco-

Community indicators: Assessing the impact of Alcohol 135 of community indicator data and highlighting the various theoretical premise that reducing alcohol use and alcohol- challenges of collecting data on alcohol at the community related problems requires a focus on the community system level. The literature on community indicators of alcohol use and structural factors influencing alcohol use rather than on and harms is expansive, spanning a large number of disciplines individual-level treatment and prevention (Holder 1998b; and extending back for numerous decades. As such, it is Holder et al. 2005; Treno and Lee 2002). beyond the scope of this article to provide a comprehensive Putting these definitions of community into practice review of all the literature and measures pertaining to com- when attempting to define and use community indicators is munity indicators on alcohol. Rather, this article provides not without its challenges and has direct implications for background information relevant to the use of community data collection. When defining the boundaries of the com- indicators in general and in relation to alcohol use and munity for the purpose of generating community indicators, harms, providing examples of some of the most common it is necessary to consider data availability, methodological measures used by alcohol researchers. In addition, the article requirements of research (i.e., having sufficient cases for mentions notable methodological and technological advances meaningful analyses), the catchment area in terms of service that have characterized this field of study over the past few provision, other geographic boundaries according to which decades, while highlighting the ongoing challenges faced by data are routinely collected by a community, and local stake- researchers and community stakeholders interested in assessing holder perspectives on their understanding of community alcohol use and alcohol-related harm at the local level. This (Gruenewald et al. 1997). These considerations do not article draws on extensive knowledge regarding community always coincide (e.g., available data may not match the indicator data on alcohol use and harms that has emerged catchment area of interest to community stakeholders), making from key community-based intervention trials, such as the it necessary to weigh the relative importance of these factors Saving Lives project led by Hingson (Hingson et al. 1996), when defining the boundaries of the community under the Community Trials project led by Holder (Grube 1997; study (Gruenewald et al. 1997). Holder 2000; Holder and Reynolds 1997; Holder and Treno 1997; Holder et al. 1997a, 1997b, 2000; Millar and Gruenewald 1997; Reynolds et al. 1997; Saltz and Stanghetta Data Sources for Community indicators on Alcohol 1997; Treno and Holder 1997; Voas 1997; Voas et al. 1997), Community indicators relating to alcohol use and harms are and the Communities Mobilizing for Change on Alcohol typically gleaned from two main types of data sources: (1) (CMCA) project led by Wagenaar (Wagenaar et al. 1994, archival sources collected for purposes other than addressing 1999, 2000 , 2000 ). The sections that follow outline some a b research questions on the impact of alcohol on communities of the main community indicators emerging from this literature (e.g., data from police and hospital records; crash data from and other relevant research in reference to four main topics— traffic safety databases); and (2) primary data collected by alcohol use, patterns, and problems; alcohol availability; researchers for the purpose of assessing, understanding, and alcohol-related health outcomes/trauma; and alcohol-related addressing alcohol use and related harms. These different crime and enforcement. sources of data have inherent advantages and disadvantages in terms of their utility for assessing the community-level What is A Community? impact of alcohol use. A number of different definitions of community have been Archival data proposed and used in the social sciences since the 1800s (for Archival data are an important source of community indicator a helpful overview of the various ways in which community data. Examples of these archival data sources include admin- has been defined historically, see Holder 1992). Generally istrative and surveillance databases maintained by local city speaking, the concept of community implies both geographic departments, community organizations, municipal/national and social proximity. Gruenewald and colleagues (1997) agencies, schools, hospitals, and police/law enforcement define a community as “a contiguous geopolitical area over- departments, in addition to larger health data–recording seen by a common political structure with common policing systems and traffic crash databases (e.g., the Healthcare and enforcement agencies and common educational and Cost and Utilization Project [HCUP] databases and the utility systems, and in which individuals are in daily physical Fatality Analysis Reporting System [FARS]). A wide range contact for the purposes of economic and social exchange” of indicators produced from archival data are used to assess (pp. 10–11). Holder (1992, 1998b) provides a similar defi- various alcohol-related issues and harms at the community nition based on a community-systems perspective and theo- level (examples and discussion of common indicators are pre- retically geared toward the prevention of alcohol problems. sented in the section Community Indicators on Alcohol and Community, in this context, is conceptualized as a dynamic, Alcohol-Related Harm; see also the table). complex, and adaptive system consisting of “a set or sets of A main benefit of using archival sources to produce persons engaged in shared socio-cultural-politico-economic community indicators is that they can be a cost-effective processes” (Holder 1998b, p. 12). This definition informs the means of documenting alcohol use and harms, offering a

136 Alcohol Research: Current Reviews large volume of retrospective data. In addition, unlike many Moreover, in the case of health-related indicators, the prob- of the constructs and measures used in social and epidemio- lem of low incidence is compounded by the fact that most logical research, archival data often result in indicators that health-related harms associated with alcohol use are only are straightforward, understandable, and of interest to the partially attributable to alcohol (Rehm et al. 2003). Although community, making them easier to use in community plan- researchers have developed approaches for estimating the ning (Gabriel 1997; Gruenewald et al. 1997; Mansfield and proportion of a given outcome that is attributable to alcohol Wilson 2008). Despite these advantages, there also are several as a specific risk factor (i.e., the attributable fraction, AF) limitations associated with using archival data to assess alcohol (see English et al. 1995; Martin et al. 2010; Rehm et al. 2003; use/harms in a community. By definition, these data are not Single et al. 1999; Stockwell et al. 2000; World Health gathered for research purposes and thus raise concerns relating Organization [WHO] 2000), these types of analyses require to both reliability and validity. Most notably, archival data a large volume of data and are typically only conducted at are subject to various sources of measurement error conse- higher levels of aggregation (e.g., State, Federal). quent to the fact that they are not collected according to the systematic and rigorous procedures that characterize social Primary data and epidemiological research. In addition, for some measures, Given that archival data often are unavailable or insufficient the involvement of alcohol may not be explicitly identified. to assess alcohol use and harm at the community level, primary For instance, hospital staff and police typically do not sys- data are collected to enhance knowledge of the community- tematically record data on alcohol consumption as part of level impact of alcohol use (Gruenewald et al. 1997; Stockwell routine practice (Brinkman et al. 2001; Gruenewald et al. et al. 2000). Population or subpopulation surveys are the 1997; Stockwell et al. 2000). When alcohol data are recorded predominant source of primary data used to produce alcohol- in community settings, they may be collected in an inconsis- related community indicators. Surveys offer the advantage of tent manner, influenced by subjective judgments and local allowing researchers to define the constructs of interest and practices (Brinkman et al. 2001). These limitations affect use psychometrically sound measures, including measures the extent to which researchers can confidently use existing that have been used in other community-level, State, or Federal data such as hospital records or police data to assess alcohol surveys, thereby facilitating comparisons. Surveys also permit involvement in injury or crime. Moreover, access to such the collection of self-report data that cannot be gleaned from data requires cooperation of local community agencies and/or archival data, such as individual-level alcohol use patterns; municipal or regional departments, which may not be underage access to alcohol; and beliefs, attitudes, and percep- always possible. tions surrounding alcohol. These data allow for individual Another important caveat relates to the use of archival and group-level risk factors to be determined and permit data for conducting community comparisons. Differences analyses on subpopulations of interest, such as adolescents or across communities in policies and data recording systems young adults (Gruenewald et al. 1997; Stockwell et al. 2000). (Gruenewald et al. 1997; Brinkman et al. 2001; Stockwell et In some instances, it may be possible to extract community- al. 2000) can make it difficult to conduct comparisons across level data from surveys conducted at higher levels of aggrega- communities. For example, when using arrest data on alcohol- tion (e.g., State or national surveys). However, the time related crime such as public intoxication or disorderly conduct, frames of State and national surveys often do not meet the indicator will reflect the definition used by the police community or research needs. For example, timing of data department (itself dependent on local or regional statutes) as collection is an essential factor when monitoring the impact well as on local enforcement capacity and practices, including of local policy changes or community initiatives, which may levels of police discretion. Thus, data on arrests may not be not coincide with national survey data collection (Mansfield directly comparable across communities, even if the commu- and Wilson 2008). Moreover, when attempting to glean nities themselves are well matched on demographic or other information from national or State-level surveys, sample sizes important baseline measures (Gruenewald et al. 1997). for smaller communities often are insufficient to permit Changes in recording systems or policies also present problems valid conclusions about specific communities or population for researchers interested in examining patterns over time subgroups within a community (Gruenewald et al. 1997; within communities. For example, variation over time in Mansfield and Wilson 2008; Stockwell et al. 2000). For the number of alcohol-related arrests may reflect changes in these reasons, surveys implemented at the community level enforcement, recording practices, or policies rather than true are key to developing local indicators of alcohol use and variations in alcohol-related crime (Gruenewald et al. 1997). harms. Surveys have been widely used in community-based Events with low levels of incidence present another chal- research projects, including both general population surveys lenge relating to use of archival data for assessing the impact and surveys of particular population groups, such as college of alcohol on communities. For instance, although alcohol- students (discussed below in Community Indicators on related morbidity and mortality are of great interest to Alcohol and Alcohol-Related Harm; see also the table). communities, these types of indicators may be difficult When conducting surveys to produce community indi- to provide at the community level, particularly for smaller cators, it is necessary to consider the limitations of the survey communities, because of their relatively low baseline rate. method. Recent evidence suggests that population surveys

Community indicators: Assessing the impact of Alcohol 137 table Examples, strengths and limitations of Community indicators from archival and Primary Data sources indicator indicators from Archival Sources indicators from P rimary Data Sources Category Examples of indicators Strengths Limitations Examples of indicators Strengths Limitations

Alcohol use, Per capita generated from Does not capture self-reported drinking offer individual- and general limitations patterns and alcohol available sales patterns of access behavior and group-level data of surveys and self- problems consumption data or use problems (youth, unavailable from report measures adults) archival sources that - high cost of Excludes “surrogate” - age at fi rst use can be aggregated surveys alcohols (homemade, - drinking prevalence to community level, - possible biases illegal, alcohol not - drinking volume including drinking (selection bias, intended for - heavy episodic pattern social desirability consumption) drinking (i.e., bias, recall bias, binge drinking) ability to implement coverage bias) Data may not be - hazardous or scientifi cally valid available at the local harmful drinking and reliable measures level (depends on employed in other catchment area of alcohol dependence communities and research and defi nition other levels of of “community”) aggregation (state, Federal) for comparison purposes

Alcohol Formal access Data on outlet Data do not capture alcohol purchase Capture events not Persuasiveness of availability - number of active licenses are sales to minors attempts at alcohol visible in archival results potentially outlet licenses generally outlets by data and not affected undermined by the per 100,000 maintained with Data do not capture pseudo-underage by self-report biases fact that buyers population good geographic social access customers are actually of - concentration/ specificity by useful in legal age spatial distribu- alcohol Control Data do not capture evaluations of tion of outlets boards differences between strategies to reduce - excise taxes outlets with respect to youth access to on alcoholic sales (e.g., small outlets alcohol beverages versus large outlets) - price of alcoholic beverages Community estimates may be affected by migratory patterns and purchases in communities of non-residence

Price data difficult to obtain

self-report data Provides data general limitations collected from under- unavailable from of self-report data age youth on ability archival sources to purchase alcohol at alcohol outlets

social access self-report data from Data on a high-risk general limitations underage youth on group unavailable of self-report data social sources of from archival data and surveys alcohol (friends, family sources members, bought by additional concerns someone else, took with coverage bias from someone for telephone else’s home) surveys due to high rates of cell phone use among youth and young adults

138 Alcohol Research: Current Reviews table Examples, strengths and limitations of Community indicators from archival and Primary Data sources (continued) indicator indicators from Archival Sources indicators from Primary Data Sources Category Examples of indicators Strengths Limitations Examples of indicators Strengths Limitations

Alcohol-related hospital discharge Capture serious low base rates of self-reported health general strengths general limitations health and data health/trauma mortality from alcohol harms and trauma of surveys and of self-report data trauma - rates of direct outcomes – at the community level experiences related self-report data and surveys alcohol mortality strong impact to alcohol or morbidity: for communities multiple causes of baC data provides Difficulty obtaining alcohol death often poorly ED surveys objective measure- permission for ED cardiomyopathy, nighttime recorded in archival - baC measurement ment of alcohol surveys alcohol cirrhosis emergency data - self reported alcohol involvement in of liver, alcoholic department (ED) consumption prior to injury presentations psychoses, presentations Proportion of mortali- ED presentation to ED accidental and nighttime ty/morbidity events ethyl alcohol single vehicle attributable to alcohol self-reported alcohol poisoning, etc. traffic crashes difficult to estimate at consumption shown - rates of indirectly- are reliable the community level to be valid measure related alcohol surrogates of of alcohol use deaths: certain alcohol-involved hospital/ED cases capture malignant tumors, trauma only the most severe cirrhosis, pancre- cases atitis, etc. blood alcohol concen- nighttime presen- trations (baC) not tations of trauma routinely recorded in from violence or hospital/emergency traffic accidents settings (surrogate measures) baC not always alcohol-involved measured in injury- traffic crashes producing/fatal crashes

single-vehicle Fatal crashes rare at nighttime traffic community level crashes

Alcohol-related Calls to police for if cooperation heavily dependent on self-reported crime self-reported crime general limitations crime nighttime assaults can be obtained, police enforcement and - alcohol consumption captures incidents of self-report data arrest or Ems accuracy in recording prior to driving/driving not reported to Challenges of Calls to emergency records are a while intoxicated police implementing medical services cost-effective Difficult to determine if - violence perpetration roadside surveys for alcohol-related source of data changes are due to after drinking baC provides an - can be difficult injury that is meaningful changes in police Roadside survey data objective measure to obtain police to community enforcement, valid - baC readings of alcohol cooperation Calls to police for members changes in crime, or consumption - high cost public drunkenness prevention programs - generally not or disorderly contact random (not in community prevention representative arrest rates for driving trials or when communities of community) under the influence are interested in com- - can be difficult to parisons, different find appropriate arrest rates for statutes or operational comparison nighttime assaults policies affect ability to communities compare communities alcohol-related arrests arrests represent only a as a percentage of proportion of offenses – total arrests underestimates harm

Community indicators: Assessing the impact of Alcohol 139 can underestimate the prevalence of alcohol use and associated Community indicators on Alcohol and Alcohol- harms because of selection bias, response bias, and coverage Related Harm bias (e.g., exclusion of homeless people) (Shield and Rehm 2012; see also Curtin et al. 2005; Dillman et al. 2002; Kempf Table 1 provides a summary of common community indicators and Remington 2007). The growth in use of voicemail, of alcohol use and related harms measured in community- caller ID, cell phones, and do-not-call lists, along with a based research. These indicators are organized into four growing aversion to aggressive telemarketing (Galesic et al. broad areas: alcohol use, patterns, and problems; alcohol 2006), have contributed to a notable decline in telephone availability; alcohol-related health outcomes/trauma; and survey response rates (Dillman et al. 2002; Hartge 1999; alcohol-related crime/enforcement. Although this table does Kempf and Remington 2007; see also Galea and Tracy not provide an exhaustive list of all possible measures used to 2007). Young people may be particularly underrepresented assess alcohol use and alcohol-related harm at the commu- in population surveys, given their high reliance on cell phones nity level, it provides common measures used in community and nonuse of landlines (Blumberg et al. 2007). Large-scale research (see Saltz et al. 1992). For each category, examples surveys can also be expensive and time consuming to implement. of indicators produced using archival and primary data When collecting primary data on alcohol use and harms, sources are provided, and general strengths and limitations it is also important to consider the limitations of self-report associated with these data are noted. data on drinking behavior and harms associated with drinking. Although self-report data on alcohol use generally are believed Alcohol use, Patterns, and Problems to be adequately valid and reliable and are widely used in social At the community level, indicators of alcohol use, patterns, and epidemiological research, they have been found to be and problems commonly are produced from individual-level susceptible to recall error as well as intentional distortion related self-report (i.e., survey) data. Existing community-based in part to social desirability (Del Boca and Darkes 2003). studies have examined a wide range of self-report measures Despite these limitations, surveys are key to answering of alcohol use, including, for example, lifetime drinking, specific questions about alcohol use and harms in the absence drinking frequency, heavy episodic drinking (or binge drink- of suitable archival data and are central for cross-validating ing) and hazardous or harmful drinking, alcohol problems, data gleaned from other sources. Moreover, extensive work and alcohol dependence (see Dent et al. 2005; Flewelling on conducting surveys as part of community prevention trials et al. 2005; Harrison et al. 2000; Hawkins et al. 2009; Perry has led to important methodological and statistical innova- et al. 1996, 2000, 2002; Saltz et al. 2009, 2010; Spera et al. tions, producing advanced knowledge of how to design and 2010; Wagenaar et al. 2006; see table 1). It is beyond the analyze surveys better (see Murray 1998; Murray and Short scope of this article to discuss the many different instruments 1995, 1996; Murray et al. 2004). used and all of the methodological challenges associated with In addition to surveys, other forms of primary data used measuring self-reported drinking and problems. Choice in to produce community indicators include pseudo-patron how to measure indicators of use, patterns, and problems studies designed to assess sales of alcohol to individuals appear- will depend on the research question being asked and the ing underage in both off-premise and on-premise alcohol population under examination. The strengths and limitations of various specific measures of alcohol consumption have outlets (see, for example, Freisthler et al. 2003; Saltz and been discussed extensively in the literature (see Dawson 2003; Stanghetta 1997; Toomey et al. 2008; Treno et al. 2006; Gmel et al. 2006a; Graham et al. 2004; Greenfield 2000; Wagenaar et al. 2000a) and roadside breath testing to assess Rehm 1998; Rehm et al. 1999), and recommendations for drinking and driving (e.g., McCartt et al. 2009; Roeper and measurement have been put forward elsewhere (see Dawson Voas 1998). These methods and their strengths and limitations and Room 2000). are discussed in later sections on alcohol availability and Drinking behavior among youth often is of particular crime/enforcement, respectively. interest to both researchers and communities. Evidence suggests Overall, although primary data, particularly surveys, that youth are more likely than adults to engage in risky allow for the use of psychometrically sound measures, they patterns of drinking (Adlaf et al. 2005) and to experience suffer from potential biases that researchers must take into harms from drinking, including harms to brain develop- account when assessing the impact of alcohol use on a com- ment, physical health, financial well-being, and social life munity. Alternatively, archival data sources can provide use- (Adlaf et al. 2005; Kolbe et al. 1993; Toumbourou et al. ful data on alcohol’s effects on local communities but require 2007; White and Swartzwelder 2004). Moreover, drinking careful interpretation and application and do not always allow at a young age can become an ingrained pattern of behavior, researchers to answer questions of interest. Each data source with youth who engage in risky drinking being more likely thus offers unique strengths and limitations, such that trian- to exhibit problem drinking later in life (Jefferis et al. 2005). gulation of both types of data is a common approach taken For these reasons, measuring alcohol use and alcohol-related by alcohol researchers when assessing the impact of alcohol problems among youth often is prioritized in prevention and on communities. early-intervention initiatives designed to reduce harm from

140 Alcohol Research: Current Reviews alcohol at both the individual and community levels (see result of the low base rate of harms at the community level DeJong et al. 2009; Nelson et al. 2010). The well-known and in part from challenges associated with obtaining sales prevention initiative CMCA (Wagenaar et al. 1994, 1999, data at the community level compared with the State level. 2000a, b) is notable for its focus on community-level strate- gies for reducing alcohol use and problems among youth Availability and its development of indicators of alcohol use and harms Measuring the availability of alcohol at the community level to evaluate program effectiveness. is essential for assessing the impact of policies designed to Surveys on youth drinking have commonly captured reduce alcohol use and alcohol-related harms (see Babor et these populations in their educational environments, including al. 2003). Availability commonly is measured in terms of elementary, high school, and college or university settings. commercial access (including alcohol outlet density, days The priority of addressing alcohol use among college students and hours of sales, and price of alcohol) as well as social is well evidenced by the NIAAA’s Rapid Response to College access (i.e., informal sources of alcohol, such as peers). Drinking Problems initiative, which produced recommenda- With respect to commercial access, although the evidence tions for reducing heavy drinking by this subgroup (see DeJong on the effects of limiting alcohol outlet density on alcohol et al. 2009; Nelson et al. 2010). Alcohol use, patterns, and consumption is somewhat mixed (see Livingston et al. 2007), problems have been measured in the implementation and studies generally have found significant positive relationships evaluation of alcohol prevention trials in school and college between alcohol outlet density and a range of problems at settings (see reviews by Saltz 2011 for college-based preven- the community level, including rates of violence, drinking tion approaches and Stigler et al. 2011 for elementary and and driving, motor vehicle accidents, medical harms, and high school programs). Examples of measures of alcohol use crime (Britt et al. 2005; Campbell et al. 2009; Gruenewald and problems among college and school-age students include and Remer 2006; Gruenewald et al. 2006; Livingston et al. self-reported alcohol use (i.e., measures of frequency of 2007; Toomey et al. 2012). Evidence also suggests a positive drinking, drinking patterns, and binge drinking) (Flewelling relationship between days (Middleton et al. 2010) and hours et al. 2005; Harrison 2000; Hawkins et al. 2009; Perry et al. (Hahn et al. 2010) of sale and alcohol consumption and 1996, 2000, 2002; Saltz et al. 2009, 2010), the incidence alcohol-related harms (see also Edwards et al. 1994). Alcohol and likelihood of intoxication at off-campus drinking estab- prices and taxes are inversely related to alcohol consumption lishments (Saltz et al. 2010), age of onset of drinking (Hawkins and heavy drinking (Chaloupka et al. 2002; Edwards et al. et al. 2009), and perceptions and experiences of negative 1994; Osterberg 2004; Wagenaar et al. 2009), although the consequences associated with drinking (Flewelling et al. 2005; extent of the impact of price changes depends to some Saltz et al. 2009, 2010). Significantly, although surveys of extent on cultural context (i.e., drinking norms) and prevail- college and university students may provide communities ing social and economic circumstances, among other factors with estimates of alcohol use, patterns, and problems among (Osterberg 2004; see also Babor et al. 2003). Researchers this segment of the population, these surveys are inherently have used indicators of commercial access to evaluate whether limited to the sampling frame of youth attending these insti- changes in State policies have an impact on alcohol use/ tutions. As a result, they fail to capture youth from the broader problems in communities (see Babor et al. 2003; Edwards et community not attending educational institutions and thus al. 1994; Hahn et al. 2010; Middleton et al. 2010). cannot offer community prevalence data for that age range. Community indicators of economic availability commonly With respect to archival data on alcohol use, this type are produced using archival data sources, including alcohol of information is less commonly available at the community price and tax (excise and sales) data from State departments level compared with higher levels of aggregation. Most and alcohol-control boards, although the quality of these notable in this regard is the use of sales data to examine per data and their utility for research at the community level capita alcohol consumption. WHO (2000) has recom- varies substantially across States (Gruenewald et al. 1997). mended that alcohol use among populations be monitored Archival data on retail alcohol prices are difficult to obtain using reliable estimates of per capita alcohol consumption at the State level, and even more so at the community level. derived from alcohol sales data, in addition to monitoring Evidence suggests that available data are prone to substantial through population surveys of alcohol use. Sales data commonly measurement error (Young and Bielinska-Kwapisz 2003), have been used at the State, regional, and Federal levels to leading many researchers to rely on tax data instead. When examine the link between per capita alcohol consumption making comparisons across communities or over time, and various health harms, including suicide (Kerr et al. researchers generally also prefer to use tax rates over price data 2011b, Landberg 2009), mortality and morbidity (Kerr et al. to avoid conflating price differences with differing tax rates 2011a; Nordstrom and Ramstedt 2005; Polednak 2012), across space and over time. Liquor licensing information and traffic crashes (Gruenewald and Ponicki 1995). These from alcohol-control boards commonly is used to generate types of analyses, however, generally are restricted to large indicators of commercial availability—namely, number of populations (Dawson 2003) and thus are less applicable to outlets/population rates and concentration of on- and off- alcohol researchers interested in community indicators (i.e., premise outlets (Sherman et al. 1996; see also Gruenewald et measures below the State level of aggregation), in part as a al. 1997). However, counts of active licenses represent only

Community indicators: Assessing the impact of Alcohol 141 an indirect measure of alcohol availability and can underestimate data, used to produce indicators of hospitalizations and alcohol sales (Gruenewald et al. 1992). Geographic Information emergency department (ED) visits associated with acute or System (GIS) mapping has emerged as an innovative means chronic alcohol use; (2) traffic fatality data, used to estimate of generating community indicators of outlet density (includ- alcohol involvement in crashes; and (3) household or ing off- and on-premise outlets) and to examine alcohol outlet subpopulation surveys, used to generate indicators from density and locations in relation to alcohol-related problems, self-reported data on alcohol-involved injuries (including such as assaults and sale of alcohol to minors (see Gruenewald violence). As shown in table 1, each of these data sources has et al. 2002; Millar and Gruenewald et al. 1997). strengths and limitations pertaining to their utility for pro- One major caveat relating to measures of commercial ducing community indicators on alcohol-related harms. access to alcohol is that archival data obscure who is making purchases, who is consuming the alcohol purchased, and Hospital and Ed data. Archival hospital data allow for how (in what patterns) the alcohol is being consumed. documentation of cases of alcohol-related health outcomes Therefore, important information about risky drinking behavior and trauma requiring urgent or emergent care. Such data (i.e., binge drinking) and populations who engage in such can provide powerful information for use by communities behavior remains unknown from data on alcohol availability. (e.g., in educational or prevention campaigns) because of This limitation is particularly salient for measuring drinking their severity and corresponding psychological impact among youth, who commonly obtain alcohol from social (Stockwell et al. 2000). Despite this appeal, notable challenges rather than commercial sources (see Wagenaar et al. 1993). exist to using archival data to produce community In light of this limitation, and the fact that early preven- indicators on health outcomes and trauma associated with tion of alcohol use and alcohol-related problems is often a alcohol. First, as stated above, one of the major caveats high priority for communities and researchers, other data with measuring alcohol-related mortality and morbidity at collection strategies have been implemented to measure access the community level is the rarity of cases (Giesbrecht et al. to alcohol among youth. Access surveys involving pseudo- 1989; Stockwell et al. 2000), meaning that there may be underage youth purchase attempts have produced indicators insufficient numbers for meaningful analysis at the community of youth commercial access, often as part of the evaluation level. Second, it often is quite difficult to obtain access of community prevention initiatives (see Chen et al. 2010; to hospital or ED data within communities, particularly Grube 1997; McCartt et al. 2009; Paschall et al. 2007; Perry data of reasonable quality for developing valid and reliable et al. 1996, 2000, 2002; Toomey et al. 2008; Wagenaar et al. estimates. Third, it often is challenging or impossible to 1994, 1999, 2000a, b). Self-reported social access to alcohol determine the extent of alcohol involvement in health has also been measured in school or community surveys of outcomes. As previously noted, many chronic health youth, with participants asked to report on sources from harms associated with alcohol, including those leading to which they obtain alcohol (i.e., commercial [on- or off-premise hospitalization and mortality, are only partially attributable outlets] versus social [friends, family, etc.] sources) (see Dent to this risk factor (Rehm et al. 2003). In terms of emergency et al. 2005; Harrison et al. 2000; Hearst et al. 2007; Jones- cases, archival data frequently do not capture alcohol Webb et al. 1997; Wagenaar et al. 1994). Some studies also involvement (Giesbrecht et al. 1989; Stockwell et al. 2000). have examined perceived availability of alcohol among youth Blood alcohol concentration (BAC) is not routinely (Flewelling et al. 2005; Perry et al. 1996, 2000, 2002; Treno assessed in hospitals or urgent-care centers in relation to et al. 2008). traumatic presentations, given that staff generally are operating under time and resource constraints that preclude Health outcomes/trauma systematic testing for alcohol use. Staff also may be As stated previously, evidence reveals a strong and consistent hesitant to make conclusions about intoxication because association between alcohol consumption and a variety of of insurance and liability concerns (Giesbrecht et al. 1989, negative health outcomes, including morbidity, early mortality, 1997; Stockwell et al. 2000; Treno and Holder 1997). As a and increased risk of trauma such as burns, falls, drowning, result, archival data of emergency cases likely underestimate and injury from interpersonal violence (Cherpitel 1995; the role of alcohol in trauma requiring emergent care. In Gmel et al. 2006b; Rehm et al. 2003, 2006; Treno et al. 1997). cases where BAC is recorded, determining the role of Collectively, alcohol-related health harms and traumas impose alcohol in a traumatic event is complicated by time notable demands on local emergency and hospital services. elapsed since the incident and by alcohol consumed after Documenting alcohol-related morbidity, mortality, and the incident (Young et al. 2004). trauma is thus often a priority for communities and researchers, In the face of challenges associated with lack of docu- with such research informing initiatives geared toward pre- mentation of alcohol involvement in archival data, researchers venting alcohol-related harm and efforts to reduce health costs. commonly turn to surrogate measures of alcohol-related Both archival and primary data have been used to pro- trauma. Such measures have been well studied using interna- duce community indicators relating to fatal and nonfatal tional data. For instance, Young and colleagues (2004) found alcohol-involved health harms. Data sources and types of that being male, unmarried, younger than age 45, and pre- indicators emerging from these data include (1) hospital senting at EDs in the late night or early morning hours on

142 Alcohol Research: Current Reviews Fridays, Saturdays, or Sundays were most highly associated McCarroll and Haddon [1962]; Perrine et al. [1971]; with alcohol consumption prior to injury (based on BAC Zador [1991]; and Zador et al. [2000]). Relative risk data and self-reported alcohol consumption within 6 hours prior to such as these have been widely used to support alcohol injury). The strongest predictor of alcohol-related injury was safety legislation, including the lowering of BAC driving time of day of presentation (odds ratio of 4.92 for presenta- limits (see review by Mann et al. 2001). tions occurring between midnight and 4:59 a.m.). It follows The FARS (formerly the Fatal Accident Reporting that, in the absence of reliable BAC data, proxy measures System) (see http://www.nhtsa.gov/FARS), initially established that take into account time-of-day presentation and demo- in 1975, is a reliable database of all fatal crashes in the graphic variables may offer a means for estimating alcohol- United States and includes the BACs of drivers involved related trauma in a community (Brinkman et al. 2001; in fatal crashes. When chemical tests of driver BACs are Treno et al. 1996). Such estimates require access to medical not performed in fatal crashes, FARS provides imputed data records that include time-of-day presentation and detailed (see Subramanian 2002). FARS data can be disaggregated to demographic information. the level of the county (see Voas et al. 1998; Williams 2006). Archival data on hospitalizations and ED visits are becom- Studies using FARS or State traffic safety department ing more readily available for use in the development of databases have generated indicators of various levels of driver community indicators. For example, the Healthcare Cost BAC associated with traffic fatalities (e.g., Hingson et al. and Utilization Project (see Steiner et al. 2002, http://www. 2005, et al. 2006; Wagenaar and Wolfson 1995). However, hcup-us.ahrq.gov/) consists of a series of health care fatal crashes are relatively rare events (Voas et al. 1997), and databases that provide data on inpatient, ambulatory, and thus aggregation of events over a long time period may be ED cases for community hospitals in participating States needed to produce sufficient cases for analysis at the com- since 1988. These databases permit research on topics such munity level (e.g., see Wagenaar et al. 2000a). as diagnoses; procedures; mortality; cost of health services; Researchers commonly also use fatal single-vehicle access to health care programs; and treatment outcomes at nighttime crashes as a surrogate for alcohol-involved traffic the national, State, and local levels (http://www.hcup-us. fatalities, which can be a useful strategy when data on alcohol ahrq.gov/). Some participating States allow the release of involvement in crashes are unavailable for the community of hospital and patient-level geographic data that may permit interest or too few cases have been documented. These data analysis at the community level (Steiner et al. 2002). have been shown to be a reliable proxy for alcohol-related Researchers have also produced indicators on alcohol- fatalities. They often are available from local or State sources involved trauma at the community level from ED surveys, involving the collection of interview and breathalyzer data (e.g., police departments or departments of transportation) from ED patients (see Cherpitel 1994 and 1993 for reviews and, depending on the size of the community, may occur in of ED studies; see also Busset al. 1995; Cherpitel et al. 2009; sufficient numbers for analysis (see Hingson et al. 1996; Holder et al. 2000; Treno and Holder 1997). Cherpitel Roeper and Voas 1998; Treno et al. 2006; Wagenaar and (1995) measured alcohol-related problems and injuries or Holder 1991; Wagenaar et al. 2000a, 2006). Nevertheless, illnesses for which emergency medical care was sought in a caution is warranted when interpreting traffic crash data, countywide representative study of ED data. When compar- particularly in the absence of BAC data, given the myriad of ing these data to a general population sample, Cherpitel other factors that stand to be involved in crashes, including (1995) found no difference in frequency of drunkenness road conditions, speeding, and use of seat belts. The use of related to injury between the two samples, suggesting that multiple data sources for triangulation of data (Gruenewald ED surveys may be a useful approach for measuring these et al. 1997) can help overcome the limitations of any one issues. However, obtaining ED cooperation and producing measure of alcohol-involved vehicle crashes. representative ED samples is a notable challenge faced by researchers when endeavoring to conduct ED surveys Population Survey data. Population or community (Holder et al. 2000). surveys are used to measure self-reported alcohol-related health outcomes and trauma. An advantage of these traffic Fatality data. Alcohol-related traffic fatalities are surveys is that they can detect events not resulting in an important form of trauma in the community-indicator fatalities or hospital admissions (Gruenewald et al. 1997). literature on alcohol-related harm. Consistent evidence These data are thus useful for documenting less severe confirms that alcohol is a leading cause of traffic crashes, cases, which are more common than fatal or near-fatal particularly those resulting in fatal and nonfatal injuries cases. However, the number of self-reported events (e.g., (Hingson and Winter 2003). Research has demonstrated injury) may still be insufficient for analysis, particularly that the relative risk of fatal injury and fatal crash involvement in small communities. General limitations of population rises with increasing driver BAC (see the classic Grand surveys apply to these data, including the cost and time Rapids study by Borkenstein et al. [1974] and subsequent required to conduct them, as well as reporting and studies by Hurst [1973]; Krüger and Vollrath [2004]; coverage biases that may result in underestimates of Mathijssen and Houwing [2005]; Mayhew et al. [1986]; alcohol-related harms.

Community indicators: Assessing the impact of Alcohol 143 Crime/Enforcement BACs than those who consent to a breath test (Lestina et al. 1999). Conversely, overestimates of impaired driving may Both primary and archival data sources have been used to occur if roadways characterized by high volumes of alcohol- generate measures of alcohol-related crime in communities. related crashes are targeted for surveys (Lestina et al. 1999). At the community level, household, telephone, and school In evaluations of alcohol-safety programs (and other alcohol surveys have been conducted to measure various self-reported interventions), it is necessary to compare the intervention crimes, including driving under the influence (DUI) (e.g., Clapp et al. 2005; Saltz et al. 2009; Wagenaar et al. 2006), community with a comparison community in which the underage alcohol purchases (e.g., Harrison et al. 2000), program was not implemented to determine whether alcohol-related violence (Greenfield and Weisner 1995), and changes in drinking and driving can be attributed to the public drunkenness (Greenfield and Weisner 1995). The intervention. However, finding adequate comparison sites general strengths and limitations of surveys and self-report can be a challenge, given the need for a community with measures of alcohol use have been discussed previously. similar population characteristics and policies and the fact Therefore, this section will focus on roadside surveys and that comparison (“non-experimental”) communities may arrest data. have their own campaigns to reduce drinking and driving Roadside surveys involve stopping motorists at roadside (see Voas 1997). checkpoints for the purpose of collecting breath alcohol Arrest data on DUI as well as other alcohol-related measurements. Two key purposes of roadside surveys are to offenses also represent valuable indicators for communities. track drinking and driving trends and to evaluate alcohol Numerous researchers have used archival police and justice safety programs (Lange et al. 1999; Lestina et al. 1999). The records to produce community indicators of alcohol-related majority of roadside studies conducted to track trends in crimes, including DUI, liquor law violations, assault, public drinking and driving have occurred at the national level (e.g., drunkenness, and disorderly conduct (e.g., Breen et al. 2011; in the United States, Canada, Britain, Germany, Sweden, Duncan et al. 2002; Sherman et al. 1996; Treno et al. 2006; Norway, Belgium, and the Netherlands) (see Lacey et al. Wagenaar et al. 2000a) (see table 1). When using archival 2008; Lestina et al. 1999; Lund and Wolfe 1991; Voas et al. data to assess levels of alcohol-related crime, it is important 1998; Wolfe 1974 for information on the U.S. National to recognize that such arrests represent only offenses brought Roadside Surveys). These national surveys typically do not to the attention of the police that they have acted upon. provide sufficient data at the community level for assessment Some criminal events (e.g., violent crime) are not commonly of local drinking and driving because of the exclusion of reported to the police, or there may be insufficient cause for smaller communities and/or roadways with low daily traffic police to file an arrest report (Brinkman et al. 2001). counts (Voas et al. 1998). At the community level, roadside Moreover, by definition, arrest data are dependent on local surveys primarily have been used in the evaluation of com- and State statutes and also are highly sensitive to enforce- munity prevention trials (e.g., McCartt et al. 2009; Roeper ment capacity and practices as well as operational changes and Voas 1998). They allow researchers to assess changes and recording practices, including police discretion in drinking-and-driving behavior in relation to prevention (Gruenewald et al. 1997). These factors are thus critical to campaigns when fatality and crash data are unavailable (Roeper consider when making comparisons over time or across and Voas 1998). In instances where fatality and crash data communities. As noted previously, changes in alcohol-related are available, roadside survey data may still be useful to arrests can represent changes in actual crime, changes in confirm that changes in crash data reflect valid changes in enforcement or recording practices, or changes in policies drinking-and-driving behavior rather than other changes and laws (Gruenewald et al. 1997). In some instances, con- not related to alcohol consumption (e.g., roadway improve- founding variables (such as police discretion in making ments) (Roeper and Voas 1998). arrests) are difficult if not impossible to measure. Two main strategies are used to implement roadside sur- Another problem with police data is that for many types veys at the community level: (1) “piggybacking” on existing of crime (e.g., violence), police do not formally measure police sobriety check points; and (2) using roadside check alcohol involvement (i.e., through a breath test). Although points dedicated entirely to research. In both instances, some research has measured alcohol-involved crime through cooperation of local police is imperative, which may create archival records of cases that police have flagged for alcohol a challenge in communities lacking widespread support for involvement (Wagenaar et al. 2000a), these data are unlikely the research (Howard and Barofsky 1992). In addition to to be systematic and rely in large part on police discretion the notable cost associated with conducting roadside surveys, (see discussion by Brinkman et al. 2001). To partially address there are several limitations and challenges associated with such concerns, surrogate measures have been used to pro- this method of data collection (Lestina et al. 1999). For duce indicators of alcohol-related crime from archival data. example, many high-BAC drivers are able to avoid roadside For example, nighttime assaults have been used as a proxy survey check points by driving alternate routes, resulting in for alcohol-related violence, given that temporal data are underestimates of local levels of drinking and driving likely to be recorded in police records and violent assaults (Lestina et al. 1999). Drivers also may refuse to provide a during nighttime hours have a high likelihood of being breath sample, and these people may be likely to have higher alcohol related (Brinkman et al. 2001).

144 Alcohol Research: Current Reviews Indicators of enforcement are also related to measure- community indicator systems consisting of data on a range ment of alcohol-related crime at the community level. of factors (e.g., social, economic, and environmental) to Some investigators have measured enforcement activities in allow a detailed examination of influences on community community-based research projects, often for the purpose of well-being (Besleme and Mullin 1997; Ramos and Jones evaluating policy changes or prevention efforts (e.g., Grube 2005). National initiatives such as the 2008 Community 1997; McCartt et al. 2009; Voas, Holder and Gruenewald Health Status Indicators (CHSI) project (see Heitgerd et al. 1997; see also Wagenaar and Wolfson 1995) (see table 1). 2008; Metzler et al. 2008; see also www.communityhealth. Indicators of enforcement can provide communities with hhs.gov), the Community Assessment Initiative (http:// data on enforcement capacity and, if tracked over time, can www.cdc.gov/ai/index.html), and the National Neighborhood allow for an assessment of the impact of enforcement on Indicators Partnership (http://www.neighborhoodindicators. reducing alcohol-related crime. org), for example, have sought to improve access to local data and inform use of data in planning efforts and evalua- tion of health policies and interventions. At the international Conclusion level, the Community Indicators Consortium, established in 2003, represents one of the most extensive efforts to engage Measuring alcohol use and harm in communities is complex stakeholders from around the world and to document and and requires researchers to make choices and find creative share knowledge on community indicators (see Ramos and ways of assessing the local-level impact of alcohol. The data Jones 2005; http://www.communityindicators.net). Some source and indicator used will depend on data availability, projects included in the Community Indicators Consortium the purpose of the research (e.g., to provide a community database of indicator projects specifically include risky alcohol with descriptive data versus evaluation of an intervention), consumption as part of their examination of community and, in many cases, community support for the research to well-being (see http://www.communityindicators.net). These facilitate access to archival data or cooperation in primary types of initiatives suggest that community indicators, includ- data collection efforts. ing indicators of alcohol use and harm, will continue to grow Whether using archival or primary data to produce in the coming years as an area of interest and innovation. community indicators, it is important for both researchers Community indicators are certainly not a panacea for and community stakeholders to be aware of the strengths either investigators or community stakeholders. However, and potential limitations of the data. They must also recog- when produced with a thorough understanding of the local nize the value of combining data from multiple sources community system and through thoughtful application of when making conclusions about the impact of alcohol on advanced methodological knowledge, they can serve as a communities. Indeed, many community-based projects have powerful tool for understanding, assessing, and addressing relied on both primary and archival data to assess alcohol use alcohol-related problems within their local context. ■ and harms in communities and to evaluate the impact of intervention initiatives. Triangulation of indicators is key for validating measures and thus drawing accurate conclusions Acknowledgements about research findings. Despite the limitations and challenges associated with Support to CAMH for salary of scientists and infrastructure assessing alcohol use and alcohol-related harms at the com- is provided by the Ontario Ministry of Health and Long- munity level, many significant advances have been made in Term Care. The views expressed here do not necessarily the field, including important advances in statistical methods reflect those of the Ministry of Health and Long-Term Care. (e.g., Murray 1998; Murray and Short 1995, 1996; Murray et al. 2004), refinement of surrogate measures (e.g., Treno et al. 1994, 1996, 1997), and spatial analysis (e.g., Gruenewald Financial Disclosure et al. 2002; Millar and Gruenewald 1997). Another example of an innovative approach that currently is being employed The authors declare that they have no competing financial to develop community indicators involves use of a mobile interests. research laboratory to collect social, epidemiological, and biological data in diverse communities in the province of Ontario, Canada. Led by a multidisciplinary team of researchers, References this project involves collection of local data and the develop- aDlaF, E.m.; bEgin, P.; anD saWKa, E. Canadian Addiction Survey (CAS): A National Survey ment of a community indicator database relating to mental of Canadians’ Use of Alcohol and Other Drugs: Prevalence of Use and Related Harms: health and addictions in participating communities, includ- Detailed Report. ottawa, Canada: Canadian Centre on substance abuse, 2005. ing indicators of alcohol use and harms (see Wells et al. 2011). baboR, t.; CaEtano, R.; CassWEll, s.; Et al. Alcohol: No Ordinary Commodity. Research Building on these types of innovations and the rich history and Public Policy. new york: oxford university Press, 2003. of social indicators in the United States, a number of com- bEslEmE, K., anD mullin, m. Community indicators and healthy communities. National munities recently have sought to develop comprehensive Civic Review 86(1):43–52, 1997.

Community indicators: Assessing the impact of Alcohol 145 blumbERg, s.J., luKE, J.V., Cynamon, m.l., anD FRanKEl, m.R. Recent trends in household DunCan, s.C.; DunCan, t.E.; anD stRyCKER, l.a. a multilevel analysis of neighborhood con- telephone coverage in the united states. in: lepkowski, J.m., Et al, Eds. Advances in text and youth alcohol and drug problems. Prevention Science 3(2):125–133, 2002. Telephone Survey Methodology. hoboken, nJ: Wiley & sons, 2007, pp. 56–86. PmiD: 12088137

boRKEnstEin, R.F.; CRoWthER, R.F.; shumatE, R.P.; Et al. the role of the drinking driver in EDWaRDs, g.; anDERson, P.; baboR, t.; Et al. Alcohol Policy and the Public Good. new york: traffic accidents (the grand Rapids study). Blutalkohol11(suppl. 1):7–13, 1974. oxford university Press, 1994.

bREEn, C.; shaKEshaFt, a.; slaDE, t.; Et al. Do community characteristics predict alcohol- English, D.R.; holman, C.D.J.; milnE, E.; Et al. The Quantification of Drug Caused related crime? Alcohol and Alcoholism 46(4):464–470, 2011. PmiD: 21546376 Morbidity and Mortality in Australia 1995. Canberra, australia: Commonwealth Department of human services and health, 1995. bRinKman, s.; ChiKRitzhs, t.; stoCKWEll, t.; Et al. an indicator approach to the measure- ment of alcohol-related violence. in: Williams, P., Ed. Alcohol, Young Persons and FlEWElling, R.l.; austin, D.; halE, K.; Et al. implementing research-based substance abuse Violence. australian institute of Criminology Research and Public Policy series no. 35. prevention in communities: Effects of a coalition-based prevention initiative in Vermont. Canberra, australia: australian institute of Criminology, 2001, pp. 61–84. Journal of Community Psychology 33(3):333–335, 2005.

bRitt, h.R.; CaRlin, b.P.; toomEy, t.l.; Et al. neighborhood level spatial analysis of the FREisthlER, b.; gRuEnEWalD, P.J.; tREno, a.J.; anD lEE, J. Evaluating alcohol access and the relationship between alcohol outlet density and criminal violence. Environmental and alcohol environment in neighbourhood areas. Alcoholism: Clinical and Experimental Ecological Statistics 12:411–426, 2005. Research 27(3):477–484, 2003. PmiD: 12658114

buss, t.F.; abDu, R.; anD WalKER, J.R. alcohol, drugs, and urban violence in a small city trau- gabRiEl, R.m. Community indicators of substance abuse: Empowering coalition planning ma center. Journal of Substance Abuse Treatment 12(2):75–83, 1995. PmiD: 7623393 and evaluation. Evaluation and Program Planning 20:335–343, 1997.

CamPbEll, C.a.; hahn, R.a.; ElDER, R.; Et al. the effectiveness of limiting alcohol outlet galEa, s., anD tRaCy, m. Participation rates in epidemiologic studies. Annals of density as a means of reducing excessive alcohol consumption and alcohol-related Epidemiology 17(9):643–653, 2007. PmiD: 17553702 harms. American Journal of Preventive Medicine 37(6):556–569, 2009. PmiD: 19944925 galEsiC, m.; touRangEau, R.; anD CouPER, m.P. Complementing random-digit-dial tele- ChalouPKa, F.J.; gRossman, m.; anD saFFER, h. the effects of price on alcohol consumption phone surveys with other approaches to collecting sensitive data. American Journal of and alcohol-related problems. Alcohol Research & Health 26(1):22–34, 2002. PmiD: Preventive Medicine 31(5):437–443, 2006. PmiD: 17046416 12154648 giEsbRECht, n.; gonzalEz, R.; gRant, m.; Et al. Drinking and Casualties: Accidents, ChEn, m.-J.; gRubE, J.W.; anD gRuEnEWalD, P.J. Community alcohol outlet density and Poisonings and Violence in an International Perspective. london, uK: Routledge, 1989. underage drinking. Addiction 105(2):270–278, 2010. PmiD: 20078485 gmEl, g.; bissERy, a.; gammEtER, R.; Et al. alcohol-attributable injuries in admissions to a ChERPitEl, C.J. alcohol and injuries resulting from violence: a review of emergency room swiss emergency room: an analysis of the link between volume of drinking, drinking studies. Addiction 89(2):157–165, 1994. PmiD: 8173481 patterns, and preattendance drinking. Alcoholism: Clinical and Experimental Research ChERPitEl, C.J. alcohol and injuries: a review of international emergency room studies. 30(3):501–509, 2006a. PmiD: 16499491 Addiction 88(7):923–937, 1993. PmiD: 8358264 gmEl g.; gRaham K.; KuEnDig h.; anD KuntsChE, s. measuring alcohol consumption: should ChERPitEl, C.J. alcohol and casualties: Comparison of county-wide emergency room the ‘graduated frequency’ approach become the norm in survey research? Addiction data with the county general population. Addiction 90(3):343–350, 1995. PmiD: 101(1):16–30, 2006b. PmiD: 16393189 7735019 gRaham, K.; DEmERs, a.; REhm, J.; anD gmEl, g. Problems with the graduated frequency ChERPitEl, C.J.; boRgEs, g.; giEsbRECht, n.; Et al. (EDs.). Alcohol and Injuries: Emergency approach to measuring alcohol consumption: Results from a pilot study in toronto, Department Studies in an International Perspective. geneva, switzerland: World health Canada. Alcohol and Alcoholism 39(5):455–462, 2004. PmiD: 15289203 organization, 2009. gREEnFiElD, t.K. Ways of measuring drinking patterns and the difference they make: ClaPP, J.D.; Johnson, m.; Voas, R.b.; Et al. Reducing Dui among us college students: Experience with graduated frequencies. Journal of Substance Abuse 12(1–2):33–49, Results of an environmental prevention trial. Addiction 100(3):327–334, 2005. PmiD: 2000. PmiD: 11288473 15733246 gREEnFiElD, t.K., anD WEisnER, C. Drinking problems and self-reported criminal behavior, CuRtin, R.; PREssER, s.; anD singER, E. Changes in telephone survey nonresponse over the arrests and convictions: 1990 us alcohol and 1989 county surveys. Addiction past quarter century. Public Opinion Quarterly 69(1):87–98, 2005. 90(3):361–373, 1995. PmiD: 7735021

DaWson, D.a. methodological issues in measuring alcohol use. Alcohol Research & gRoss, b.m., anD stRaussman, J. the social indicators movement. Social Policy Health 27(1):18–29, 2003. PmiD: 15301397 5(september–october):43–54, 1974.

DaWson, D.a., anD Room, R. towards agreement on ways to measure and report drinking gRubE, J.W. Preventing sales of alcohol to minors: Results from a community trial. patterns and alcohol-related problems in adult general population surveys: the skarpö Addiction 92(suppl. 2):s251–s260, 1997. PmiD: 9231448 conference overview. Journal of Substance Abuse 12(1–2):1–21, 2000. PmiD: 11288465 gRuEnEWalD, P.J., anD PoniCKi, W.R. the relationship of the retail availability of alcohol and alcohol sales to alcohol-related traffic crashes. Accident Analysis and Prevention DEJong, W.; laRimER, m.E.; WooD, m.D.; anD haRtman, R. niaaa’s Rapid Response to 27(2):249–259, 1995. PmiD: 7786392 College Drinking Problems initiative: Reinforcing the use of evidence-based approaches in college alcohol prevention. Journal of Studies on Alcohol and Drugs gRuEnEWalD, P.J., anD REmER, l. Changes in outlet densities affect violence rates. Alcoholism: 16(suppl.):5–11, 2009. PmiD: 19538907 Clinical and Experimental Research 30(7):1184–1193, 2006. PmiD: 16792566

DEl boCa, F.K., anD DaRKEs, J. the validity of self-reports of alcohol consumption: state of gRuEnEWalD, P.J.; FREisthlER, b.; REmER, l.; Et al. Ecological models of alcohol outlets and the science and challenges for research. Addiction 98(suppl. 2):1–12, 2003. PmiD: violent assaults: Crime potentials and geospatial analysis. Addiction 101(5):666–677, 14984237 2006. PmiD: 16669900

DEnt, C.W.; gRubE, J.W.; anD biglan, a. Community-level alcohol availability and enforce- gRuEnEWalD, P.J.; maDDEn, P.; anD JanEs, K. alcohol availability and the formal power and ment of possession laws as predictors of youth drinking. Preventive Medicine resources of state alcohol beverage control agencies. Alcoholism: Clinical and 40(3):355–362, 2005. PmiD: 15533551 Experimental Research 16(3):591–597, 1992. PmiD: 1626661

Dillman, D.a.; EltingE, J.l.; gRoVEs, R.m.; Et al. survey nonresponse in design, data col- gRuEnEWalD, P.J.; REmER, l.; anD liPton, R. Evaluating the alcohol environment: lection, and analysis. in: groves, R.m.; Dillman, D.a.; Eltinge, J.s.; little, R.J.a., Eds. Community geography and alcohol problems. Alcohol Research & Health 26(1):42–48, Survey Nonresponse. new york: John Wiley, 2002, pp. 3–26. 2002. PmiD: 12154650

146 Alcohol Research: Current Reviews gRuEnEWalD, P.J.; tREno, a.J.; taFF, g.; Et al. Measuring Community Indicators: A Systems and howard, J., Eds. Community Prevention Trials for Alcohol Problems: Methodological Approach to Drug and Alcohol Problems. (applied social Research methods series Vol. Issues. Westport, Ct: Praeger, 1992, pp.209–226. 45.) thousand oaks, Ca: sage Publications, 1997. huRst, P.m. Epidemiological aspects of alcohol in driver crashes and citations. Journal hahn, R.a.; KuzaRa, J.l.; ElDER, R.W.; Et al. Effectiveness of policies restricting hours of of Safety Research 5(3):130–148, 1973. alcohol sales in preventing excessive alcohol consumption and related harms. JEFFERis, b.J.; PoWER, C.; anD manoR, o. adolescent drinking level and adult binge drink- American Journal of Preventive Medicine 39(6):590–604, 2010. PmiD: 21084080 ing in a national birth cohort. Addiction 100(4):543–549, 2005. PmiD: 15784069 haRRison, P.a.; FulKERson, J.a.; anD PaRK, E. the relative importance of social versus com- mercial sources in youth access to tobacco, alcohol, and other drugs. Preventive JonEs-WEbb, R.; toomEy, t.l.; shoRt, b.; Et al. Relationships among alcohol availability, Medicine 31(1):39–48, 2000. PmiD: 10896842 drinking location, alcohol consumption, and drinking problems in adolescents. Substance Use & Misuse 32(10):1261–1285, 1997. PmiD: 9286000 haRtgE, P. Raising response rates: getting to yes. Epidemiology 10(2):105–107, 1999. PmiD: 10069242 KEmPF, a.m., anD REmington, P.l. new challenges for telephone survey research in the twenty-first century. Annual Review of Public Health 28:113–126, 2007. PmiD: 17094769 haWKins, J.D.; oEstERlE, s.; bRoWn, E.C.; Et al. Results of type 2 translational research trial to prevent adolescent drug use and delinquency: a test of Communities that Care. KERR, W.C.; KaRRiKER-JaFFE, K.; subbaRaman, m; anD yE, y. Per capita alcohol consumption Archives of Pediatrics & Adolescent Medicine 163(9):789–798, 2009. PmiD: 19736331 and ischemic heart disease mortality in a panel of us states from 1950 to 2002. Addiction 106(2):313–322, 2011a. PmiD: 21059185 hEaRst, m.o.; FulKERson, J.a.; malDonalDo-molina, m.m.; Et al. Who needs liquor stores when parents will do? the importance of social sources of alcohol among young urban KERR, W.C.; subbaRaman, m.; anD yE, y. Per capita alcohol consumption and suicide mor- teens. Preventive Medicine 44(6):471–476, 2007. PmiD: 17428525 tality in a panel of us states from 1950 to 2002. Drug and Alcohol Review 30(5):473–480, 2011b. PmiD: 21896069 hEitgERD, J.l.; DEnt, a.l.; holt, J.b.; Et al. Community health status indicators: adding a geospatial component. Preventing Chronic Disease 5(3), 2008. PmiD: 18558046 KolbE, l.J.; Kann, l.; anD Collins, J.l. overview of the youth Risk behavior surveillance system. Public Health Reports 108(suppl. 1):2–10, 1993. PmiD: 8210269 hingson, R., anD WintER, m. Epidemiology and consequences of drinking and driving. Alcohol Research & Health 27(1):63–78, 2003. PmiD: 15301401 KRügER, h.P., anD VollRath, m. the alcohol-related accident risk in germany: Procedure, methods and results. Accident Analysis and Prevention 36(1):125–133, 2004. PmiD: hingson, R.; mCgoVERn, t.; hoWlanD, J.; Et al. Reducing alcohol-impaired driving in 14572834 massachusetts: the savings lives Program. American Journal of Public Health 86(6):791–797, 1996. PmiD: 8659651 laCEy, J.h.; KEllEy-baKER, t.; FuRR-holDEn, D.; Et al. 2007 National Roadside Survey of Alcohol and Drug Use by Drivers: Alcohol Results. Washington, D.C.: national highway hingson, R.W.; zaKoCs, R.C.; hEEREn, t.; Et al. Effects on alcohol related fatal crashes of a traffic safety administration, 2008. community based initiative to increase substance abuse treatment and reduce alcohol availability. Injury Prevention 11(2):84–90, 2005. PmiD: 15805436 lanD, K.C., anD sPilERman, s. Social Indicator Models. new york: sage, 1975. holDER, h.D. What is a community and what are implications for prevention trials for lanDbERg, J. Per capita alcohol consumption and suicide rates in the u.s., 1950–2002. reducing alcohol problems? in: holder, h.D., and howard. J.m., Eds. Community Suicide & Life-Threatening Behavior 39(4):452–459, 2009. PmiD: 19792986 Prevention Trials for Alcohol Problems. Westport, Ct: Praeger, 1992, pp. 15–33. langE, J.E.; lauER, E.m.; anD Voas, R.b. a survey of the san Diego-tijuana cross-border holDER, h.D. Planning for alcohol-problem prevention through complex systems model- binging: methods and analysis. Evaluation Review 23(4):378–398, 1999. PmiD: 10558392 ing: Results from simCom. Substance Use & Misuse 33(3):669–692, 1998a. PmiD: lEstina, D.; gREEnE, m.; Voas, R.b.; anD WElls, J. sampling procedures and survey 9533735 methodologies for the 1996 survey with comparisons to earlier national roadside sur- holDER, h.D. Alcohol and the Community: A Systems Approach to Prevention. Cambridge, veys. Evaluation Review 23(1):28–46, 1999. PmiD: 10346071 uK: Cambridge university Press, 1998b. liVingston, m.; ChiKRitzhs, t.; anD Room, R. Changing the density of alcohol outlets to holDER, h.D. Community prevention of alcohol problems. Addictive Behaviors 25(6):843– reduce alcohol-related problems. Drug and Alcohol Review 26(5):557–566, 2007. 859, 2000. PmiD: 11125775 PmiD: 17701520

holDER, h.D., anD REynolDs, R.i. application of local policy to prevent alcohol problems: lunD, a.K., anD WolFE, a.C. Changes in the incidence of alcohol-impaired driving in the Experiences from a community trial. Addiction 92(suppl. 2):s285–s292, 1997. PmiD: united states, 1973–1986. Journal of Studies on Alcohol 52(4): 293–301, 1991. 9231451 PmiD: 1875700

holDER, h.D., anD tREno, a.J. media advocacy in community prevention: news as a maCRaE, D. Policy Indicators: Links between Social Science and Public Debate. Chapel means to advance policy change. Addiction 92(suppl. 2):s189–s199, 1997. PmiD: hill, nC: university of north Carolina Press, 1985. 9231444 mann, R.E.; maCDonalD, s.; stoDuto, l.g.; Et al. the effects of introducing or lowering holDER, h.D.; gRuEnEWalD, P.J.; PoniCKi, W.R.; Et al. Effect of community-based interven- legal per se blood alcohol limits for driving: an international review. Accident Analysis tions on high-risk drinking and alcohol-related injuries. JAMA: Journal of the American and Prevention 33(5):569–583, 2001. PmiD: 11491238 Medical Association 284(18):2341–2347, 2000. PmiD: 11066184 mansFiElD, C.J., anD Wilson, J.l. Community-level data. North Carolina Medical Journal holDER, h.D.; saltz, R.F.; gRubE, J.W.; tREno, a.J.; Et al. summing up: lessons from a 69(2):142–145, 2008. PmiD: 18605166 comprehensive community prevention trial. Addiction 92(suppl. 2):s293–s301, 1997a. PmiD: 9231452 maRtin, J.; baRRy, J.; goggin, D.; Et al. alcohol-attributable mortality in ireland. Alcohol and Alcoholism 45(4):379–386, 2010. PmiD: 20530495 holDER, h.D.; saltz, R.F.; gRubE, J.W.; Voas, R.b.; Et al. a community prevention trial to reduce alcohol-involved accidental injury and death: overview. Addiction 92(suppl. 2): mathiJssEn, R., anD houWing, s. The Prevalence and Relative Risk of Drink and Drug s155–s171, 1997b. PmiD: 9231442 Driving in the Netherlands: A Case-control in the Tilburg Police District. (2nd ed.). leidschendam, the netherlands: sWoV institute for Road safety Research, 2005. holDER, h.D.; tREno, a.; anD lEVy, D. Community systems and ecologies of drug and alco- hol problems. in: stockwell, t.; gruenewald, P.J.; toumbourou, J.W., and loxley, W., mayhEW, D.R.; DonElson, a.C.; bEiRnEss, D.J.; anD simPson, h.m. youth, alcohol and rela- Eds. Preventing Harmful Substance Use: The Evidence Base for Policy and Practice. tive risk of crash involvement. Accident Analysis and Prevention 18(4):272–287, 1986. West sussex, uK: John Wiley and sons, 2005, pp. 149–162. PmiD: 3741579 hoWaRD, J., anD baRoFsKy, i. implementing research designs: Protecting the scientific mCCaRRoll, J.R., anD haDDon, W.J. a controlled study of fatal automobile accidents in integrity of community intervention studies: Confronting social realities. in: holder, h.D., new york City. Journal of Chronic Diseases 15(8):811–826, 1962.

Community indicators: Assessing the impact of Alcohol 147 mCCaRtt, a.t.; hEllinga, l.a.; anD WElls, J.K. Effects of a college community campaign on REhm, J.; mathERs, C.; PoPoVa, s.; Et al. global burden of disease and injury and econom- drinking and driving with a strong enforcement component. Traffic Injury Prevention ic cost attributable to alcohol use and alcohol-use disorders. Lancet 10(2):141–147, 2009. PmiD: 19333826 373(9682):2223–2233, 2009. PmiD: 19560604

mEtzlER, m.; KanaREK, n., highsmith, K., Et al. Community health status indicators Project: REhm, J.; Room, R.; gRaham, K.; Et al. the relationship of average volume of alcohol con- the development of a national approach to community health. Preventing Chronic sumption and patterns of drinking to burden of disease: an overview. Addiction Disease 5(3):a94, 2008. PmiD: 18558044 98(9):1209–1228, 2003. PmiD: 12930209

miDDlEton, J.C.; hahn, R.a.; KuzaRa, J.l.; Et al. Effectiveness of policies maintaining or REhm, J.; tayloR, b.; anD PatRa, J. Volume of alcohol consumption, patterns of drinking restricting days of alcohol sales on excessive alcohol consumption and related harms. and burden of disease in the European region 2002. Addiction 101(8):1086–1095, American Journal of Preventive Medicine 39(6):575–589, 2010. PmiD: 21084079 2006. PmiD: 16869838

millaR, a.b., anD gRuEnEWalD, P.J. use of spatial models for community program evalua- REynolDs, R.i.; holDER, h.D.; anD gRuEnEWalD, P.J. Community prevention and alcohol tion of changes in alcohol outlet distribution. Addiction 92(suppl. 2):s273–s283, retail access. Addiction 92(suppl. 2):s261–s272, 1997. PmiD: 9231449 1997. PmiD: 9231450 RoEPER, P., anD Voas, R.b. alcohol consumption measured at roadside surveys and varia- muRRay, D.m. Design and Analysis of Group-Randomized Trials. new york: John Wiley tions in traffic injury crashes. Accident Analysis and Prevention 30(4):409–416, 1998. and sons, 1998. PmiD: 9666237

muRRay, D.m., anD shoRt, b. intraclass correlation among measures related to alcohol Room, R.; baboR, t.; anD REhm, J. alcohol and public health. Lancet 365(9458):519–530, use by young adults: Estimates, correlates and applications intervention studies. 2005. PmiD: 15705462 Journal of Studies on Alcohol 56(6):681–694, 1995. PmiD: 8558900 saltz, R.F. Environmental approaches to prevention in college settings. Alcohol muRRay, D.m., anD shoRt, b. intraclass correlation among measures related to alcohol Research & Health 34(2):204–209, 2011. PmiD: 22330219 use by school aged adolescents: Estimates, correlates, and applications in intervention studies. Journal of Drug Education 26(3):207–230, 1996. PmiD: 8952207 saltz, R.F., anD stanghEtta, P. a community-wide Responsible beverage service program in three communities: Early findings. Addiction 92(suppl. 2):s237–s249, 1997. PmiD: muRRay, D.m.; VaRnEll, s.P.; anD blitstEin, J.l. Design and analysis of group-randomized 9231447 trials: a review of recent methodological developments. American Journal of Public Health 94(3):423–432, 2004. PmiD: 14998806 saltz, R.F.; gRuEnEWalD, P.J.; anD hEnnEssy, m. Candidate alcohol problems and implica- tions for measurement: general alcohol problems, outcome measures, instrumenta- nElson, t.F.; toomEy, t.l.; lEnK, K.m.; Et al. implementation of niaaa College Drinking tion, and surrogates. in: holder, h.D. and howard, J.m., Eds. Community Prevention task Force recommendations: how are colleges doing 6 years later? Alcoholism: Trials for Alcohol Problems. Westport, Ct: Praeger, 1992. Clinical and Experimental Research 34(10):1687–1693, 2010. PmiD: 20626728 saltz, R.F.; PasChall, m.J.; mCgaFFigan, R.P.; Et al. alcohol risk management in college noRDstRom, t., anD RamstEDt, m. mortality and population drinking: a review of the litera- settings: the safer California universities Randomized trial. American Journal of ture. Drug and Alcohol Review 24(6):537–547, 2005. PmiD: 16361210 Preventive Medicine 39(6):491–499, 2010. PmiD: 21084068

ostERbERg, E. Effects of price on taxation. in: heather, n. and stockwell, t., Eds. The saltz, R.F.; WElKER, l.R.; PasChall, m.J.; Et al. Evaluating a comprehensive campus: Essential Handbook of Treatment and Prevention of Alcohol Problems. West sussex, uK: Community prevention intervention to reduce alcohol-related problems in a college John Wiley and sons, 2004, pp. 199–212. population. Journal of Studies on Alcohol and Drugs (suppl. 16):21–27, 2009. PmiD:

PasChall, m.J.; gRubE, J.W.; blaCK, C.; Et al. alcohol outlet characteristics and alcohol 19538909 sales to youth: Results of alcohol purchase surveys in 45 oregon communities. shERman, R.E.; gillEsPiE, s.; anD Diaz, J.a. use of social indicators in assessment of local Prevention Science 8(2):153–159, 2007. PmiD: 17243019 community alcohol and other drug dependence treatment needs within Chicago.

PERRinE, m.W. Methodological Considerations in Conducting and Evaluating Roadside Substance Use & Misuse 31(6):691–728, 1996. PmiD: 8816117 Research Surveys. (tech. Rep. no. Doths–800–471.) Washington, DC: national shiElD, K.D., anD REhm, J. Difficulties with telephone-based surveys on alcohol consump- highway traffic safety administration, 1971. tion in high-income countries: the Canadian example. International Journal of Methods PERRy, C.l.; Williams, C.l.; KomRo, K.a.; Et al. Project northland high school interventions: in Psychiatric Research 21(1):17–28, 2012. PmiD: 22337654 Community action to reduce adolescent alcohol use. Health Education & Behavior singlE, E.; Robson, l.; REhm, J.; anD XiE, X. morbidity and mortality attributable to alcohol, 27(1):29–49, 2000. PmiD: 10709791 tobacco, and illicit drug use in Canada. American Journal of Public Health PERRy, C.l.; Williams, C.l.; KomRo, K. a.; Et al. Project northland: long-term outcomes of 89(3):385–390, 1999. PmiD: 10076491 community action to reduce adolescent alcohol use. Health Education Research sPERa, C.; FRanKlin, K.; uEKaWa, K.; Et al. Reducing drinking among junior enlisted air 17(1):117–132, 2002. PmiD: 11888042 Force members in five communities: Early findings of the EuDl program’s influence on PERRy, C.l.; Williams, C.l.; VEblEn-moRtEnson, s.; Et al. Project northland: outcomes of a self-reported drinking behaviors. Journal of Studies on Alcohol and Drugs community-wide alcohol use prevention program during early adolescence. American 71(3):373–383, 2010. PmiD: 20409431 Journal of Public Health 86(7):956–965, 1996. PmiD: 8669519 stEinER, C.; EliXhausER, a.; anD sChnaiER, J. the healthcare Cost and utilization Project: an PolEDnaK, a.P. u.s. mortality from liver cirrhosis and alcoholic liver disease in overview. Effective Clinical Practice 5(3):143–151, 2002. PmiD: 12088294 1999–2004: Regional and state variation in relation to per capita alcohol consumption. stiglER, m.h.; nEusEl, E.; anD PERRy, C.l. school-based programs to prevent and reduce Substance Use & Misuse 47(3):202–213, 2012. PmiD: 22217123 alcohol use among youth. Alcohol Research & Health 34(2):157–162, 2011. PmiD: Ramos, o.t., anD JonEs, K. Comprehensive community indicators systems. National Civic 22330213 Review 94(2):74–77, 2005. stoCKWEll, t.; ChiKRitzhs, t.; anD bRinKman, s. the role of social and health statistics in REhm, J. measuring quantity, frequency, and volume of drinking. Alcoholism: Clinical measuring harm from alcohol. Journal of Substance Abuse 12(1–2):139–154, 2000. and Experimental Research 22(suppl. 2):4s–14s, 1998. PmiD: 9603301 PmiD: 11288467

REhm, J.; gREEnFiElD, t.K.; Walsh, g.; Et al. assessment methods for alcohol consump- subRamanian, R. Transitioning to Multiple Imputation: A new method to Impute Missing tion, prevalence of high risk drinking and harm: a sensitivity analysis. International Blood Alcohol Concentration (BAC) Values in FARS. (Dot hs 809 403). Washington, Journal of Epidemiology 28(2):219–224, 1999. PmiD: 10342682 DC: national highway traffic safety administration, 2002.

148 Alcohol Research: Current Reviews toomEy, t.l.; ERiCKson, D.J.; CaRlin, b.P.; Et al. the association between density of alco- WagEnaaR, a.C.; ERiCKson, D.J.; haRWooD, E.m.; anD o’mallEy, P.m. Effects of state coali- hol establishments and violent crime within urban neighbourhoods. Alcoholism: Clinical tions to reduce underage drinking. a national evaluation. American Journal of and Experimental Research 36(8):1468–1473, 2012. PmiD: 22587231 Preventive Medicine 31(4):307–315, 2006. PmiD: 16979455

toomEy, t.l.; ERiCKson, D.J.; lEnK, K.m.; Et al. a randomized trial to evaluate a manage- WagEnaaR, a.C.; FinnEgan, J.R.; WolFson, m.; Et al. Where and how adolescents obtain ment training program to prevent illegal alcohol sales. Addiction 103(3):405–413, alcoholic beverages. Public Health Reports 108(4):459–464, 1993. PmiD: 8341780 2008. PmiD: 18190669 WagEnaaR, a.C.; gEhan, J.P.; JonEs-WEbb, R.; Et al. Communities mobilizing for Change on toumbouRou, J.W.; stoCKWEll, t.; nEighboRs, C.; Et al. interventions to reduce harm associ- alcohol: lessons and results from a 15-community randomized trial. Journal of ated with adolescent substance use. Lancet 369(9570):1391–1401, 2007. PmiD: 17448826 Community Psychology 27(3):315–326, 1999. tREno, a.J., anD holDER, h.D. measurement of alcohol-involved injury in community pre- WagEnaaR, a.C.; muRRay, D.m.; gEhan, J.P.; Et al. Communities mobilizing for Change on vention: the search for a surrogate iii. Alcoholism: Clinical and Experimental Research 21(9):1695–1703, 1997. PmiD: 9438532 alcohol: outcomes from a randomized community trial. Journal of Studies on Alcohol 61(1):85–94, 2000a. PmiD: 10627101 tREno, a.J., anD lEE, J.P. approaching alcohol problems through local environmental interventions. Alcohol Research & Health 26(1):35–40, 2002. PmiD: 12154649 WagEnaaR, a.C.; muRRay, D.m.; anD toomEy, t.l. Communities mobilizing for Change on alcohol (CmCa): Effects of a randomized trial on arrests and traffic crashes. Addiction tREno, a.J.; CooPER, K.; anD RoEPER, P. Estimating alcohol involvement in trauma patients: 95(2):209–17, 2000b. PmiD: 10723849 search for a surrogate. Alcoholism: Clinical and Experimental Research 18(6):1306–1311, 1994. PmiD: 7695022 WagEnaaR, a.C.; muRRay, D.m.; WolFson, m.; Et al. Communities mobilizing for change on alcohol: Design of a randomized community trial. Journal of Community Psychology tREno, a.J.; gRuEnEWalD, P.J.; anD PoniCKi, W.R. use of iCD-9-Cm codes in the estimation of alcohol-involved injury: search for a surrogate ii. Alcoholism: Clinical and (special issue):79–101, 1994.

Experimental Research 20(2):320–326, 1996. PmiD: 8730224 WagEnaaR, a.C.; salois, m.J.; anD KomRo, K.a. Effects of beverage alcohol price and tax tREno, a.J.; gRuEnEWalD, P.J.; anD PoniCKi, W.R. the contribution of drinking patterns to levels on drinking: a meta-analysis of 1003 estimates from 112 studies. Addiction the relative risk of injury in six communities: a self-report based probability approach. 104(2):179–190, 2009. PmiD: 19149811 Journal of Studies on Alcohol 58(4):372–381, 1997. PmiD: 9203118 WElls, s.; Flynn, a.; gRaham, K.; Et al. using a mobile laboratory to study mental health, tREno, a.J.; gRuEnEWalD, P.J.; WooD, D.s.; anD PoniCKi, W.R. the price of alcohol: a con- addictions, and violence: a research plan. Challenges 2(1):1–18, 2011. sideration of contextual factors. Alcoholism: Clinical and Experimental Research 30(10):1734–1742, 2006. PmiD: 17010140 WhitE, a.m., anD sWaRtzWElDER, h.s. hippocampal function during adolescence: a unique target of ethanol effects. Annals of the New York Academy of Sciences 1021:206–220, tREno, a.J.; PoniCKi, W.R.; REmER, l.g.; anD gRuEnEWalD, P.J. alcohol outlets, youth drink- 2004. PmiD: 15251891 ing, and self-reported ease of access to alcohol: a constraints and opportunities approach. Alcoholism: Clinical and Experimental Research 32(8):1372–1379, 2008. Williams, a.F. alcohol-impaired driving and its consequences in the united states: the PmiD: 18540922 past 25 years. Journal of Safety Research 37(2):123–138, 2006. PmmiD: 16647085

Voas, R.b. Drinking and driving prevention in the community: Program planning and WolFE, a. 1973 u.s. national Roadside breathtesting survey: Procedures and Results. implementation. Addiction 92(suppl. 2):s201–s219, 1997. PmiD: 9231445 Washington, DC: u.s. Department of transportation, 1974.

Voas, R.b.; holDER, h.D.; anD gRuEnEWalD, P.J. the effect of drinking and driving interven- World health organization (Who). International Guide for Monitoring Alcohol tions on alcohol-involved traffic crashes within a comprehensive community trial. Consumption and Related Harm. geneva, switzerland: Who, 2000. available at: Addiction 92(suppl. 2):s221–s236, 1997. PmiD: 9231446 http://whqlibdoc.who.int/hq/2000/who_msd_msb_00.4.pdf. accessed august 31, 2012. Voas, R.b.; langE, J.; anD tREno, a.J. Documenting community-level outcomes: lessons young, D.J., anD biElinsKa-KWaPisz, a. alcohol consumption, beverage prices and mea- from drinking and driving. Evaluation Review 21(2):191–208, 1997. PmiD: 10183274 surement error. Journal of Studies on Alcohol 64(2):235–238, 2003. PmiD: 12713197 Voas, R.b.; WElls, J.; lEstina, D.; Et al. Drinking and driving in the united states: the young, D.J.; stoCKWEll, t.; ChERPitEl, C.J.; Et al. Emergency room injury presentations as 1996 national Roadside survey. Accident Analysis and Prevention 30(2):267–275, 1998. PmiD: 9450130 an indicator of alcohol-related problems in the community: a multilevel analysis of an international study. Journal of Studies on Alcohol 65(5):605–612, 2004. PmiD: 15536770 WagEnaaR, a.C., anD holDER, h.D. Effects of alcoholic beverage server liability on traffic crash injuries. Alcoholism: Clinical and Experimental Research 15(6):942–947, 1991. zaDoR, P.l. alcohol-related relative risk and fatal driver injuries in relation to driver age PmiD: 1789390 and sex. Journal of Studies on Alcohol 52(4):302–310, 1991. PmiD: 1875701

WagEnaaR, a.C., anD WolFson, m. Deterring sales and provision of alcohol to minors: a zaDoR, P.l.; KRaWChuK, s.a.; anD Voas, R.b. alcohol-related relative risk of driver fatalities study of enforcement in 295 counties in four states. Public Health Reports and driver involvement in fatal crashes in relation to driver age and gender: an update 110(4):419–427, 1995. PmiD: 7638329 using 1996 data. Journal of Studies on Alcohol 61(3):387–395, 2000. PmiD: 10807209

Community indicators: Assessing the impact of Alcohol 149 drinking prior to the injury event, is substantial. To accu- Focus on: the Burden of Alcohol rately assess the relationship between alcohol use and injury risk, ED studies generally have used probability sampling use—trauma and Emergency designs, in which all times of day and days of the week are represented equally. This approach circumvents biases associ- outcomes ated with sampling that might occur, for example, if samples were identified only on weekend evenings, when a higher prevalence of drinking and, possibly, of injury might be expected. Although the high prevalence rates mentioned Cheryl J. Cherpitel, Dr.P.H. above suggest that alcohol is an important risk factor for injury, they do not provide the information necessary to hospital emergency departments (EDs) see many patients with evaluate the actual level of risk for injury at which drinking alcohol-related injuries and therefore frequently are used to places the individual. assess the relationship between alcohol consumption and injury Data to establish drinking-related risk of both intentional risk. these studies typically use either case–control or and unintentional injury in ED samples generally have come case–crossover designs. Case–control studies, which compare from two types of study design: case–control studies and injured ED patients with either medical ED patients or the case–crossover studies. This article summarizes the findings general population, found an increased risk of injury after of these studies and explores specific aspects of the relation- alcohol consumption, but differences between the case and ship between alcohol use and injury risk. control subjects partly may account for this effect. Case–crossover designs, which avoid this potential confounding factor by using the injured patients as their own control Risk of injury in ED Studies subjects, also found elevated rates of injury risk after alcohol consumption. however, the degree to which risk is increased can vary depending on the study design used. other factors Case–Control Studies influencing injury risk include concurrent use of other drugs and Two types of case–control studies have been used to estimate drinking patterns. additional studies have evaluated cross- the risk of injury from drinking for patients treated in the country variation in injury risk as well as the risk by type (i.e., ED. The most commonly used type of case–control study intentional vs. unintentional) and cause of the injury. Finally, ED uses noninjured (i.e., medical) patients attending the same studies have helped determine the alcohol-attributable fraction ED during the same period of time as quasi-control subjects. of injuries, the causal attribution of injuries to drinking, and the These patients presumably come from the same geographic impact of others’ drinking. although these studies have some area as the injured patients and likely share other characteris- limitations, they have provided valuable insight into the tics (e.g., socioeconomic status). Researchers conducted a association between drinking and injury risk. kEY WoRDS: Alcohol meta-analysis of 15 ED studies conducted in 7 countries that consumption; alcohol-related injury; alcohol and drug participated in the Emergency Room Collaborative Alcohol related–injury; alcohol-attributable fractions; risk factors; Analysis Project (ERCAAP) (Cherpitel et al. 2003a) and alcohol and other drug–induced risk; hospital; emergency which all used the same methodology and instrumentation. department; emergency room; emergency care; trauma; The studies only included those patients who arrived at the injury; intentional injury; unintentional injury; patients; ED within 6 hours of the injury event and excluded those case–control studies; case–crossover studies medical patients who primarily were admitted to the ED for alcohol intoxication or withdrawal symptoms. The meta- analysis found a pooled odds ratio (OR) of injury associated lcohol consumption is a leading risk factor for morbidity with a positive BAC (≥0.01 percent) of 2.4 (95% CI = 1.9– and mortality related to both intentional (i.e., violence- 3.0);1 moreover, the OR was higher (OR = 2.9) for patients Arelated) and unintentional injury. In 2000, 16.2 percent with higher BAC levels (≥0.10 percent) (Ye and Cherpitel of deaths and 13.2 percent of disability-adjusted life years (DALYs) from injuries, worldwide, were estimated to be attributable to alcohol (Rehm et al. 2009). Alcohol affects 1 the oR is the ratio between the risk that a person with a certain characteristic (e.g., positive baC) experiences a certain outcome (e.g., an injury) and the risk that a person without that characteristic psychomotor skills, including reaction time, as well as cogni- experiences the same outcome. in other words, an oR of 2.4 indicates that people who have a posi- tive skills, such as judgment; as a result, people drinking alcohol tive baC are 2.4 times as likely to be injured as people without a positive baC. often place themselves in high-risk situations for injury. Much of the data linking alcohol with nonfatal injuries have come from studies conducted in hospital emergency departments (EDs). As described in this article, in these set- Cheryl J. Cherpitel, Dr.P.H. is a senior scientist in the tings the prevalence of alcohol involvement in the patients’ Alcohol Research Group, Emeryville, California. injuries, as measured by a positive blood alcohol concentra- tion (BAC) at the time of arrival in the ED or self-reported

150 Alcohol Research: Current Reviews 2009). A similar likelihood of injury (OR = 2.1) was found • The usual-frequency approach. This approach compares for patients who reported drinking within 6 hours prior to the patients’ drinking in the 6 hours preceding the injury the injury event, regardless of time of arrival in the ED. to their expected drinking during that time, based on One concern with this approach of using medical patients their usual frequency of drinking. In a study using this as control subjects for injured patients is the possibility of approach that included 28 EDs across 16 countries, the underestimating the true risk of drinking associated with estimated ORs ranged from 1.05 (Canada) to 35.0 injury. Noninjured patients have been found to be heavier (South Africa), with a pooled estimate of 5.69 (95% drinkers than people in the general population from which CI = 4.04–8.00) (Borges et al. 2006a). they come who do not seek emergency care (Cherpitel 1993). Thus, these patients may be attending the ED for conditions related to their drinking (in addition to those Comparison of Methods to Estimate Risk associated with alcohol intoxication or withdrawal). In the second type of case–control study used to estimate The results described above indicate that the estimates of risk risk of injury from drinking in ED patient samples, people of injury in samples from the same country can vary depend- in the general population of the community from which the ing on the method used. For example, in analyses across ED patients come are used as control subjects. These indi- eight countries participating in ERCAAP, analyses using the viduals presumably are free of conditions that may be related case–control method found that the pooled OR of injury for to their drinking. Only four such studies have been reported self-reported drinking prior to the event was 2.1, compared to date, including two from Australia (Mcleod et al. 1999; with an OR of 5.2 when the usual-frequency method of Watt et al. 2004) and one each from the United States (Vinson case–crossover analysis was used (Ye and Cherpitel 2009). et al. 2003) and Mexico (Borges et al. 1998). In these studies, Furthermore, the World Health Organization (WHO) the ORs ranged from 6.7 in the Mexican study to 3.1 in the Collaborative Study on Alcohol and Injury, which used the U.S. study and around 2.0 in the Australian studies. Moreover, case–crossover method across 12 countries, found a pooled both the U.S. and the Australian studies demonstrated a OR of injury of 6.8 using the usual-frequency approach, dose-response relationship. compared with 5.7 using the matched-interval approach (Borges et al. 2006b). Case–control designs may underesti- Case–Crossover Studies mate the risk of injury if noninjured control subjects are presenting to the ED with other conditions related to their The second study design used to estimate the risk of injury drinking, whereas both the matched-interval and usual- from alcohol consumption is the case–crossover study frequency approaches to the case–crossover design are subject (Maclure 1991). This approach is thought to circumvent to recall bias of drinking in the past. at least some of the problems raised with the case–control design, such as demographic and others differences between case and control subjects that may be related to both alcohol consumption and likelihood of injury. There are two approaches Effects of other Factors on Risk of injury to the case–crossover design, both of which use injured patients as their own control subjects, thereby theoretically reducing confounding of the alcohol–injury relationship Effects of other drug use from stable risk factors, such as age and gender. None of these estimates of risk of injury related to drinking • The matched-interval approach. Studies using the matched- have taken into consideration other drug use at the time interval approach compare drinking within 6 hours prior of injury, although multiple substances commonly are used to the injury event with drinking during a predetermined together in ED populations (Buchfuhrer and Radecki 1996). control period, such as the same 6-hour period during Other drug use might be expected to elevate the risk of the previous day or previous week. Such studies have reported injury, either alone or in combination with alcohol; however, ORs ranging from 3.2 (based on any drinking at the this may not be the case. One study found an OR of 3.3 for same time the previous day) (Vinson et al. 2003) to 5.7 drinking within 6 hours prior to injury and an OR of 3.0 in a 10-country study (based on any drinking at the same for drinking in combination with other drug use during the time the previous week) (Borges et al. 2006b). Both studies same time; in contrast, drug use alone had no significant demonstrated a dose-response relationship. Thus, the analysis effect on risk (Cherpitel et al. 2012b). It is important to con- of Vinson and colleagues (2003) determined ORs ranging sider that in this study the majority of drug users reported from 1.8 with consumption of 1 to 2 drinks prior to injury using marijuana. However, given their different pharmaco- to 17 with consumption of 7 or more drinks. Likewise, logical properties, all drugs would not be expected to act in a Borges and colleagues (2006b) found ORs ranging from similar manner, either alone or in combination with alcohol. 3.3 with consumption of one to two drinks to 10.1 with Consequently, in other populations with different drug use consumption of six or more drinks prior to injury. patterns the findings might be different.

Focus on trauma and Emergency outcomes 151 Effects of usual drinking Patterns been found to have a higher risk of injury related to alcohol The risk of injury from drinking prior to the event (i.e., than those with lower DDP scores (Cherpitel et al. 2005b). acute consumption) also is influenced by the drinker’s usual drinking patterns (i.e., chronic consumption). Cherpitel and colleagues (2004) found that the risk of injury from drink- Risk by type and Cause of injury ing prior to the event was lower among frequent heavy Risk of injury from alcohol also varies by type (i.e., inten- drinkers than among infrequent heavy drinkers, suggesting tional vs. unintentional) and cause of injury. For example, that heavier drinkers may have developed tolerance against some adverse effects of alcohol that lead to injury. Likewise, Macdonald and colleagues (2006) found that the risk was in an analysis by Gmel and colleagues (2006), the risk of highest for violence-related (i.e., intentional) injuries. A injury was greater among usual light drinkers who occasion- case–crossover analysis using the usual-frequency approach ally drink heavily (i.e., report episodic heavy drinking) than that included data from 15 countries in the ERCAAP and among people who usually drink heavily but report no WHO projects found that greater variations across countries episodic heavy drinking or among people who usually drink existed in risk of an intentional injury than in risk of unin- heavily as well as report episodic heavy drinking. tentional injury; this difference was at least in part explained by the level of DDP in a country (Cherpitel and Ye 2010). Overall, the pooled OR for intentional injury related to Risk of Alcohol-Related injury drinking in these countries was 21.5, compared with 3.37 for unintentional injury (Borges et al. 2009). Furthermore, Although acute alcohol consumption, modified by drinking the risk of intentional injury showed a greater dose–response pattern, has been found to be associated with risk of injury, association than the risk of unintentional injury (Borges et drinking pattern also has been found to be associated with al. 2009). Thus, the ORs for intentional injuries ranged risk of an alcohol-related injury2 (defined as drinking within from 11.14 for one to two drinks prior to injury to 35.57 6 hours prior to injury), with frequency of drinking among non–heavy drinkers (Cherpitel et al. 2003b) and both for five or more drinks during this time, whereas the ORs episodic and frequent heavy drinking predictive of alcohol- for unintentional injuries ranged from 3.86 to 6.4, respec- related injury (Cherpitel et al. 2012c). An analysis of com- tively. Among the unintentional injuries, the risk also varied bined data from ERCAAP and from the WHO Collaborative depending on the cause of the injury. For example, the OR Study on Alcohol and Injury across 16 countries found the was 5.24 for traffic-related injuries, compared with 3.39 for pooled risk of alcohol-related injury was increased with injuries related to falls. heavy episodic drinking (OR = 2.7) as well as with chronic high-volume drinking (OR = 3.5); moreover, the risk was highest for people reporting both patterns of drinking (OR Alcohol-Attributable Fraction = 6.1) (Ye and Cherpitel 2009). Another variable that has been studied in the context of assessing the risk of injuries after drinking is the alcohol- Cross-country variation in Risk of injury attributable fraction (AAF). This variable represents the proportional reduction in injury that would be expected if A great deal of variation has been found across countries in the risk factor (i.e., drinking prior to injury) was absent; it risk of injury and risk of alcohol-related injury, and this het- reflects the burden of injury in a given society that results erogeneity seems to be associated with a country’s level of from alcohol use. The AAF also varies across countries in detrimental drinking pattern (DDP). The DDP score, ED studies, because it is related to both the risk of injury which is based on aggregate survey data and key informant and the prevalence of alcohol-related injury. In a case–control surveys, is a measure developed for comparative risk assess- study of 14 EDs from six countries in ERCAAP, the AAF ment in the WHO’s Global Burden of Disease study (Rehm based on self-reported drinking within 6 hours prior to the et al. 2004). It includes such indicators of drinking patterns injury event varied from 0.5 percent to 18.5 percent for all as heavy drinking occasions, drinking with meals, and drink- ing in public places. The DDP has been assessed in a large types of injury, and from 19.1 percent to 83.3 percent for number of countries around the world as a measure of the intentional injury (Cherpitel et al. 2005a). The pooled esti- “detrimental impact” on health, and other drinking-related mate from all EDs for the AAF was 5.8 percent for all types harms, at a given level of alcohol consumption (Rehm et al. of injury and 42.5 percent for intentional injury. In other 2001, 2003). Countries with a higher level of DDP have words, more than 40 percent of all intentional injuries would not have occurred if the people involved had not been drinking. Moreover, the investigators determined higher AAF estimates for male than female subjects for both unintentional injuries 2 as used here, the term “alcohol-related injury” refers to injuries where the patient reported using alcohol in the 6-hour period immediately preceding the injury; in contrast, the term “injury” is used (5.5 percent vs. 1.7 percent) and intentional injuries (50.0 here to refer to any injury, regardless of whether it was preceded by alcohol use or not. percent vs. 7.7 percent).

152 Alcohol Research: Current Reviews Causal Attribution the ERCAAP and WHO studies across 16 countries (Cherpitel et al. 2007). The ED studies in the ERCAAP and WHO projects also The studies reported here all have been conducted in assessed the patients’ causal attribution of their injuries to EDs, rather than in trauma centers that generally treat the their drinking—that is, patients were asked whether they most serious injury cases and, consequently, are less con- believed the injury would have occurred if they had not been ducive to the detailed data collection effort required in studies drinking. In an evaluation that included 15 countries, one- of alcohol and injury, unless the patient is admitted to the half of the patients who reported drinking prior to injury hospital. It is unknown how this may affect the resulting also reported a causal attribution (Cherpitel et al. 2006). conclusions regarding the rates of the risk of injury from This information was used to establish a subjective AAF—an drinking, because the literature has been mixed regarding AAF derived from the patient’s own causal attribution of alcohol’s association with injury severity. their injury to drinking. This subjective AAF then was com- As noted earlier, some limitations also apply to the methods pared to the AAF obtained using the standard formula based that have been used to estimate the risk of injury related to on the relative risk of injury from alcohol and prevalence of alcohol consumption. Case–control studies may underesti- drinking in the 6-hour period (i.e., the objective AAF) from mate this risk because the medical patient controls also may the six ERCAAP countries, as described above. This compar- have drinking-related conditions. The matched-interval approach ison found that for unintentional injuries, the subjective to case–crossover analyses eliminates the heaviest drinkers AAF generally was somewhat higher than the objective AAF. (i.e., those who report drinking both during the period pre- For intentional injuries, however, the subjective AAF was ceding the injury and during the control period), which may substantially lower (i.e., 5.9 percent to 46.7 percent) than lead to underestimates of the risk of injury for these drinkers. the objective AAF (i.e., 24.9 percent to 83.3 percent) (Bond Likewise, the usual-frequency approach may underestimate and Macdonald 2009). the risk of injury for heavy drinkers because of the increase in expected drinking occasions for the heaviest drinkers. In addition, when estimating risk of injury using the case– others’ Drinking crossover approach, it is important to consider the activity in which the patient was engaged at the time of injury. For example, Researchers also increasingly are interested in studying the for a patient injured in a motor vehicle accident who had harm, including injury, resulting from other people’s drink- been drinking, the comparison with the control time interval ing. Evaluating these so-called externalities is important for a only would be valid if the patient also had been in a motor fuller understanding of the burden of alcohol-related injury vehicle during the control interval. Otherwise, the patient in society. To assess such externalities, investigators for the would not have been exposed to the risk of incurring a motor ED studies in the WHO project also obtained data on whether vehicle–related injury, regardless of whether he or she had the patient being treated for a violence-related injury believed been drinking. This is an important consideration in future the other person had been drinking. Across the 14 countries, studies that seek to examine risk of injury related to alcohol. from 14 percent to 73 percent of the victims believed that Lastly, the risk of injury related to drinking likely is others definitely had been drinking. Based on these data, the affected by a number of individual-level characteristics such pooled estimate for the AAF was 38.8 percent when both as age, gender, and risk-taking disposition, as well as by soci- victim and perpetrator were considered, compared with an etal-level characteristics such as detrimental drinking pattern, AAF of 23.9 percent when only the patient was considered as discussed above. Estimates of AAFs for injury, which are (Cherpitel et al. 2012a). required for determining the global burden of disease for injury related to alcohol, generally have not taken these vari- Considerations and Limitations in Estimating Risk ables into consideration, and this is a necessary direction for future research on the burden alcohol-related injury puts on of injury society. ■ The data reported here on the risk of injury primarily were derived from patients’ self-reports of drinking prior to injury. Although the ED studies all estimated the patient’s BAC at Acknowledgements the time of ED admission based on breath alcohol levels, Work on this article was supported by National Institute on self-reports seem to be a better measure of drinking, because Alcohol Abuse and Alcoholism grant 2–RO1–AA–013750–04. in many cases a substantial period of time may have lapsed between the patient’s last drink, the injury event, and arrival at the ED. As a result, the BAC may be negative even though Financial Disclosure the patient reports drinking prior to injury. Indeed, this dis- crepancy has been found in an analysis of the concordance The author declares that she has no competing financial between self-reported drinking and BAC measurements in interests.

Focus on trauma and Emergency outcomes 153 References ChERPitEl, C.J.; yE, y.; bonD, J.; Et al. attribution of alcohol to violence-related injury: self and other’s drinking in the event. Journal of Studies on Alcohol and Drugs bonD, J., anD maCDonalD, s. Causality and causal attribution of alcohol in injuries. in: 73:227–284, 2012a. PmiD: 22333335 Cherpitel, C.; borges, g.; giesbrecht, n.; et al., Eds. Alcohol and Injuries: Emergency ChERPitEl, C.J.; yE, y.; bonD, J.; Et al. multi-level analysis of injury risk and drinking pat- Department Studies in an International Perspective. geneva, switzerland: World health tern: Emergency department data from 19 countries. Addiction 107(7):1263–1272, organization, 2009, pp. 27–40. 2012c. PmiD: 22236278 boRgEs, g.; ChERPitEl, C.J.; mEDina-moRa, m.E.; Et al. alcohol consumption in emergency room patients and the general population: a population based study. Alcoholism: ChERPitEl, C.J.; yE, y.; WattERs, K.; Et al. Risk of injury from alcohol and drug use in the Clinical and Experimental Research 22(9):1986–1991, 1998. PmiD: 9884142 emergency department: a case-crossover study. Drug and Alcohol Review 31(4):431– 438, 2012b. PmiD: 21824208 boRgEs, g.; ChERPitEl, C.J.; oRozCo, R.; Et al. acute alcohol use and the risk of non-fatal injury in sixteen countries. Addiction 101(7):993–1002, 2006a. PmiD: 16771891 gmEl, g.; bissERy, a.; gammEtER, R.; Et al. alcohol-attributable injuries in admissions to a swiss emergency room—an analysis of the link between volume of drinking, drinking boRgEs, g.; ChERPitEl, C.J.; oRozCo, R.; Et al. multicentre study of acute alcohol use and non-fatal injuries: Data from the Who collaborative study on alcohol and injuries. patterns, and preattendance drinking. Alcoholism: Clinical and Experimental Research Bulletin of the World Health Organization 84(6):453–460, 2006b. PmiD: 17199729 30(3):501–509, 2006. PmiD: 16499491

boRgEs, g.; maCDonalD, s.; ChERPitEl, C.J.; Et al. Variation in alcohol-related injury by type maCDonalD, s.; ChERPitEl, C.J.; DEsouza, a.; Et al. Variations of alcohol impairment in dif- and cause. in: Cherpitel, C.; borges, g.; giesbrecht, n.; et al., Eds. Alcohol and Injuries: ferent types, causes, and contexts of injuries: Results of emergency room studies from Emergency Department Studies in an International Perspective. geneva, switzerland: 16 countries. Accident: Analysis and Prevention 38(6):1107–1112, 2006. PmiD: 16828047 World health organization, 2009, pp. 15–25. maCluRE, m. the case-crossover design: a method for studying transient effects on the buChFuhRER, l.a., anD RaDECKi, s.E. alcohol and drug abuse in an urban trauma center: risk of acute events. American Journal of Epidemiology 133(2):144–153, 1991. PmiD: Predictors of screening and detection. Journal of Addictive Diseases 15(1):65–74, 1985444 1996. PmiD: 8729147 mClEoD, R.; stoCKWEll, t.; stEVEns, m.; anD PhilliPs, m. the relationship between alcohol ChERPitEl, C.J. alcohol consumption among emergency room patients: Comparison of consumption patterns and injury. Addiction 94(11):1719–1734, 1999. PmiD: 10892010 county/community hospitals and an hmo. Journal of Studies on Alcohol 54(4):432–440, 1993. PmiD: 8341045 REhm, J.; montEiRo, m.; Room, R.; Et al. steps towards constructing a global comparative risk analysis for alcohol consumption: Determining indicators and empirical weights for ChERPitEl, C.J., anD yE, y. alcohol and violence-related injuries among emergency room patients in an international perspective. Journal of the American Psychiatric Nurses patterns of drinking, deciding about theoretical minimum, and dealing with different Association 16(4):277–235, 2010. PmiD: 20824198 consequences. European Addiction Research 7(3):138–147, 2001. PmiD: 11509844

ChERPitEl, C.J.; bonD, J.; yE, y.; Et al. a cross-national meta-analysis of alcohol and REhm, J.; PoPoVa, s.; anD PatRa, J. alcohol-attributable injury in a global perspective. in: injury: Data from the Emergency Room Collaborative alcohol analysis Project (ERCaaP). Cherpitel, C.; borges, g.; giesbrecht, n.; et al., Eds. Alcohol and Injuries: Emergency Addiction 98(9):1277–1286, 2003a. PmiD: 12930215 Department Studies in an International Perspective. geneva, switzerland: World health organization, 2009, pp. 41–51. ChERPitEl, C.J.; bonD, J.; yE, y.; Et al. alcohol-related injury in the ER: a cross-national meta-analysis from the Emergency Room Collaborative alcohol analysis Project REhm, J.; REhn, n.; Room, R.; Et al. the global distribution of average volume of alcohol (ERCaaP). Journal of Studies on Alcohol 64(5):641–649, 2003b. PmiD: 14572186 consumption and patterns of drinking. European Addiction Research 9(4):147–156,

ChERPitEl, C.J.; bonD, J.; yE, y.; Et al. multi-level analysis of causal attribution of injury to 2003. PmiD: 12970583 alcohol and modifying effects: Data from two international emergency room projects. REhm, J.; Room, R.; montEiRo, m.; Et al. alcohol use. in: Ezzati, m.; lopez, a.D.; Rodgers, Drug and Alcohol Dependence 82(3):258–268, 2006. PmiD: 16257137 a.; and murray, C.J.l.; eds. Comparative Quantification of Health Risks: Global and ChERPitEl, C.J.; yE, y.; anD bonD, J. alcohol and injury: multi-level analysis from the Regional Burden of Disease Attributable to Selected Major Risk Factors. Vol. 1. geneva, Emergency Room Collaborative alcohol analysis Project (ERCaaP). Alcohol and switzerland: World health organization, 2004, pp. 959–1108. Alcoholism 39(6):552–558, 2004. PmiD: 15351747 Vinson, D.C.; maCluRE, m.; REiDingER, C.; anD smith, g.s. a population-based case- ChERPitEl, C.J.; yE, y.; anD bonD, J. attributable risk of injury associated with alcohol use: crossover and case-control study of alcohol and the risk of injury. Journal of Studies on Cross-national data from the Emergency Room Collaborative alcohol analysis Project. Alcohol 64(3):358–366, 2003. PmiD: 12817824 American Journal of Public Health 95(2):266–272, 2005a. PmiD: 15671463 Watt, K.; PuRDiE, D.m.; RoChE, a.m.; anD mCCluRE, R.J. Risk of injury from acute alcohol ChERPitEl, C.J.; yE, y.; bonD, J.; Et al. Validity of self-reported drinking before injury com- consumption and the influence of confounders. Addiction 99(10):1262–1273, 2004. pared with a physiological measure: Cross-national analysis of emergency-department data from 16 countries. Journal of Studies on Alcohol and Drugs 68(2):296–302, PmiD: 15369564

2007. PmiD: 17286349 yE, y., anD ChERPitEl, C.J. Risk of injury associated with alcohol and alcohol-related

ChERPitEl, C.J.; yE, y.; bonD, J.; Et al. multi-level analysis of alcohol-related injury among injury. in: Cherpitel, C.; borges, g.; giesbrecht, n.; et al., Eds. Alcohol and Injuries: emergency department patients: a cross-national study. Addiction 100(12):1840– Emergency Department Studies in an International Perspective. geneva, switzerland: 1850, 2005b. PmiD: 16367985 World health organization, 2009, pp. 3–13.

154 Alcohol Research: Current Reviews Reduce the Harmful Use of Alcohol, which was passed in Chronic Diseases and May 2010. Of growing concern are noncommunicable chronic diseases and conditions that have been shown to Conditions Related to contribute substantially to the alcohol-attributable burden of disease (Rehm et al. 2009). Specifically, in 2004 an estimated 35 million deaths and 603 million disability-adjusted life- Alcohol Use years (DALYs) lost were caused by chronic diseases and con- ditions globally (WHO 2008); alcohol was responsible for 3.4 percent of the deaths and 2.4 percent of DALYs caused by these conditions (Parry et al. 2011). To address the burden kevin D. Shield, M.H.Sc.; Charles Parry, Ph.D.; and of chronic diseases and conditions, the United Nation (UN) General Assembly passed Resolution 64/265 in May of Jürgen Rehm, Ph.D. 2010, calling for their prevention and control (UN 2010). This resolution is intended to garner multisectoral commitment alcohol consumption is a risk factor for many chronic diseases and facilitate action on a global scale to address the fact that and conditions. the average volume of alcohol consumed, alcohol (together with tobacco, lack of exercise, and diet) consumption patterns, and quality of the alcoholic beverages plays a significant role in chronic diseases and conditions. It consumed likely have a causal impact on the mortality and is noteworthy that cardiovascular diseases, cancers, and diabetes morbidity related to chronic diseases and conditions. twenty- in particular have been highlighted for targeted action (UN five chronic disease and condition codes in the international 2010) because alcohol is a risk factor for many cardiovascular Classification of Disease (iCD)-10 are entirely attributable to diseases and cancers and has both beneficial and detrimental alcohol, and alcohol plays a component-risk role in certain effects on diabetes and ischemic cardiovascular diseases,1 cancers, other tumors, neuropsychiatric conditions, and depending on the amount of alcohol consumed and the numerous cardiovascular and digestive diseases. Furthermore, patterns of consumption. alcohol has both beneficial and detrimental impacts on diabetes, Building on previous reviews concerning alcohol and disease ischemic stroke, and ischemic heart disease, depending on the (Rehm et al. 2003a, 2009), this article presents an up-to- overall volume of alcohol consumed, and, in the case of date and in-depth overview of the relationship of alcohol ischemic diseases, consumption patterns. however, limitations consumption and high-risk drinking patterns and the initia- exist to the methods used to calculate the relative risks and alcohol-attributable fractions. Furthermore, new studies and 1 ischemic cardiovascular diseases are those caused by a blockage of blood vessels, resulting in a confounders may lead to additional diseases being causally loss of blood supply to the tissue serviced by the affected blood vessels. linked to alcohol consumption, or may disprove the relationship between alcohol consumption and certain diseases that currently are considered to be causally linked. these limitations do not affect the conclusion that alcohol consumption significantly contributes to the burden of chronic diseases and kevin D. Shield, M.H.Sc., is a Ph.D. student in medical conditions globally, and that this burden should be a target for science at the University of Toronto. He also is affiliated intervention. kEY WoRDS: Alcohol consumption; alcohol use with the Centre for Addiction and Mental Health (CAMH) frequency; chronic diseases; disorders; mortality; morbidity; in Toronto, Canada. alcohol-attributable fractions (AAF); risk factors; relative risk; AoD-induced risk; cancers; neuropsychiatric disorders; Charles Parry, Ph.D., is the director of the Alcohol & cardiovascular diseases; digestive diseases; diabetes; Drug Abuse Research Unit at the South African Medical ischemic stroke; ischemic heart disease; burden of disease Research Council, Cape Town, South Africa, and extraordi- nary professor in the Department of Psychiatry, Stellenbosch University, Cape Town, South Africa. lcohol has been a part of human culture for all of recorded history, with almost all societies in which Jürgen Rehm, Ph.D., is director of the Social and Aalcohol is consumed experiencing net health and social Epidemiological Research Department at the Centre for problems (McGovern 2009; Tramacere et al. 2012b, c). Addiction and Mental Health, chair and professor in the With the industrialization of alcohol production and the Dalla Lana School of Public Health, University of Toronto, globalization of its marketing and promotion, alcohol con- Canada, and section head at the Institute for Clinical sumption and its related harms have increased worldwide (see Alcohol Consumption Trends, in this issue). This has Psychology and Psychotherapy, Technische Universität prompted the World Health Organization (WHO) to pass Dresden, Dresden, Germany. multiple resolutions to address this issue over the past few years, including the World Health Assembly’s Global Strategy to

Chronic Diseases and Conditions Related to Alcohol Use 155 tion/exacerbation and treatment of various chronic diseases These include the following: and conditions. It also assesses the methods used to calculate • The toxic and beneficial biochemical effects of beverage the impact of alcohol consumption on chronic diseases and alcohol (i.e., ethanol) and other compounds found in conditions. alcoholic beverages; • The consequences of intoxication; and Alcohol Consumption As a Risk Factor for Chronic • The consequences of alcohol dependence. Diseases and Conditions These intermediate mechanisms have been reviewed in Figure 1 presents a conceptual model of the effects of alcohol more detail by Rehm and colleagues (2003a). consumption on morbidity and mortality and of the influ- ence of both societal and demographic factors on alcohol consumption and alcohol-related harms resulting in chronic Chronic Diseases and Conditions Related to Alcohol diseases and conditions (adapted from Rehm et al. 2010a). According to this model, two separate, but related, measures Chronic diseases and Conditions Entirely Attributable of alcohol consumption are responsible for most of the to Alcohol causal impact of alcohol on the burden of chronic diseases and conditions—overall volume of alcohol consumption Of the chronic diseases and conditions causally linked with and patterns of drinking. The overall volume of alcohol con- alcohol consumption, many categories have names indicat- sumption plays a role in all alcohol-related diseases, whereas ing that alcohol is a necessary cause—that is, that these drinking patterns only affect ischemic cardiovascular dis- particular diseases and conditions are 100 percent alcohol eases. In addition to the overall volume and pattern of con- attributable. Of these, alcohol use disorders (AUDs)—that sumption, the quality of the alcoholic beverages consumed is, alcohol dependence and the harmful use of alcohol as also may influence mortality and morbidity from chronic defined by the International Classification of Disease, Tenth diseases and conditions. However, this pathway is of less Edition (ICD–10)—certainly are the most important cate- importance from a public health perspective (Lachenmeier gories, but many other diseases and conditions also are and Rehm 2009; Lachenmeier et al. 2007) because it has a entirely attributable to alcohol (see table 1). much smaller impact than the other two factors. The effects of overall volume of alcohol consumed, con- Chronic diseases and Conditions for Which Alcohol sumption patterns, and quality of the alcoholic beverages Is a Component Cause consumed on mortality and morbidity from chronic diseases Alcohol is a component cause for more than 200 other dis- and conditions are mediated by three main mechanisms. eases and conditions with ICD–10 three-digit codes—that

Figure 1 Causal model of alcohol consumption, intermediate mechanisms, and long-term consequences, as well as of the influence of societal and demographic factors on alcohol consumption and alcohol-related harms resulting in chronic diseases and conditions.

souRCE: adapted from Rehm et al. 2010a.

156 Alcohol Research: Current Reviews is, alcohol consumption is not necessary for the diseases to Thus, two recent meta-analyses found no association between develop (Rehm et al. 2010a). For these conditions, alcohol alcohol drinking status (i.e., drinkers compared with non- shows a dose-response relationship, where the risk of onset of drinkers) and risk of gastric cardia adenocarcinoma (Tramacere or death from the disease or condition depends on the total et al. 2012a, d). However, one meta-analysis did find an volume of alcohol consumed (Rehm et al. 2003a). Table 2 association between heavy alcohol consumption and the risk outlines these chronic diseases and conditions that are associ- of this type of cancer (Tramacere et al. 2012a). ated with alcohol consumption and lists the source of the For several types of cancer investigators have found a relative risk (RR) functions if the chronic disease or condi- nonsignificant positive association with alcohol consump- tion is included as an alcohol-attributable harm in the 2005 tion, including endometrial (Bagnardi et al. 2001; Rota et al. Global Burden of Disease (GBD) Study.2 Several of these 2012), ovarian (Bagnardi et al. 2001), and pancreatic cancers chronic diseases and conditions are singled out for further (Bagnardi et al. 2001). However, because the relationship at discussion in the following sections to highlight alcohol’s least between alcohol consumption and endometrial and causative or protective role. pancreatic cancer is modest (i.e., the point estimates of RR are low, even at high levels of average daily alcohol consump- tion), additional studies with large numbers of participants Specific Chronic Diseases and Conditions are needed to accurately assess the relationship (Bagnardi et Associated With Alcohol Consumption al. 2001). The relationship between alcohol consumption and bladder and lung cancers is even less clear, with one meta-analysis finding that alcohol significantly increases the Malignant neoplasms risk for both types of tumors (Bagnardi et al. 2001), whereas The relationship between alcohol consumption and cancer already was suggested in the early 20th century, when Lamy table 1 Chronic Diseases and Conditions that are, by Definition, (1910) observed that patients with cancer either of the alcohol attributable (i.e., Require alcohol Consumption as a esophagus or of the cardiac region were more likely to be necessary Cause) alcoholics. The accumulation of evidence supporting the iCD–10 relationship between ethanol and cancers led the International Code Disease Agency for Research on Cancer (IARC) to recognize the cancer-inducing potential (i.e., carcinogenicity) of ethanol in F10 mental and behavioral disorders attributed to the use animal models and to conclude that alcoholic beverages are of alcohol carcinogenic to humans (IARC 2008). Specifically, the GBD F10.0 acute intoxication study found that alcohol increased the risk of cancers of the F10.1 harmful use upper digestive track (i.e., mouth and oropharynx, esopha- F10.2 Dependence syndrome gus, and larynx), the lower digestive track (i.e., colon, rec- F10.3 Withdrawal state tum, and liver), and the female breast (see figure 2). More F10.4 Withdrawal state with delirium up-to-date systematic reviews and meta-analyses on alcohol F10.5 Psychotic disorder consumption and the risk of developing cancer have been F10.6 amnesic syndrome published by Fedirko and colleagues (2011) for colorectal F10.7 Residual and late-onset psychotic disorder cancer, Islami and colleagues (2011) for esophageal squamous F10.8 other mental and behavioral disorders cell carcinoma, Islami and colleagues (2010) for laryngeal F10.9 unspecified mental and behavioral disorder cancer, and Tramacere and colleagues (2010) and Turati and g31.2 Degeneration of nervous system attributed to alcohol colleagues (2010) for oral and pharyngeal cancers. g62.1 alcoholic polyneuropathy A recent meta-analysis also has indicated that alcohol g72.1 alcoholic myopathy consumption is significantly linked to an increased risk of i42.6 alcoholic cardiomyopathy developing prostate cancer in a dose-dependent manner K29.2 alcoholic gastritis (Rota et al. 2012); this observation is consistent with previ- K70 alcoholic liver disease ous meta-analyses concluding that alcohol consumption and K70.0 alcoholic fatty liver the risk for prostate cancer are significantly correlated (Dennis K70.1 alcoholic hepatitis 2000; Fillmore et al. 2009). Additional research, however, is K70.2 alcoholic fibrosis and sclerosis of liver required on the biological pathways to prove the role of alcohol K70.3 alcoholic cirrhosis of liver consumption in the development of this type of cancer. K70.4 alcoholic hepatic failure Evidence also has suggested that stomach cancer may be K70.9 alcoholic liver disease, unspecified linked to ethanol consumption (Bagnardi et al. 2001; Tramacere K85.2 alcohol-induced acute pancreatitis et al. 2012a); however, the findings have not been unequivocal. K86.0 alcohol-induced chronic pancreatitis P04.3 Fetus and newborn affected by maternal use of alcohol

2 the gbD study is a project that aims to provide a consistent and comparative description of the Q86.0 Fetal alcohol syndrome (dysmorphic) global burden of diseases and injuries and the risk factors that cause them.

Chronic Diseases and Conditions Related to Alcohol Use 157 Table 2 Chronic Diseases and Conditions for Which Alcohol Consumption Is a Component Cause, Identified by Various Meta-Analyses and Reviews and Listed in the 2005 Global Burden of Disease (GBD) Study

No. of 2005 Disease ICD–10 Effect Level of Evidence Used if Included in the GBD Study GBD Code Meta-Analysis

IIA Malignant neoplasms

IIA1 Mouth cancer C00–C08 Detrimental Causally related International Agency for Research on Cancer 2008 (based on relative risks from Corrao et al. 2004)

IIA2 Nasopharynx cancer C09–C13 Detrimental Causally related International Agency for Research on Cancer 2008 and other pharynx (based on relative risks from Corrao et al. 2004) cancers

IIA3 Esophagus cancer C15 Detrimental Causally related International Agency for Research on Cancer 2008 (based on relative risks from Corrao et al. 2004)

IIA4 Stomach cancer C16 Detrimental Insufficient causal evidence

IIA5 Colon and rectum C18–C21 Detrimental Causally related International Agency for Research on Cancer 2008 cancers (based on relative risks from Corrao et al. 2004)

IIA6 Liver cancer C22 Detrimental Causally related International Agency for Research on Cancer 2008 (based on relative risks from Corrao et al. 2004)

IIA9 Larynx cancer C32 Detrimental Causally related International Agency for Research on Cancer 2008 (based on relative risks from Corrao et al. 2004)

IIA10 Trachea, bronchus, C33–C34 Detrimental Insufficient causal and lung cancers evidence

IIA13 Breast cancer (women C50 Detrimental Causally related International Agency for Research on Cancer 2008 only) (based on relative risks from Corrao et al. 2004)

IIA16 Ovarian cancer C56 Detrimental Insufficient causal evidence

IIA17 Prostate cancer C61 Detrimental Insufficient causal evidence

IIA19 Kidney and other urinary C64–C66, Beneficial (renal Insufficient causal organ cancers C68 (except cell carcinoma evidence C68.9) only)

IIA23 Hodgkins lymphoma C81 Beneficial Insufficient causal evidence

IIA24 Non-Hodgkins C82–C85, Beneficial Insufficient causal lymphoma C96 evidence

IIB Other neoplasms D00–D48 Detrimental (except D09.9, D37.9, D38.6, D39.9, D40.9, D41.9, 48.9)

158 Alcohol Research: Current Reviews Table 2 Chronic Diseases and Conditions for Which Alcohol Consumption Is a Component Cause, Identified by Various Meta-Analyses and Reviews and Listed in the 2005 Global Burden of Disease (GBD) Study (Continued)

No. of 2005 Disease ICD–10 Effect Level of Evidence Used if Included in the GBD Study GBD Code Meta-Analysis

IIC Diabetes E10–E13 Beneficial Causally related (Baliunas et al. 2009) (however, this depends on drinking patterns and volume of consumption)

IIE Mental and behavioral disorders

IIE1 Unipolar depressive F32–F33, Detrimental Causally related disorders F34.1

IIF Neurological conditions

IIF1 Alzheimer’s disease F01–F03, Conflicting Insufficient causal and other dementias G30–G31 evidence evidence (mainly beneficial)

IIF3 Epilepsy G40–G41 Detrimental Causally related (Samokhvalov et al. 2010a)

IIH Cardiovascular and circulatory diseases

IIH2 Hypertensive heart I11–I13 Detrimental Causally related (Taylor et al. 2010) disease (however, this depends on drinking patterns and volume of consumption)

IIH3 Ischemic heart disease I20–I25 Beneficial Causally related (Roerecke and Rehm 2010) (however, this depends on drinking patterns and volume of consumption)

IIH4 Cerebrovascular diseases

IIH4a Ischemic stroke I63–I67, Beneficial Causally related (Patra et al. 2010) I69.3 (however, this depends on drinking patterns and volume of consumption)

Chronic Diseases and Conditions Related to Alcohol Use 159 Figure 2 The relationship between increasing amounts of average daily alcohol consumption and the relative risk for cancer, with lifetime abstainers serving as the reference group.

SOURCE: Lim et al. 2012.

160 Alcohol Research: Current Reviews more recent meta-analyses have found no significant associa- ate alcohol consumption protects against the disorder by tion between alcohol consumption and the risk of bladder increasing insulin sensitivity (Hendriks 2007). Such an cancer (Pelucchi et al. 2012) or the risk of lung cancer in alcohol-related increase in insulin sensitivity has been found individuals who had never smoked (Bagnardi et al. 2001). in observational studies as well as in randomized controlled These conflicting results may stem from the studies in the trials (Davies et al. 2002; Kiechl et al. 1996; Lazarus et al. more recent meta-analyses adjusting for smoking status when 1997; Mayer et al. 1993; Sierksma et al. 2004). Alternative assessing the risk relationship between alcohol and these cancers explanations for the protective effect of moderate alcohol within individual observational studies (Bagnardi et al. 2001; consumption involve an increase in the levels of alcohol Pelucchi et al. 2012). metabolites, such as acetaldehyde and acetate (Sarkola et al. The biological pathways by which alcohol increases the 2002); an increase in high-density lipoprotein (HDL)4 risk of developing cancer depends on the targeted organ and (Rimm et al. 1999); and the anti-inflammatory effects of are not yet fully understood. Factors that seem to play a role alcohol consumption (Imhof et al. 2001). It is important to include the specific variants of alcohol-metabolizing enzymes note, however, that although there is reason to believe that (i.e., alcohol dehydrogenase [ADH], aldehyde dehydrogenase alcohol consumption is causally linked to reduced risk of [ALDH], and cytochrome P450 2E1) a person carries or type 2 diabetes, it currently is unclear whether alcohol con- the concentrations of estrogen as well as changes in folate sumption itself is a protective factor or if moderate drinking metabolism and DNA repair (Boffetta and Hashibe 2006). is a marker for healthy lifestyle choices that may account for For example, a deficiency in ALDH2 activity in people carrying some of the observed protective effect. a gene variant (i.e., allele) called ALDH2 Lys487 contributes Furthermore, the effects of alcohol consumption on risk to an elevated risk of esophageal cancer from alcohol con- of diabetes are dose dependent (see figure 3). Thus, in obser- sumption (Brooks et al. 2009). Because the ALDH2 Lys487 vational studies consumption of large amounts of alcohol allele is more prevalent in Asian populations (i.e., Japanese, has been related to an increased risk of type 2 diabetes because Chinese, and Koreans) (Eng et al. 2007), and ALDH2 is higher consumption levels may increase body weight, the hypothesized to impact the risk associated with alcohol for concentrations of certain fats (i.e., triglycerides) in the blood, all cancers, studies should account for the presence of this and blood pressure (Wannamethee and Shaper 2003; allele when assessing the risk relationship between alcohol Wannamethee et al. 2003). consumption and the development of any form of cancer. However, it is important to note that alcohol not only increases the risk of cancer but also may lower the risk of certain Neuropsychiatric Conditions types of cancer. For example, meta-analyses of observational studies have found that alcohol significantly decreases the One of the neuropsychiatric conditions associated with alcohol risk of renal cell carcinoma (Bellocco et al. 2012; Song et al. consumption is epilepsy, which is defined as an enduring 2012), Hodgkin’s lymphoma (Tramacere et al. 2012c), and predisposition for epileptic seizures and requires the occurrence non-Hodgkin’s lymphoma (Tramacere et al. 2012b). Alcohol’s of at least one seizure for a diagnosis. Alcohol consumption protective effect for renal cancer is thought to be mediated is associated with epilepsy, whereas alcohol withdrawal can by an increase in insulin sensitivity, because light to moder- cause seizures but not epilepsy (Hillbom et al. 2003).5 ate alcohol consumption has been associated with improved Observational research has found that a consistent dose- insulin sensitivity (Davies et al. 2002; Facchini et al. 1994; response relationship exists between alcohol consumption Joosten et al. 2008). Insulin resistance may play a key role in and the risk of epilepsy (see figure 3). Multiple possible path- the development of renal cancer because people with diabetes, ways may underlie this relationship. In particular, alcohol which is characterized by insulin resistance, have an increased consumption may have a kindling effect, where repeated risk of renal cancers (Joh et al. 2011; Lindblad et al. 1999). withdrawals from alcohol consumption by heavy drinkers The mechanisms underlying alcohol’s protective effect on may lower the threshold for inducing an epileptic episode the risk of developing Hodgkin’s lymphoma and non-Hodgkin’s (Ballenger and Post 1978). Alternatively, heavy alcohol lymphoma currently are unknown (Tramacere et al. 2012b, c). consumption may increase the risk of epilepsy by causing Thus, these observed protective effects should be interpreted shrinkage of brain tissue (i.e., cerebral atrophy) (Dam et al. with caution because the underlying biological mechanisms 1985), cerebrovascular infarctions, lesions, head traumas, are not understood and confounding factors and/or misclas- and changes in neurotransmitter systems and ionic balances sification of abstainers within observational studies may be responsible for these effects. 3 there are two main types of diabetes. type 1 diabetes results from the body’s failure to produce insulin, and patients therefore regularly must inject insulin. this type also is known as juvenile diabetes because of its early onset, or insulin-independent diabetes. type 2 diabetes results from insulin resis- diabetes tance, which develops when the cells fail to respond properly to insulin. it develops with age and therefore also is referred to as adult-onset diabetes. Research has found that moderate alcohol consumption is 4 associated with a reduced risk of type 2 diabetes3 (Baliunas et hDls are certain types of compounds consisting of both fat (i.e., lipid) and protein components that al. 2009). Because development of insulin resistance is key in are involved in cholesterol metabolism in the body. hDls also are referred to as “good cholesterol.” the pathogenesis of type 2 diabetes, it is thought that moder- 5 seizures are excluded from the 2005 gbD study definition of epilepsy.

Chronic Diseases and Conditions Related to Alcohol Use 161 (Barclay et al. 2008; Dam et al. 1985; Freedland and McMicken prevent ischemic events in the circulatory system (Peters et 1993; Rathlev et al. 2006). al. 2008; Tyas 2001). However, studies of these associations Another neuropsychiatric disorder considered to be have generated highly heterogeneous results, and the design causally linked to alcohol consumption is unipolar depressive and statistical analyses of these studies make it impossible to disorder. This association is supported by the temporal order rule out the potential effects of confounding factors (Panza of the two conditions, consistency of the findings, reversibility et al. 2008; Peters et al. 2008). with abstinence, biological plausibility, and the identification of a dose-response relationship. One study determined the Cardiovascular and Circulatory Diseases risk of depressive disorders to be increased two- to threefold in alcohol-dependent people (see Rehm and colleagues Alcohol consumption affects multiple aspects of the cardio- [2003a] for an examination of the causal criteria). The vascular system, with both harmful and protective effects. alcohol-attributable morbidity and mortality from unipolar These include the following (figure 4): depressive disorder currently cannot be calculated because • Increased risk of hypertension (at all consumption levels the relationship may be confounded by several factors, for men and at higher consumption levels for women); including a genetic predisposition, environmental factors (i.e., an underlying disorder or environmental exposure that • Increased risk of disorders that are caused by abnormalities may contribute to both heavy alcohol use and depressive in the generation and disruption of the electrical signals disorders), and potential self-medication with alcohol by that coordinate the heart beat (i.e., conduction disorders individuals with unipolar depressive disorders (Grant and and other dysrhythmias); Pickering 1997; Rehm et al. 2004). Research findings sug- gest that all of these pathways may play a role. The pathways • Increased risk of cardiovascular disease, such as stroked for the association between alcohol and unipolar depressive caused by blockage of blood vessels in the brain (i.e., disorder in which alcohol does not play a causal role only ischemic stroke) (at a higher volume of consumption) or affect the measurement of the alcohol-based RR for unipolar rupture of blood vessels (i.e., hemorrhagic stroke); and depressive disorder; however, they do not contradict the notion that alcohol is causally related to the development of • Protective effects (at lower levels of consumption) against unipolar depressive disorder via other pathways. This conclusion hypertension in women and against ischemic heart results from the observation that depressive symptoms increase disease and ischemic stroke in both men and women. markedly during heavy-drinking occasions and disappear or lessen during periods of abstinence (Rehm et al. 2003a). Numerous studies also have examined the association between alcohol and Alzheimer’s disease and vascular dementia.6 6 Vascular dementia, the second most common form of dementia after Alzheimer’s disease, is caused These analyses generally have determined a beneficial effect by problems in the supply of blood to the brain. Its most common symptoms include problems with thinking, concentration, and communication; depression and anxiety; physical weakness or paralysis; of alcohol, which has been attributed to alcohol’s ability to memory problems; seizures; and periods of severe confusion.

Figure 3 The relationship between increasing amounts of average daily alcohol consumption and the relative risk for diabetes and epilepsy, with lifetime abstainers serving as the reference group.

SOURCE: Lim et al. 2012.

162 Alcohol Research: Current Reviews Figure 4 The relationship between increasing amounts of average daily alcohol consumption and the relative risk for cardiovascular diseases (i.e., hypertension, conduction disorders, and ischemic and hemorrhagic stroke), with lifetime abstainers serving as the reference group. For both hypertension and hemorrhagic and ischemic stroke, the relationship differs between men and women. Moreover, for both ischemic and hemorrhagic stroke, the influence of alcohol consumption on mortality is much greater than the influence on morbidity, at least in women. In men, no such difference appears to exist.

SOURCE: Lim et al. 2012.

Chronic Diseases and Conditions Related to Alcohol Use 163 The specific biological pathways through which alcohol Alcohol interacts with the ischemic system to decrease consumption interacts with the cardiovascular system are not the risk of ischemic stroke and ischemic heart disease at low always clear, but several mechanisms have been identified levels of consumption; however, this protective effect is not that may play a role. These include increased blood concen- observed at higher levels of consumption. As mentioned trations of HDLs, effects on cellular signaling, decreased above, alcohol exerts these effects mainly by increasing levels blood clot formation by platelets, and increased blood clot of HDL, preventing blood clots, and increasing the rate of dissolution through enzyme action (Zakhari 1997). More breakdown of blood clots. However, binge drinking, even indirect effects also may play a role. For example, alcohol by light to moderate drinkers, leads to an increased risk of may increase the risk of hypertension by enhancing the activity ischemic events by increasing the probability of clotting and of the sympathetic nervous system, which results in greater abnormal contractions of the heart chambers (i.e., ventricular constriction of the blood vessels and makes the heart contract fibrillation). As with hemorrhagic stroke, alcohol has different more strongly. In addition, alcohol possibly decreases the effects on morbidity than on mortality related to ischemic sensitivity of the body’s internal blood pressure sensors (i.e., events (see figure 5). Thus, meta-analyses of alcohol con- baroreceptors), thereby diminishing its ability to regulate sumption and the risk of ischemic heart disease (Roerecke blood pressure. and Rehm 2012) and ischemic stroke (Taylor et al. 2009) Alcohol’s protective effects against the risk of ischemic found a larger protective effect for morbidity than for mor- heart disease as well as against hypertension in women is tality related to these conditions. One possible explanation hypothesized to result from its ability to increase HDL levels for this observation, in addition to those listed above for and/or reduce platelet aggregation on arterial walls. Differences hemorrhagic stroke, is that patients in the morbidity studies in the effects of alcohol in men and women may stem from may be younger at the time of the stroke than those in mor- differing drinking patterns, with men more likely to engage tality studies. Despite the increased risk for ischemic heart in binge drinking, even at low average levels of consumption. disease at higher levels of alcohol consumption noted in These heavy-drinking occasions may lead to an increased observational studies (see Roerecke and Rehm 2012 for the risk of hypertension for men compared with women at similar most up-to-date meta-analysis), there was not enough evidence alcohol consumption levels (Rehm et al. 2003b). for a detrimental effect of alcohol consumption on ischemic Alcohol’s effect on hypertension also contributes to the heart disease for it to be modeled in the 2005 GBD study. risk of hemorrhagic stroke (Taylor et al. 2009), with a Moreover, the observational studies investigating the hypothesized dose-response effect. The mortality and morbidity link between alcohol consumption and ischemic events had from alcohol-attributable hemorrhagic stroke differ by sex several methodological flaws, and the RR functions for (see figure 4). As with hypertension, differences in drinking ischemic events, especially ischemic heart disease, therefore pattern between men and women most likely are responsible are not well defined. A meta-analysis conducted by Roerecke for the differing RR functions for hemorrhagic stroke by sex. and Rehm (2012) observed a substantial degree of hetero- Three possible explanations have been put forth to explain geneity among all consumption levels, pointing to a possible the effects of drinking pattern on RR: confounding effect of heavy drinking. In addition, previous observational studies have been limited by the inclusion of • Heavy drinkers also may have other comorbidities that “sick quitters” in the reference groups, who have an increased may increase the probability of a fatal hemorrhagic stroke. risk of ischemic events compared with lifetime abstainers.

• Alcohol consumption may worsen the disease course digestive diseases through biological mechanisms and by decreasing compliance with medication regimens. Alcohol is associated with various liver diseases and is most strongly related to fatty liver, alcoholic hepatitis, and cirrhosis. • Alcohol’s effects on morbidity may be underestimated The association between the risk of liver cirrhosis and alcohol because of a stigmatization of heavy alcohol consumption consumption has long been recognized (see figure 6). The in women, thereby potentially decreasing the probability main biological mechanism contributing to this liver damage that female heavy drinkers will be treated for stroke. likely involves the breakdown of ethanol in the liver through oxidative and nonoxidative pathways that result in the pro- Large cohort studies and meta-analyses have shown that duction of free radicals, acetaldehyde, and fatty acid ethyl alcohol consumption leads to an increase in the risk for con- esters, which then damage liver cells (Tuma and Casey 2003). duction disorders and dysrhythmias (Samokhvalov et al. Given the same amount of alcohol consumption, alcohol 2010b). These effects are caused by changes in the electrical increases the risk of mortality from liver cirrhosis more activity of the heart, including direct toxic effects of alcohol steeply than the risk of morbidity because it worsens the on the heart (i.e., cardiotoxicity), excessive activity of the course of liver disease and has a detrimental effect on the sympathetic nervous system (i.e., hyperadrenergic activity) immune system (Rehm et al. 2010c). during drinking and withdrawal, impairment of the Alcohol consumption also has been linked to an increase parasympathetic nervous system (i.e., of vagal tone), and in the risk for acute and chronic pancreatitis. Specifically, increase of intra-atrial conduction time (Balbão et al. 2009). heavy alcohol consumption (i.e., more than, on average,

164 Alcohol Research: Current Reviews Figure 5 The relationship between increasing amounts of average daily alcohol consumption and the relative risk for ischemic heart disease, with lifetime abstainers serving as the reference group. Low to moderate alcohol consumption has a beneficial effect on both mortality and morbidity from ischemic heart disease. However, the specific effects depend on both the gender and the age of the drinker, with the greatest beneficial effects of low-to-moderate consumption seen on morbidity from ischemic heart disease in women ages 15 to 34.

SOURCE: Lim et al. 2012.

Chronic Diseases and Conditions Related to Alcohol Use 165 48 grams pure ethanol, or about two standard drinks, per activators, such as cyclin D1 and keratinocyte growth factor, day) leads to a noticeably elevated risk of pancreatitis, that could lead to excessive multiplication of skin cells (i.e., whereas consumption below 48 grams per day is associated epidermal hyperproliferation). Finally, alcohol may exacer- with a small increase in risk of pancreatitis (see figure 6). bate disease progression by interfering with compliance with Higher levels of alcohol consumption may affect the risk of treatment regimens (Gupta et al. 1993; Zaghloul and pancreatitis through the same pathways that cause liver dam- Goodfield 2004). age, namely the formation of free radicals, acetaldehyde, and fatty acid ethyl esters during the metabolism of alcohol in Alcohol’s Effects on Other Medication Regimens damaged pancreatic acinar cells (Vonlaufen et al. 2007). Alcohol can affect cognitive capacity, leading to impaired judgment and a decreasing ability to remember important Psoriasis information, including when to take medications for other Psoriasis is a chronic inflammatory skin disease caused by conditions (Braithwaite et al. 2008; Hendershot et al. 2009; the body’s own immune system attacking certain cells in Parsons et al. 2008). Although the relationship between the body (i.e., an autoimmune reaction). Although there is alcohol consumption and adherence to treatment regimens insufficient biological evidence to indicate that alcohol is mainly has been studied in regards to adherence to HIV causally linked with psoriasis, many observational studies antiretroviral treatment (Braithwaite and Bryant 2010; have determined a detrimental impact of drinking on Hendershot et al. 2009; Neuman et al. 2012), research also psoriasis, especially in male patients. Alcohol is hypothesized has shown that alcohol consumption and alcohol misuse to induce immune dysfunction that results in relative impact adherence to medications for other chronic diseases, immunosuppression. In addition, alcohol may increase with significant or almost-significant effects (Bates et al. the production of inflammatory cytokines and cell cycle 2010; Bryson et al. 2008; Coldham et al. 2002; Verdoux et

Figure 6 The relationship between increasing amounts of average daily alcohol consumption and the relative risk for digestive diseases (i.e., liver cirrhosis and pancreatitis), with lifetime abstainers serving as the reference group. For liver cirrhosis, alcohol’s effects on mortality are greater than those on morbidity, and slight differences exist between the effects in men and women.

SOURCE: Lim et al. 2012.

166 Alcohol Research: Current Reviews al. 2000). Thus, for diseases or conditions managed by phar- Estimating the Alcohol-Attributable Fractions of macotherapy, alcohol consumption likely is associated with Chronic Diseases and Conditions increased morbidity and even mortality (if nonadherence to the medication could be fatal) if drinking results in nonad- When assessing the risk of chronic diseases and conditions herence to medication regimens. that are related to alcohol consumption in some, but not all, cases, one of the variables frequently analyzed is the alcohol- Impact of Sex, Race, and Age on the Association attributable fraction (AAF)—that is, the proportion of cases of Alcohol Consumption with Chronic diseases that can be attributed to the patient’s alcohol consumption. Because alcohol consumption can be modeled using a con- Given the same amount of alcohol consumed, men and tinuous distribution (Kehoe et al. 2012; Rehm et al. 2010b), women can have differing morbidity and mortality from the calculation of the alcohol-attributable burden of disease alcohol-related chronic disease and conditions. These differ- uses a continuous RR function.7 Thus, the AAFs for chronic ences may be related to the pharmacokinetics of alcohol in diseases and conditions can be calculated using the following men and women. Women generally have a lower body water formula: content than men with the same body weight, causing women to reach higher blood alcohol concentrations than men after drinking an equivalent amount of alcohol (Frezza et al. 1990; Taylor et al. 1996). Moreover, women appear to eliminate alcohol from the blood faster than do men, possibly In this formula, Pabs represents the prevalence of the disease because they have a higher liver volume per unit body mass among lifetime abstainers, P is the prevalence among former (Kwo et al. 1998; Lieber 2000). In addition to these phar- form drinkers, RRform is the relative risk for former drinkers, P(x) is macokinetic factors, hormonal differences also may play a the prevalence among current drinkers with an average daily role because at least in the case of liver disease, alcohol- alcohol consumption of x, and RR(x) is the relative risk for attributable harm is modified by estrogen. However, hormonal current drinkers with an average daily alcohol consumption influences on alcohol-related risks are not yet fully understood of x. These AAFs vary greatly depending on alcohol exposure (Eagon 2010). levels. (For examples of AAFs and information on the calcu- As noted previously, a deficiency of the ALDH2 enzyme lation of the 95 percent confidence intervals for chronic in people carrying the ALDH2 Lys487 allele contributes to diseases and conditions see Gmel and colleagues [2011]). an elevated risk of cancer from alcohol consumption. Because alcohol metabolism also plays a role in many other chronic diseases, the ALDH2 Lys487 allele also may increase the risk Limitations of RR Functions for Chronic Diseases for digestive diseases. The heterogeneity of risk caused by and Conditions this allele, which is more prevalent in Asian populations, The chronic disease RR functions outlined in figures 2 to 6 may lead to incorrect measurements of the risk for cancer are derived from the most up-to-date and rigorous meta- and digestive disease outcomes in countries with a small analyses in which the risk of a disease (i.e., for mortality and, Asian population, and will lead to incorrect results if the RRs where possible, morbidity) was provided by alcohol con- from these countries are applied to Asian populations (Lewis sumption as a continuous function (for more details on the and Smith 2005; Oze et al. 2011). meta-analysis methods used in each study, see the original Because the pathology of alcohol-related ischemic heart articles cited in table 2). However, the RR functions and the disease is affected by the age of the drinker (Lazebnik et al. relationship between alcohol consumption and the risk of 2011), differences also may exist in the risk of ischemic heart chronic diseases and conditions are biased by multiple factors. disease in different age groups. Preliminary research assessing First, the RRs can be limited by poor measurement of alcohol this issue across multiple studies has found that the association exposure, outcomes, and confounders. Research on alcohol between alcohol consumption and the resulting risk for ischemic consumption patterns and disease is scarce, and only few heart disease does indeed differ by age (see figure 5). However, studies have investigated the effects of drinking patterns on no meta-analyses to date have investigated the effects of alco- chronic diseases and conditions. Thus, the chronic disease hol consumption on the risk of morbidity and mortality in and condition RRs presented in this article may be confounded by drinking patterns, which are correlated to overall volume different age groups for other chronic diseases and condi- of alcohol consumption. Additionally, the volume of alcohol tions. Accordingly, research is needed to assess if the varying consumed generally is poorly measured, with many medical relationship between alcohol consumption and ischemic epidemiology studies measuring alcohol consumption only heart disease in different age groups results from biological at baseline. As a result, these analyses do not include measures differences in pathology or from differences in drinking pat- terns. Additionally, research is needed to assess if age modi- 7 the exception to this approach is tuberculosis because only data on categorical alcohol exposure fies the risk relationships between alcohol and other diseases. risks are available.

Chronic Diseases and Conditions Related to Alcohol Use 167 of the volume of alcohol consumed during the medically relevant Limitations of AAFs for Chronic Diseases and time period, which may encompass several years. For example, Conditions in the case of cancer, the cumulative effects of alcohol may take many years before an outcome is observed. Likewise, In most studies assessing AAFs for chronic diseases and con- many of the larger cohort studies only use single-item, semi- ditions, the AAF for an outcome is calculated as if the health quantitative food questionnaires that measure either frequency consequences of alcohol consumption are immediate. or volume of consumption. Indeed, for most chronic diseases and conditions, including Second, medical epidemiology studies typically suffer even diseases such as cirrhosis, a large degree of the effects from poorly defined reference groups (Rehm et al. 2008). caused by changes in alcohol consumption can be seen Thus, such studies typically only measure alcohol consump- immediately at the population level (Holmes et al. 2011; tion at one point or during one time period in a participant’s Leon et al. 1997; Zatonski et al. 2010; for a general discus- life and classify, for example, all people who do not consume sion see Norström and Skog 2001; Skog 1988). For cancer, alcohol during the reference period as abstainers, even though however, the situation is different. The effects of alcohol it is essential to separate ex-drinkers, lifetime abstainers, and consumption on the risk of cancer only can be seen after years, very light drinkers. As a result, these measurements of alcohol and often as long as two decades. Nevertheless, for the pur- consumption may lead to incorrect risk estimates because pose of illustrating the entire alcohol-attributable burden of the groups of nondrinkers in these studies have heterogeneous disease it is important to include cancer deaths, because they risks for diseases (Shaper and Wannamethee 1998). The account for a substantial burden. For example, a recent large potential significance of this issue is underscored by previous study found that in Europe 1 in 10 cancers in men and 1 research indicating that more than 50 percent of those par- in 33 cancers in women were alcohol related (Schütze et al. ticipants who identified themselves as lifetime abstainers in 2011). Therefore, in the interpretation of alcohol’s effect on medical epidemiology studies also had reported lifetime mortality and burden of disease in this article, the assump- drinking in previous surveys (Rehm et al. 2008). tion that there has been uniform exposure to alcohol for at Third, chronic disease and condition outcomes in medical least the previous two decades must be kept in mind. epidemiology studies also frequently are poorly measured, Another limitation to calculating the burden of chronic most often by means of self-reporting. Additionally, other diseases and conditions attributable to alcohol consumption confounding factors, such as relevant, non-substance use–related is the use of mainly unadjusted RRs to determine the AAFs. confounders, often are not controlled for. The RR formulas were developed for risks and were adjusted Fourth, RR estimates for chronic diseases and conditions only for age (see Flegal et al. 2006; Korn and Graubard resulting from alcohol consumption frequently are hampered 1999; Rockhill and Newman 1998), although many other by weak study designs that base estimates of alcohol-related socio-demographic factors are linked with both alcohol con- risks on nonexperimental designs (i.e., case-control and sumption and alcohol-related harms (see figure 1). However, cohort studies). These study designs are limited by factors two arguments can be made to justify the use of mainly that cannot be controlled for and which may lead to incor- unadjusted RR formulas in the 2005 GBD study. First, in rect results. For example, experimental studies on the effects risk analysis studies (Ezzati et al. 2004) almost all of the of antioxidants have failed to confirm the protective effects underlying studies of the different risk factors only report unadjusted risks. Relying on adjusted risks would severely bias of such agents found in observational studies (Bjelakovic et the estimated risk functions because only a small proportion al. 2008). Furthermore, the sampling methodology of many of generally older studies could be included. Second, most of of the cohort studies that were used in the meta-analyses for the analyses of alcohol and the risk of chronic diseases and the above-presented RRs is problematic, especially when conditions show no marked differences after adjustment (see studying the effects of alcohol consumption. Many of the Rehm et al. 2010b). However, the need for adjustment to cohorts in these studies were from high-income countries the RRs may change when other dimensions of alcohol con- and were chosen based on maximizing follow-up rates. Although sumption, such as irregular heavy-drinking occasions, are the chosen cohorts exhibited variation in average daily alcohol considered with respect to ischemic heart disease. consumption, little variation was observed in drinking patterns and other potential moderating lifestyle factors. The overall effect of these limitations on the RRs and Conclusions AAFs, and on the estimated burden of mortality and morbidity calculated using these RRs, currently is unclear. In order to There are limitations to the current ability to estimate the investigate the effect of these biases, studies should be under- burden of chronic diseases and conditions attributable to taken that combine better exposure measures of alcohol alcohol consumption. The comparative risk assessment consumption with state-of-the-art outcome measures in study within the GBD study only can determine this burden countries at all levels of economic development. These studies based on current knowledge of alcohol consumption and are important, not only for understanding the etiology of risk and mortality patterns at a global level. More detailed, alcohol-related chronic diseases and conditions, but also for country-specific estimates often are limited by the validity of formulating prevention measures (Stockwell et al. 1997). the available consumption and mortality data. As more stud-

168 Alcohol Research: Current Reviews ies are published, it is likely that new confounders will be bElloCCo, R.; PasQuali, E.; Rota m.; Et al. alcohol drinking and risk of renal cell carcino- ma: Results of a meta-analysis. Annals of Oncology Epub ahead of print, 2012. PmiD: discovered for some of the relationships between alcohol con- 22398178 sumption and various chronic diseases and conditions. The bJElaKoViC, g.; niKoloVa, D.; gluuD, l.l.; Et al. antioxidant supplements for prevention of results from such new studies then may be used in meta-analyses mortality in healthy participants and patients with various diseases. Cochrane Database of the effect of alcohol in diseases where alcohol only plays a of Systematic Reviews 16(2):CD007176, 2008. PmiD: 18425980 small role, such as bladder, endometrial, and ovarian cancer. boFFEtta, P., anD hashibE, m. alcohol and cancer. Lancet Oncology 7(2):149–156, 2006. New studies also may lead to the recognition of a causal link PmiD: 16455479 between alcohol consumption and other diseases. Further- bRaithWaitE, R.s., anD bRyant, K.J. influence of alcohol consumption on adherence to and more, new confounders and new studies may disprove the toxicity of antiretroviral therapy and survival. Alcohol Research & Health 33(3):280– relationship between alcohol consumption and certain dis- 287, 2010.

eases that currently are considered to be causally linked. bRaithWaitE, R.s.; ConigliaRo, J.; mCginnis, K.a.; Et al. adjusting alcohol quantity for mean Although there are limitations to the current methodology consumption and intoxication threshold improves prediction of nonadherence in hiV used to estimate the alcohol-attributable burden of chronic patients and hiV-negative controls. Alcoholism: Clinical and Experimental Research diseases and conditions, the limitations discussed in this article 32(9):1645–1651, 2008. PmiD: 18616666 do not affect the overall conclusion that alcohol consump- bRooKs, P.J.; EnoCh, m.a.; golDman, D.; Et al. the alcohol flushing response: an unrecog- tion is related to a considerable number of chronic diseases nized risk factor for esophageal cancer from alcohol consumption. PLOS Medicine and conditions and contributes to a substantial amount of 6(3):258–263, 2009. PmiD: 19320537 the global burden of chronic diseases and conditions. Therefore, bRyson, C.l.; au, D.h.; sun, h.; Et al. alcohol screening scores and medication nonad- alcohol consumption should be considered in developing herence. Annals of Internal Medicine 149(11):795–804, 2008. PmiD: 19047026 intervention strategies aimed at reducing the burden of ColDham, E.l.; aDDington, J.; anD aDDington, D. medication adherence of individuals with chronic diseases and conditions. a first episode of psychosis. Acta Psychiatrica Scandinavica 106(4):286–290, 2002. ■ PmiD: 12225495

CoRRao, g.; bagnaRDi, V.; zambon, a.; anD la VECChia, C. a meta-analysis of alcohol con- sumption and the risk of 15 diseases. Preventive Medicine 38(5):613–619, 2004. Acknowledgements PmiD: 15066364 Dr. Jürgen Rehm received a salary and infrastructure support Dam, a.m.; Fuglsang-FREDERiKsEn, a; sVaRRE-olsEn, u.; anD Dam, m. late-onset epilepsy: Etiologies, types of seizure, and value of clinical investigation, EEg, and computerized from the Ontario Ministry of Health and Long-Term Care tomography scan. Epilepsia 26(3):227 –231, 1985. PmiD: 4006882 and Dr. Charles Parry received the same from the South African Medical Research Council. DaViEs, m.J.; baER, D.J.; JuDD, J.t.; Et al. Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. JAMA: The Journal of the American Medical Association 287(19):2559–2562, 2002. PmiD: 12020337

Financial Disclosure DEnnis, l.K. meta-analysis for combining relative risks of alcohol consumption and prostate cancer. Prostate 42(1):56–66, 2000. PmiD: 10579799 The authors declare that they have no competing financial interests. Eagon, P.K. alcoholic liver injury: influence of gender and hormones. World Journal of Gastroenterology 16(11):1377–1384, 2010. PmiD: 20238405

Eng, m.y.; luCzaK, s.E.; anD Wall, t.l. alDh2, aDh1b, and aDh1C genotypes in asians: a References literature review. Alcohol Research & Health. 30(1):22–27, 2007. PmiD: 17718397 Ezzati, m.; loPEz, a.; RoDgERs, a.; Et al. Comparative Quantification of Health Risks. bagnaRDi, V.; blangiaRDo, m.; la VECChia, C.; anD CoRRao, g. alcohol consumption and the Global and Regional Burden of Disease Attributable to Selected Major Risk Factors. risk of cancer: a meta-analysis. Alcohol Research & Health 25(4):263–270, 2001. geneva, switzerland: World health organization, 2004. PmiD: 11910703 FaCChini, F.; ChEn, y.D.; anD REaVEn g.m. light-to-moderate alcohol intake is associated balbão, C.E.b.; DE Paola, a.a.V.; anD FEnElon, g. Effects of alcohol on atrial fibrillation: with enhanced insulin sensitivity. Diabetes Care 17(2):115–119, 1994. PmiD: 7907975 myths and truths. Therapeutic Advances in Cardiovascular Disease 3(1):53–63, 2009. FEDiRKo, V.; tRamaCERE, i.; bagnaRDi, V.; Et al. alcohol drinking and colorectal cancer risk: PmiD: 19124390 an overall and dose-response meta-analysis of published studies. Annals of Oncology baliunas, D.o.; tayloR, b.J.; iRVing, h.; Et al. alcohol as a risk factor for type 2 diabetes: a 22(9):1958–1972, 2011. PmiD: 21307158 systematic review and meta-analysis. Diabetes Care 32(11):2123–2132, 2009. PmiD: FillmoRE, K.m.; ChiKRitzhs, t.; stoCKWEll, t.; Et al. alcohol use and prostate cancer: a 19875607 meta-analysis. Molecular Nutrition & Food Research 53(2):240–255, 2009. PmiD: ballEngER, J.C., anD Post, R.m. Kindling as a model for alcohol withdrawal syndromes. 19156715

British Journal of Psychiatry 133:1–14, 1978. PmiD: 352467 FlEgal, K.m.; Williamson, D.F.; anD gRaubaRD, b.i. using adjusted relative risks to calculate attributable fractions. American Journal of Public Health 96(3):398, 2006. PmiD: 16449574 baRClay, g.a.; baRVouR, J.; stEWaRt, s.; Et al. adverse physical effects of alcohol misuse. Advances in Psychiatric Treatment 14:139–151, 2008. FREEDlanD, E.s., anD mCmiCKEn, D.b. alcohol-related seizures, Part i: Pathophysiology, dif- ferential diagnosis, and evaluation. Journal of Emergency Medicine 11(4):463–473, batEs, J.a.; WhitEhEaD, R.; bolgE, s.C.; anD Kim, E. Correlates of medication adherence 1993. PmiD: 8228111 among patients with bipolar disorder: Results of the bipolar evaluation of satisfaction and tolerability (bEst) study: a nationwide cross-sectional survey. Primary Care FREzza, m.; Di PaDoVa, C.; Pozzato, g.; Et al. high blood alcohol levels in women. the role Companion to the Journal of Clinical Psychiatry 12(5):PCC 09m00883, 2010. PmiD: of decreased gastric alcohol dehydrogenase activity and first-pass metabolism. New 21274363 England Journal of Medicine 322(2):95–99, 1990. PmiD: 2248624

Chronic Diseases and Conditions Related to Alcohol Use 169 gmEl, g.; shiElD, K.D.; FRiCK, h.; Et al. Estimating uncertainty of alcohol-attributable frac- lazaRus, R.; sPaRRoW, D.; anD WEiss, s.t. alcohol intake and insulin levels. the normative tions for infectious and chronic diseases. BMC Medical Research Methodology 11:48, aging study. American Journal of Epidemiology 145(10):909–916, 1997. PmiD: 9149662 2011. PmiD: 21496313 lazEbniK, l.b.; KomissaREnKo, i.a.; anD miKhEEVa, o.m. Cardiovascular pathology associated gRant, b.F., anD PiCKERing, R.P. Familial aggregation of Dsm–iV alcohol disorders: with digestive system diseases. Experimental & Clinical Gastroenterology 5:69–74, Examination of the primary-secondary distinction in a general population sample. 2011.(article in Russian). PmiD: 21919242 Journal of Nervous and Mental Disease 185(5):335–343, 1997. PmiD: 9171811 lEon, D.a.; ChEnEt, l.; shKolniKoV, V.m.; Et al. huge variation in Russian mortality rates guPta, m.a.; sChoRK, n.J.; guPta, a.K.; anD Ellis, C.n. alcohol intake and treatment 1984-1994: artefact, alcohol, or what? Lancet 350(9075):383–388, 1997. PmiD: responsiveness of psoriasis: a prospective study. Journal of the American Academy of 9259651 Dermatology 28(5 Pt 1):730–732, 1993. PmiD: 8496416 lEWis, s.J., anD smith, D.g. alcohol, alDh2, and esophageal cancer: a meta-analysis which hEnDERshot, C.s.; stonER, s.a.; PantalonE, D.W.; anD simoni, J.m. alcohol use and illustrates the potentials and limitations of a mendelian randomization approach. Cancer antiretroviral adherence: Review and meta-analysis. Journal of Acquired Immune Epidemiology, Biomarkers & Prevention 14(8):1967–1971, 2005. PmiD: 16103445 Deficiency Syndromes 52(2):180–202, 2009. PmiD: 19668086 liEbER, C.s. Ethnic and gender differences in ethanol metabolism. Alcoholism: Clinical hEnDRiKs, h.F.J. moderate alcohol consumption and insulin sensitivity: observations and and Experimental Research 24(4):417–418, 2000. PmiD: 10798572 possible mechanisms. Annals of Epidemiology 17(suppl.5):s40–s42, 2007. lim, s.s.; Vos, t.; FlaXman, a.D.; Et al. a comparative risk assessment of burden of dis- hillbom, m.; PiEninKERoinEn, i.; anD lEonE, m. seizures in alcohol-dependent patients: ease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, Epidemiology, pathophysiology and management. CNS Drugs 17(14):1013–1030, 1990-2010: a systematic analysis for the global burden of Disease study 2010. Lancet 2003. PmiD: 14594442 380(9859):2224–2260, 2012. PmiD: 23245609

olmEs ooth EiER Et al h , J.; b , a.; m , P.; . Time Lags in the Effects of Alcohol Policy linDblaD, P.; ChoW, W.h.; Chan, J.; Et al. the role of diabetes mellitus in the aetiology of Interventions. melbourne: international Kbs Conference; 2011. renal cell cancer. Diabetologia 42(1):107–112, 1999. PmiD: 10027588 imhoF, a..; FRoEhliCh, m..; bREnnER, h.; Et al. Effect of alcohol consumption on systemic mayER, E.J.; nEWman, b.; QuEsEnbERRy, C.P., JR.; Et al. alcohol consumption and insulin markers of inflammation. Lancet 357(9258):763–767, 2001. PmiD: 11253971 concentrations. Role of insulin in associations of alcohol intake with high-density international agency for Research on Cancer (iaRC). International Agency for Research lipoprotein cholesterol and triglycerides. Circulation 88(5 Pt 1):2190–2197, 1993. on Cancer Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume PmiD: 8222114 96. Alcoholic Beverage Consumption and Ethyl Carbamate (Urethane). lyon, France: mCgoVERn, P. Uncorking the Past: The Quest for Wine, Beer, and Other Alcoholic Beverages. iaRC; 2008. berkley; los angeles; london: the Regents of the university of California, 2009. iRVing, h.m.; samoKhValoV, a.V.; anD REhm, J. alcohol as a risk factor for pancreatitis. a nEuman, m.g.; sChnEiDER, m.; nanau, R.m.; anD PaRRy, C. alcohol consumption, progres- systematic review and meta-analysis. JOP: Journal of the Pancreas 10(4):387–392, sion of disease and other comorbidities, and response to antiretroviral medication in 2009. PmiD: 19581740 people living with hiV. AIDS Research and Treatment 2012:751827, 2012. PmiD: 22496971 islami, F.; FEDiRKo, V.; tRamaCERE, i.; Et al. alcohol drinking and esophageal squamous cell noRstRöm, t., anD sKog, o.J. alcohol and mortality: methodological and analytical issues carcinoma with focus on light-drinkers and never-smokers: a systematic review and in aggregate analyses. Addiction 96(suppl.1):s5–s17, 2001. PmiD: 11228078 meta-analysis. International Journal of Cancer 129(10):2473–2484, 2011. PmiD: 21190191 ozE, i.; matsuo, K.; WaKai, K.; Et al. alcohol drinking and esophageal cancer risk: an eval- uation based on a systematic review of epidemiologic evidence among the Japanese pop- islami, F.; tRamaCERE, i.; Rota, m.; Et al. alcohol drinking and laryngeal cancer: overall ulation. Japanese Journal of Clinical Oncology 41(5):677–692, 2011. PmiD: 21430021 and dose-risk relation—a systematic review and meta-analysis. Oral Oncology 46(11):802–810, 2010. PmiD: 20833578 Panza, F.; CaPuRso, C.; D’intRono, a.; Et al. Vascular risk factors, alcohol intake, and cog- nitive decline. The Journal of Nutrition, Health & Aging 12(6):376–381, 2008. PmiD: Joh, h.K.; WillEt, W.C.; anD Cho, E. type 2 diabetes and the risk of renal cell cancer in women. Diabetes Care 34(7):1552–1556, 2011. PmiD: 21602426 18548174 PaRRy, C.; PatRa, J.; anD REhm J. alcohol consumption and non-communicable diseases: JoostEn, m.m.; bEulEns, J.W.; KERstEn, s.; anD hEnDRiKs, h.F. moderate alcohol consump- tion increases insulin sensitivity and aDiPoQ expression in postmenopausal women: a Epidemiology and policy implications. Addiction 106(10):1718–1724, 2011. PmiD: randomised, crossover trial. Diabetologia 51(8):1375–1381, 2008. PmiD: 18504547 21819471

aRsons osoF anD ustansK KEhoE, t,; gmEl, g.J.; shiElD, K.D.; Et al. Determining the best population-level alcohol P , J.t.; R , E.; m , b. the temporal relationship between alcohol consumption model and its impact on estimates of alcohol-attributable harms. consumption and hiV-medication adherence: a multilevel model of direct and moderat- Population Health Metrics 10:6, 2012. PmiD: 22490226 ing effects. Health Psychology 27(5):628–637, 2008. PmiD: 18823189

KiEChl, s.; WillEit, J.; PoEWE, W.; Et al. insulin sensitivity and regular alcohol consump- PatRa J.; tayloR, b.; iRVing, h.; Et al. alcohol consumption and the risk of morbidity and tion: large, prospective, cross sectional population study (bruneck study). BMJ mortality for different stroke types—a systematic review and meta-analysis. BMC Public 313(7064):1040–1044, 1996. PmiD: 8898593 Health 10:258, 2010. PmiD: 20482788

KoRn, E.l., anD gRaubaRD, b.i. Analysis of Health Surveys. new york: John Wiley & sons PEluCChi, C.; galEonE, C.; tRamaCERE, i.; Et al. alcohol drinking and bladder cancer risk: a inc., 1999. meta-analysis. Annals of Oncology 23(6):1586-1593, 2012. PmiD: 22039083

KWo, P.y.; PamChanDani, V.a.; o’ConnoR, s.; Et al. gender differences in alcohol PEtERs, R.; PEtERs, J.; WaRnER, J.; Et al. alcohol, dementia and cognitive decline in the metabolism: Relationship to liver volume and effect of adjusting for body mass. elderly: a systematic review. Age and Ageing 37(5):505–512, 2008. PmiD: 18487267 Gastroenterology 115(6):1552–1557, 1998. PmiD: 9834284 RathlEV, n.K.; ulRiCh, a.s.; DElanty, n.; anD D’onoFRio, g. alcohol-related seizures. The laChEnmEiER, D.W., anD REhm, J. unrecorded alcohol: a threat to public health? Addiction Journal of Emergency Medicine 31(2):157–163, 2006. PmiD: 17044577 104(6):875–877, 2009. PmiD: 19466914 REhm, J.; baliunas, D.; boRgEs, g.l.; Et al. the relation between different dimensions of laChEnmEiER, D.W.; REhm, J.; anD gmEl, g. surrogate alcohol: What do we know and alcohol consumption and burden of disease: an overview. Addiction 105(5):817–843, where do we go? Alcoholism: Clinical and Experimental Research 31(10):1613–1624, 2010a. PmiD: 20331573 2007. PmiD: 17681034 REhm, J.; iRVing, h.; yE, y.; Et al. are lifetime abstainers the best control group in alcohol lamy, l. Étude clinique et statistique de 134 cas de cancer de l’oesophage et du cardia. epidemiology? on the stability and validity of reported lifetime abstention. American Archives des Maladies de L’Appareil Digestif 4:451–475, 1910. Journal of Epidemiology 168(8):866–871, 2008. PmiD: 18701442

170 Alcohol Research: Current Reviews REhm, J.; KEhoE, t.; gmEl, g.; Et al. statistical modeling of volume of alcohol exposure for stoCKWEll, t.; singlE, E.; haWKs, D.; Et al. sharpening the focus of alcohol policy from epidemiological studies of population health: the us example. Population Health aggregate consumption to harm and risk reduction. Addiction Research 5(1): 1–9, 1997. Metrics 8:3, 2010b. PmiD: 20202213 available at http://www.ncbi.nlm.nih.gov/ tayloR, b.; iRVing, h.m.; baliunas, D.; Et al. alcohol and hypertension: gender differences pmc/articles/PmC2841092 in dose-response relationships determined through systematic review and meta-analy- REhm, J.; mathERs, C.; PoPoVa, s.; Et al. global burden of disease and injury and econom- sis. Addiction 104(12):1981–1990, 2009. PmiD: 19804464 ic cost attributable to alcohol use and alcohol use disorders. Lancet 373(9682):2223– 2233, 2009. PmiD: 19560604 tayloR, b.; iRVing, h.m.; KantEREs, F.; Et al. the more you drink, the harder you fall: a sys- tematic review and meta-analysis of how acute alcohol consumption and injury or colli- REhm, J.; Room, R.; gRaham, K.; Et al. the relationship of average volume of alcohol sion risk increase together. Drug and Alcohol Dependence 110(1–2):108–116, 2010. consumption and patterns of drinking to burden of disease: an overview. Addiction PmiD: 20236774 98(9):1209–1228, 2003a. PmiD: 12930209 tayloR, J.l.; DolhERt, n.; FRiEDman, l.; Et al. alcohol elimination and simulator perfor- REhm, J.; Room, R.; montEiRo, m.; Et al. alcohol use. in: Ezzati, m.; lopez, a.D.; Rodgers, mance of male and female aviators: a preliminary report. Aviation, Space, and a.; et al., Eds. Comparative Quantification of Health Risks: Global and Regional Burden Environmental Medicine 67(5):407–413, 1996. PmiD: 8725465 of Disease Attributable to Selected Major Risk Factors. geneva, switzerland: World health organization, 2004, pp. 959–1109. tRamaCERE, i.; nEgRi, E.; bagnaRDi, V.; Et al. a meta-analysis of alcohol drinking and oral and pharyngeal cancers. Part 1: overall results and dose-risk relation. Oral Oncology REhm, J.; sEmPos, C.t.; anD tREVisan, m. alcohol and cardiovascular disease: more than 46(7):497–503, 2010. PmiD: 20444641 one paradox to consider. average volume of alcohol consumption, patterns of drinking and risk of coronary heart disease—a review. Journal of Cardiovascular Risk tRamaCERE, i.; nEgRi, E.; PEluCChi ,C.; Et al. a meta-analysis on alcohol drinking and gas- 10(1):15–20, 2003b. PmiD: 12569232 tric cancer risk. Annals of Oncology 23(1):28–36, 2012a. PmiD: 21536659

REhm, J.; tayloR, b.; mohaPatRa, s.; Et al. alcohol as a risk factor for liver cirrhosis: a sys- tRamaCERE, i.; PEluCChi, C.; boniFazi, m.; Et al. a meta-analysis on alcohol drinking and the tematic review and meta-analysis. Drug and Alcohol Review 29(4):437–445, 2010c. risk of hodgkin lymphoma. European Journal of Cancer Prevention 21(3):268–273, PmiD: 2063666 2012b. PmiD: 22465910

Rimm, E.b.; Williams, P.; FoshER, K.; Et al. moderate alcohol intake and lower risk of coro- tRamaCERE, i.; PEluCChi, C.; boniFazi, m.; Et al. alcohol drinking and non-hodgkin lym- nary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ phoma risk: a systematic review and a meta-analysis. Annals of Oncology Epub ahead 319(7224):1523–1528, 1999. PmiD: 10591709 of print, 2012c. PmiD: 22357444

RoCKhill, b.; nEWman, b.;anD WEinbERg, C. use and misuse of population attributable frac- tRamaCERE, i.; PEluCChi, C.; bagnaRDi, V.; Et al. a meta-analysis on alcohol drinking and tions. American Journal of Public Health 88(1):15–19, 1998. PmiD: 9584027 esophageal and gastric cardia adenocarcinoma risk. Annals of Oncology 23(2):287– 297, 2012d. PmiD: 21551004 RoERECKE, m., anD REhm, J. irregular heavy drinking occasions and risk of ischemic heart disease: a systematic review and meta-analysis. American Journal of Epidemiology tuma, D.J., anD CasEy, C.a. Dangerous byproducts of alcohol breakdown —Focus on 171(6):633–644, 2010. PmiD: 20142394 adducts. Alcohol Research & Health 27(4):285–290, 2003. PmiD: 15540799

RoERECKE, m., anD REhm, J. the cardioprotective association of average alcohol consump- tuRati, F.; gaRaVEllo, W.; tRamaCERE, i.; Et al. a meta-analysis of alcohol drinking and oral tion and ischaemic heart disease: a systematic review and meta-analysis. Addiction and pharyngeal cancers. Part 2: Results by subsites. Oral Oncology 46(10):720–726, 107(7):1246-1260, 2012. PmiD: 22229788 2010. PmiD: 20728401

Rota, m.; PasQuali, E.; sCotti, l.; Et al. alcohol drinking and epithelial ovarian cancer risk. tyas, s.l. alcohol use and the risk of developing alzheimer’s disease. Alcohol Research a systematic review and meta-analysis. Gynecologic Oncology 125(3):758–763, 2012. & Health 25(4):299–306, 2001. PmiD: 11910708 PmiD: 22449732 united nations. Prevention and Control of Non-Communicable Disease. new york: samoKhValoV, a.V.; iRVing, h.; mohaPatRa, s.; anD REhm, J. alcohol consumption, unpro- united nations; 2010. voked seizures, and epilepsy: a systematic review and meta-analysis. Epilepsia 51(7):1177–1184, 2010a. PmiD: 20074233 VERDouX, h.; lEngRonnE, J.; liRauD, F.; Et al. medication adherence in psychosis: Predictors and impact on outcome. a 2-year follow-up of first-admitted subjects. Acta samoKhValoV, a.V.; iRVing, h.m.; anD REhm, J. alcohol as a risk factor for atrial fibrillation: a Psychiatrica Scandinavica 102(3):203–210, 2000. PmiD: 11008856 systematic review and meta-analysis. European Journal of Cardiovascular Prevention and Rehabilitation 17(6):706–712, 2010b. PmiD: 21461366 VonlauFEn, a.; Wilson, J.s.; PiRola, R.C.; anD aPtE, m.V. Role of alcohol metabolism in chronic pancreatitis. Alcohol Research & Health 30(1):48–54, 2007. PmiD: 17718401 saRKola, t.; ilEs, m.R.; KohlEnbERg-muEllER, K.; anD ERiKsson, C.J. Ethanol, acetaldehyde, acetate, and lactate levels after alcohol intake in white men and women: Effect of 4- WannamEthEE, s.g.; CamaRgo, C.a.,JR.; manson, J.E.; Et al. alcohol drinking patterns and methylpyrazole. Alcoholism: Clinical and Experimental Research 26(2):239–245, 2002. risk of type 2 diabetes mellitus among younger women. Archives of Internal Medicine PmiD: 11964564 163(11):1329–1336, 2003. PmiD: 12796069 sChützE, m.; boEing, h.; PisChon, t.; Et al. alcohol attributable burden of incidence of WannamEthEE, s.g., anD shaPER, a.g. alcohol, body weight, and weight gain in middle-aged cancer in eight European countries based on results from prospective cohort study. men. American Journal of Clinical Nutrition 77(5):1312–1317, 2003. PmiD: 12716687 BMJ 342:d1584, 2011. available at http://www.bmj.com/content/342/bmj.d1584 World health organization. GBD 2004 Summary Tables. geneva, switzerland: World shaPER, a.g., anD WannamEthEE, s.g. the J-shaped curve and changes in drinking habit. health organization; 2008. Novartis Foundation Symposium 216:173–188, 1998. PmiD: 9949793 zaghloul, s.s., anD gooDFiElD, m.J. objective assessment of compliance with psoriasis siERKsma, a.; PatEl, h.; ouChi, n.; Et al. Effect of moderate alcohol consumption on treatment. Archives of Dermatology 140(4):408–414, 2004. PmiD: 15096368 adiponectin, tumor necrosis factor-alpha, and insulin sensitivity. Diabetes Care zaKhaRi, s. alcohol and the cardiovascular system: molecular mechanisms for beneficial 27(1):184–189, 2004. PmiD: 14693987 and harmful action. Alcohol Health and Research World. 21(1):21–29, 1997. PmiD: sKog, o.J. interpreting trends in alcohol consumption and alcohol related damage. 15706760 Alcohol and Alcoholism 23(3):193–202, 1988. PmiD: 3060129 zatonsKi, W.a.; sulKoWsKa, u.; manCzuK, m.; Et al. liver cirrhosis mortality in Europe, with song, D.y.; song, s.; song, y.; and lee, J.E. alcohol intake and renal cell cancer risk: a special attention to Central and Eastern Europe. European Addiction Research meta-analysis. British Journal of Cancer 106(11):1881–1890, 2012. PmiD: 22516951 16(4):193–201, 2010. PmiD: 20606444

Chronic Diseases and Conditions Related to Alcohol Use 171 Economic Costs of Excessive Alcohol Use

his brief sidebar summarizes How the Costs of Excessive costs that could be attributed to recent findings on the economic Alcohol Use Are Estimated excessive alcohol consumption.4 For tcosts of excessive alcohol use and example, researchers would consult a highlights the primary methods used As in the 1992 and 1998 reports of national health care database to to estimate these costs. The total eco- estimated costs, the most recent determine the cost of liver cirrhosis nomic cost of excessive drinking in report was based on the “cost-of-illness” and multiply this cost by a reasonable 2006,1 including costs for health AAF to determine the proportion of care, productivity losses, and costs approach. That is, costs for health such as property damage and alcohol- care, productivity losses, and other the total cost that can be attributed related crime, was estimated to be expenses for 2006 were obtained to excessive alcohol use. Likewise, $223.5 billion (see table) (Bouchery from national databases, and alcohol- costs of lost productivity, crime, and et al. 2011). Prior to this estimate, attributable fractions (AAFs) were other consequences of excessive alcohol the last comprehensive analysis reported applied to assess the proportion of use are estimated based on informa- that the estimated economic costs of excessive drinking were $148 billion in 1992 (Harwood et al. 1998). Data table Total Economic Costs of Excessive Alcohol Consumption in the United States, 2006 from that report were used to project a cost estimate of $185 billion for Cost Category total Cost 19982 (Harwood 2000). For a review of the global burden of alcohol use, Health Care Costs see Rehm and colleagues (2009). For alcohol abuse and Dependence $10,668.457 a broader examination of both the Primary Diagnoses attributable to alcohol 8,526.822 methods used to reach the current inpatient hospital 5,115.568 estimates and details on each of the Physician office and hospital ambulatory Care 1,195.946 estimated costs, as well as analysis nursing home Care 1,002.888 of the significance and limitations Retail Pharmacy and other health Professional 1,212.420 of the study, see Bouchery and Fetal alcohol syndrome 2,538.004 colleagues (2011, 2013). other health system Costs 2,822.308 Researchers used updated data, Prevention and Research 1,207.120 new data sources, and new measure- training 29.527 ment tools to develop the 2006 esti- health insurance administration 1,585.660 mate. For example, the Alcohol-Related total Health Care Costs $24,555.591 3 Disease Impact (ARDI) software, Productivity Losses created by the Centers for Disease impaired Productivity $83,695.036 Control and Prevention, was used to traditional Earnings 74,101.827 assess the relationship between excessive household Productivity 5,355.629 drinking and various health and social absenteeism 4,237.580 outcomes. This software, originally institutionalization/hospitalization 2,053.308 released in 2004, was designed to mortality 65,062.211 allow researchers to calculate alcohol- incarcerations 6,328.915 attributable deaths, years of potential Victims of Crime 2,092.886 life lost, direct health care costs, indi- Fetal alcohol syndrome 2,053.748 rect morbidity and mortality costs, total Productivity Losses $161,286.103 and non–health-sector costs associated with alcohol misuse. other Effects on Society Crime Victim Property Damage $439.766 Criminal Justice system 20,972.690 1 the most recent year for which data were available. motor Vehicle Crashes 13,718.406 2 the estimates given for 1992 and 1998 have not been Fire losses 2,137.300 adjusted for inflation. 3 the software is available online at: http://apps.nccd.cdc.gov/ Fas special Education 368.768 DaCh_aRDi/Default/Default.aspx. total other Effects $37,636.930 4 although cost-of-illness studies remain a popular tool for estimating the impact of alcohol abuse, researchers have total $223,478.624 criticized such studies as an invalid measure of alcohol’s impact on society. For a recent review of arguments against this approach, see mäkelä (2012).

172 Alcohol Research: Current Reviews tion gathered from various national Alcohol-Attributable Fractions care costs represent 11.0 percent, surveys and databases.5 criminal justice system costs make up AAFs, or the proportion of a condi- 9.4 percent, and other consequences tion or outcome that is attributed to make up 7.5 percent (see figure). ■ defining Excessive Alcohol excessive alcohol consumption, were Consumption used to estimate both health-related Acknowledgement The recent report estimated the costs costs, including deaths and health of excessive alcohol consumption, care expenditures related to excessive Source material for this article first which includes binge drinking (four drinking, and the costs of specific appeared in Bouchery et al. 2011, or more drinks per occasion for a criminal offenses. 2013. woman and five or more drinks per occasion for a man), heavy drinking Crime-Specific AAFs. 5 (more than one drink per day on Certain For information on the main data sources used in the report, alcohol-related crimes (e.g., driving see bouchery and colleagues (2011, 2013). average for a woman and more than 6 under the influence of alcohol, public although the methods used to estimate productivity losses two drinks per day on average for a attributed to premature mortality are consistent with previ- man), any alcohol consumption by drunkenness, and liquor law viola- ous cost-of-illness studies, alternative methods with greater tions) are fully attributed to alcohol. support from economists (i.e., the so-called willingness-to- persons younger than 21, and any pay approach) would yield much larger cost estimates. alcohol consumption by pregnant For other offenses, researchers women. The surveys used generally estimated the proportion attributable ask respondents about the 30 days to alcohol based on the percentage prior to the survey. The focus on of offenders intoxicated at the time excessive drinking, rather than alco- of their offense (according to self- References reported alcohol-consumption data hol use disorders, allows for a more bouChERy, E.; haRWooD, h.; saCKs, J.; Et al. Economic comprehensive look at the health from surveys of inmates). For homicide, costs of excessive alcohol consumption in the u.s., and social consequences that are researchers used the AAF from ARDI 2006. American Journal of Preventive Medicine 41 (5):516–524, 2011. PmiD: 22011424 associated with harmful drinking because it takes into account drinking by both the perpetrator and the victim. bouChERy, E.; simon, C.; anD haRWooD, h. Economic patterns, including a wide range of Costs of Excessive Alcohol Consumption in the United acute and chronic health problems, States, 2006. the lewin group, updated, 2013. productivity losses due to absen- available at: http://www.lewin.com/~/media/lewin/ the Costs of Excessive site_sections/Publications/CDC_Report_Rev.pdf. teeism, and crimes committed while accessed June 27, 2013. intoxicated. In some cases, a history Alcohol Use haRWooD, h. Updating Estimates of the Economic of alcohol abuse or dependence was Costs of Alcohol Abuse in the U.S.: Estimates, Update used as a specific indicator of exces- Of the total estimated cost of excessive Methods and Data. Report prepared by the lewin alcohol use for 2006, lost productiv- group for the national institute on alcohol abuse and sive drinking (e.g., productivity alcoholism, 2000. losses based on lost earnings). ity represents 72.2 percent,6 health haRWooD, h.; Fountain, D.; anD liVERmoRE, g. The Economic Costs of Alcohol and Drug Abuse in the U.S., 1992. Report prepared for the national institute on Drug abuse and the national institute on alcohol abuse and Other Effects alcoholism, Rockville, mD [nih Publication no. 98–4327]. 7.5% mäKElä, K. Cost-of-alcohol studies as a research pro- gramme. Nordic Studies on Alcohol and Drugs 29 Criminal Justice System (4):321–344, 2012. 9.4% REhm, J.; mathERs, C.; PoPoVa, s.; Et al. global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders. Lancet Health Care Costs 373(9682):2223–2233, 2009. PmiD: 19560604 11.0% Lost Productivity 72.2%

Figure

Economic Costs of Excessive Alcohol Use 173 and injury categories cause mortality and disability, and, Alcohol and Mortality thus, alcohol consumption causes a net burden of mortality and disability (Rothman et al. 2008). However, certain pat- terns of alcohol consumption are protective for ischemic dis- eases (Roerecke and Rehm 2012a) and diabetes (Baliunas et Global Alcohol-Attributable deaths al. 2009), and, thus, alcohol can prevent death and disability from these causes. The total mortality and disability caused From Cancer, Liver Cirrhosis, and by and prevented by the consumption of alcohol is calcu- lated by comparing the expected mortality under current Injury in 2010 conditions to a counterfactual scenario where no one has consumed alcohol (Ezzati et al. 2006; Walter 1976). Although the counterfactual scenario seems unrealistic as almost one-half of the global population consumes alcohol (for the most up-to-date statistics on alcohol consumption, see Shield et al. 2013b; World Health Organization 2011a), Jürgen Rehm, Ph.D., and kevin D. Shield, MH.Sc. recent natural experiments in countries where there has been a considerable reduction in alcohol consumption showed a alcohol consumption has long been recognized as a risk clear reduction in mortality (e.g., Russia) (Leon et al. 1997; factor for mortality. by combining data on alcohol per capita Neufeld and Rehm 2013). Accordingly, the calculations consumption, alcohol-drinking status and alcohol-drinking of the deaths and disability caused by alcohol consumption patterns, risk relationships, and mortality, the Comparative seem to correspond to real phenomena and, thus, could pre- Risk assessment study estimated alcohol-attributable dict an approximate level of reduction in mortality if alcohol mortality for 1990 and 2010. alcohol-attributable cancer, consumption were to be reduced. liver cirrhosis, and injury were responsible for the majority of This article outlines the alcohol-attributable mortality the burden of alcohol-attributable mortality in 1990 and burden from three major causes: cancer, liver cirrhosis, and 2010. in 2010, alcohol-attributable cancer, liver cirrhosis, injury. All three categories have long been identified as causally and injury caused 1,500,000 deaths (319,500 deaths linked to alcohol consumption. With respect to cancer, in among women and 1,180,500 deaths among men) and 1988 the International Agency for Research on Cancer 51,898,400 potential years of life lost (Pyll) (9,214,300 established alcohol as a carcinogen (International Agency for Pyll among women and 42,684,100 Pyll among men). this Research on Cancer 1988), and in its latest monograph has represents 2.8 percent (1.3 percent for women and 4.1 found alcohol consumption to be causally associated with percent for men) of all deaths and 3.0 percent (1.3 percent oral cavity, pharynx, larynx, esophagus, liver, colon, rectum, for women and 4.3 percent for men) of all Pyll in 2010. the and female breast cancers (International Agency for Research absolute mortality burden of alcohol-attributable cancer, liver on Cancer 2010, 2012). Studies have shown that stomach cirrhosis, and injury increased from 1990 to 2010 for both cancer may be associated with alcohol consumption, but genders. in addition, the rates of deaths and Pyll per evidence on the causal relationship between stomach cancer 100,000 people from alcohol-attributable cancer, liver cirrhosis, and injury increased from 1990 to 2010 (with the exception of liver cirrhosis rates for women). Results of this Jürgen Rehm, Ph.D., is Director, Social and paper indicate that alcohol is a significant and increasing risk Epidemiological Research (SER) Department, Centre for factor for the global burden of mortality. kEY WoRDS: Alcohol Addiction and Mental Health, Toronto, Canada; Professor, consumption; alcohol burden; alcohol-attributable Institute of Medical Science, University of Toronto, Canada; mortality; alcohol-attributable fractions; global alcohol- Chair Addiction Policy, Dalla Lana School of Public Health, attributable mortality; risk factor; cancer; liver cirrhosis; University of Toronto, Canada; Professor, Department of injury; burden of disease; Global Burden of Disease and Psychiatry, University of Toronto, Canada; Head, PAHO/WHO injury study Collaborating Centre for Mental Health & Addiction; and Head, Technische Universität Dresden, Klinische Psychologie & Psychotherapie, Dresden, Germany.

Alcohol and Mortality kevin D. Shield, MH.Sc., is a pre-doctoral fellow, Social and Epidemiological Research (SER) Department, Centre Alcohol is causally linked to more than 200 different diseases, for Addiction and Mental Health, Toronto, Canada, and conditions, and injuries (as specified in the International Ph.D. student at the Institute of Medical Science, University Classification of Diseases, Revision 10 [ICD-10] three-digit of Toronto, Canada. codes [see Rehm 2011; Rehm et al. 2009; Shield et al., 2013c [pp. 155–173 of this issue]). All of these disease, condition,

174 Alcohol Research: Current Reviews and alcohol consumption is not yet conclusive (International deaths were attributed to alcohol, which is an impossibility Agency for Research on Cancer 2012; Rehm and Shield, in as the relationship between alcohol consumption and this press). Biologically, it has been established that ethanol, and disease category is mainly protective (for details on relation- not other ingredients of alcoholic beverages, is the ingredient ship between alcohol and heart disease, see Roerecke and that mainly causes cancer (Lachenmeier et al. 2012), with Rehm 2010, 2012b). A comparison with other alcohol- acetaldehyde (the first metabolite of ethanol) likely being the attributable disease and protective effects will thus only most important biological carcinogen (International Agency be possible once the corrected CRA results are published. for Research on Cancer 2010, 2012; Rehm and Shield, in press). In addition, observational studies have found a clear dose-response relationship between alcohol consumption Methodology Underlying the Estimation of the and the risk of cancer, with no observed threshold for the Mortality Burden of Alcohol-Attributable Diseases effect of alcohol, as an elevated risk of cancer has been observed and injuries even for people who consume relatively low amounts of alcohol (Bagnardi et al. 2013; Rehm et al. 2010a). The number of alcohol-attributable cancer, liver cirrhosis, Liver cirrhosis has been associated with alcohol consump- and injury deaths in 1990 and 2010 were estimated using tion, especially heavy consumption, since the seminal work alcohol-attributable fractions (AAFs) (Benichou 2001; of Benjamin Rush (Rush 1785). The causal link between Walter 1976, 1980). AAFs are calculated by comparing alcohol consumption and liver cirrhosis is so strong and the population risk of a disease under current conditions to a important that the World Health Organization has created counterfactual scenario where no one has consumed alcohol. a specific category for alcoholic liver cirrhosis (World Health This is achieved by using information on the distribution of Organization 2007). As with cancer, there is a dose-response levels of alcohol consumption and the associated relative relationship between alcohol consumption and the risk of risks (RRs) (i.e., risks of disease for different levels of alcohol liver cirrhosis, with no lower threshold being observed (Rehm consumption versus abstainers). These calculated AAFs et al. 2010c); however, the majority of the effect can be seen then were applied to mortality data obtained from the 2010 for heavy drinking (Rehm et al. 2010c). Global Burden of Disease (GBD) Study for 1990 and 2010 The risk of injury also has been causally linked to alcohol (Lim et al. 2012). Mortality data for 1990 and 2010 were consumption, with this relationship fulfilling all of the classic modelled using data on mortality from 1980 to 2010. Data Bradford Hill criteria (e.g., consistency of the effect, tempo- on mortality were imputed for those countries with little rality, a dose-repsonse relationship with the risk of an injury or no data by using data from other countries and were [biological gradient]) (Rehm et al. 2003a). The effect of smoothed over time (in addition to other data corrections alcohol on injury is acute; the level of risk for both intentional procedures that corrected for cause of death recording errors) and unintentional injuries is clearly linked to blood alcohol (Lozano et al. 2012). level (Taylor and Rehm 2012; Taylor et al. 2010), with a very low threshold (Eckardt et al. 1998). There also is an Calculating the Alcohol-Attributable Mortality Burden association between average consumption of alcohol and of Cancer and Liver Cirrhosis injury (Corrao et al. 2004). Alcohol-attributable cancer, liver cirrhosis, and injury Alcohol consumption is causally related to mouth and constitute the majority of the burden of alcohol-attributable oropharynx cancers (ICD-10 codes: C00 to C14), esophageal mortality. Collectively, they were responsible for 89 percent cancer (C15), liver cancer (C22), laryngeal cancer (C32), of the net burden of alcohol-attributable mortality (i.e., the breast cancer (C50), colon cancer (C18), and rectal cancer mortality rate after including the beneficial effects of alcohol (C20). Alcohol RR functions for cancer were obtained from on ischemic diseases and diabetes) and for 79 percent of the Corrao and colleagues (2004) (For information about the gross burden of alcohol-attributable mortality (Shield et al. causal relationship between alcohol and cancer, see Baan 2013a) in the United States in 2005, for people 15 to 64 et al. 2007; International Agency for Research on Cancer years of age. Additionally, they were responsible for 91 percent 2010.) The alcohol RR for liver cirrhosis (ICD-10 codes: of the net alcohol-attributable mortality and 79 percent of K70 and K74) was obtained from Rehm and colleagues the gross alcohol-attributable mortality in the European (Rehm et al. 2010c). The above-noted RRs were modelled Union (Rehm et al. 2012) and 80 percent of the net alcohol- based on drinking status and average daily alcohol consump- attributable mortality and 72 percent of the gross alcohol- tion among drinkers. The same RRs were used to estimate attributable mortality globally (Rehm et al. 2009) in 2004. the AAFs by country, sex, and age for 1990 and for 2010. This article does not review the other causes of alcohol- Alcohol-drinking statuses and adult (people 15 years of attributable deaths included in the latest Comparative Risk age and older) per capita consumption data for 1990 were Assessment (CRA) Study (Lim et al. 2012). The CRA study obtained from various population surveys (Shield et al. estimates as published in December contained significant 2013b), and the Global Information System on Alcohol and errors in the calculation of alcohol-attributable cardiovascular Health (available at: http://apps.who. int/ghodata/?theme= deaths, estimating that 33 percent of all ischemic heart disease GISAH), respectively. Data on drinking status and adult per

Global Alcohol-Attributable Deaths 175

UseChronic Diseases and Conditions Related to Alcohol Use

capita consumption for 2010 were estimated by projections attributable injuries caused to nondrinkers also were estimated (performed using regression analyses) using data from years (Shield et al. 2012). prior to 2010 (Shield et al. 2013b). Average daily alcohol consumption was modelled using a gamma distribution (Rehm et al. 2010b) and data on per capita consumption for Global Consumption of Alcohol 1990, which was projected to 2010 (Shield et al. 2013b). (For more information on the methodology of how alcohol In 2005 adult per capita consumption of alcohol was 6.1 consumption was modelled, see Kehoe et al. 2012; Rehm et litres of pure alcohol. Figure 1 shows the adult per capita al. 2010b). This paper presents alcohol consumption data consumption of alcohol by country. Alcohol consumption from 2005, the latest year with actual data available. per drinker in 2005 was 17.1 litres (9.5 litres per female drinker and 20.5 litres per male drinker). Of all adults, 45.8 Calculating the Alcohol-Attributable Mortality Burden percent were lifetime abstainers (55.6 percent of female of Injuries adults and 36.0 percent of male adults), 13.6 percent were former drinkers (13.1 percent of female adults and 14.1 per- The burden of injury mortality attributable to alcohol con- cent of male adults), and 40.6 percent were current drinkers sumption was modelled according to methodology outlined (31.3 percent of female adults and 49.9 percent of male adults). by Shield and colleagues (2012), using risk information The global pattern of drinking score (a score from 1 to 5 obtained from a meta-analysis (Taylor et al. 2010) and alcohol that measures the harmfulness of alcohol drinking patterns, consumption data from 1990 and 2010. The risk of an with 1 being the least harmful and 5 being the most harmful injury caused to the drinker over a year was calculated based [Rehm et al. 2003b]) was 2.6 in 2005 and ranged from 4.9 on alcohol consumed during normal drinking occasions and for Eastern Europe to 1.5 for Western Europe. Thus, alcohol alcohol consumed during binge-drinking occasions. Alcohol- consumption in Eastern Europe can be characterized by fre-

Figure 1 adult per capita consumption of pure alcohol by country in 2005.

notE: more detailed information can be obtained from the author.

176 Alcohol Research: Current Reviews quent heavy alcohol consumption outside of meals and alcohol-attributable cancer, with 0.6 deaths and 17.1 PYLL drinking to intoxication. per 100,000 people. In 1990, alcohol-attributable cancer caused 243,000 deaths worldwide (70,700 deaths among women and 172,300 Global Alcohol-Attributable Mortality From Cancer deaths among men) and 6,405,700 PYLL (1,762,200 PYLL among women and 4,643,500 PYLL among men). This In 2010, alcohol-attributable cancer caused 337,400 deaths mortality burden is equal to 4.6 deaths per 100,000 people (91,500 deaths among women and 245,900 deaths among (2.7 deaths per 100,000 women and 6.5 deaths per 100,000 men) and 8,460,000 PYLL (2,143,000 PYLL among men) and 120.8 PYLL per 100,000 people (67.0 PYLL per women and 6,317,000 PYLL among men). This burden 100,000 women and 173.9 PYLL per 100,000 men) caused is equal to 4.9 deaths per 100,000 people (2.7 deaths per by alcohol-attributable cancer. From 1990 to 2010 the abso- 100,000 women and 7.1 deaths per 100,000 men) and lute mortality burden of alcohol-attributable cancer (mea- 122.8 PYLL per 100,000 people (62.8 PYLL per 100,000 sured in deaths and PYLL) and the rates of deaths and PYLL women and 181.9 PYLL per 100,000 men). Stated another per 100,000 people have each increased. way, alcohol-attributable cancer was responsible for 4.2 per- cent of all cancer deaths in 2010 and 4.6 percent of all PYLL caused by cancer. Figure 2 shows the number of alcohol- Global Alcohol-Attributable Mortality From Liver attributable cancer deaths per 100,000 people by region in 2010. Eastern Europe had the highest burden of mortality Cirrhosis and morbidity from alcohol-attributable cancer, with 10.3 In 2010, alcohol-attributable liver cirrhosis was responsible deaths and 272.0 PYLL per 100,000 people. North Africa for 493,300 deaths worldwide (156,900 deaths among and the Middle East had the lowest mortality burden of women and 336,400 deaths among men) and 14,327,800 PYLL

Figure 2 alcohol-attributable cancer deaths per 100,000 people in 2010 by global-burden-of-disease region.

notE: more detailed information can be obtained from the author.

Global Alcohol-Attributable Deaths 177

UseChronic Diseases and Conditions Related to Alcohol Use

(4,011,100 PYLL among women and 10,316,800 PYLL attributable liver cirrhosis (measured in deaths and PYLL) among men). This mortality burden is equal to 7.2 deaths and this mortality burden per 100,000 people have each per 100,000 people (4.6 deaths per 100,000 women and 9.7 increased (except for women, where alcohol-attributable liver deaths per 100,000 men) and 208.0 PYLL per 100,00 peo- cirrhosis deaths and PYLL per 100,000 decreased slightly). ple (117.5 PYLL per 100,000 women and 297.0 PYLL per 100,000 men) caused by alcohol-attributable liver cirrhosis in 2010. Overall, in 2010 alcohol-attributable liver cirrhosis Global Alcohol-Attributable Mortality From injury was responsible for 47.9 percent of all liver cirrhosis deaths and 47.1 percent of all liver cirrhosis PYLL. Figure 3 out- Globally in 2010, alcohol-attributable injuries were responsi- lines the number of alcohol-attributable liver cirrhosis deaths ble for 669,300 deaths (71,100 deaths among women and per 100,000 people by region in 2010, showing strong 598,200 deaths among men) and 29,110,600 PYLL regional variability. (3,060,200 PYLL among women and 26,050,400 PYLL In 1990, alcohol-attributable liver cirrhosis was responsi- among men). This mortality burden is equal to 9.7 deaths ble for 373,200 deaths worldwide (125,300 deaths among per 100,000 people (2.1 deaths per 100,000 women and women and 247,900 deaths among men) and 10,906,200 17.2 deaths per 100,000 men) and 422.6 PYLL per 100,000 PYLL (3,253,300 PYLL among women and 7,652,900 people (89.6 PYLL per 100,000 women and 750.0 PYLL PYLL among men). That is, 7.0 deaths per 100,000 people per 100,000 men). Overall, in 2010 alcohol-attributable (4.8 deaths per 100,000 women and 9.3 deaths per 100,000 injuries were responsible for 13.2 percent of all injury deaths men) and 205.7 PYLL per 100,000 people (123.7 PYLL per and 12.6 percent of all injury PYLL. Figure 4 outlines the 100,000 women and 286.6 PYLL per 100,000 men) were number of alcohol-attributable injury deaths per 100,000 caused by liver cirrhosis attributable to alcohol consumption. people in 2010. Eastern Europe had the greatest mortality From 1990 to 2010, the absolute mortality burden of alcohol- burden of alcohol-attributable injuries, with 76.7 deaths and

Figure 3 alcohol-attributable liver cirrhosis deaths per 100,000 people in 2010 by global-burden-of-disease region.

notE: more detailed information can be obtained from the author.

178 Alcohol Research: Current Reviews

UseChronic Diseases and Conditions Related to Alcohol Use

3,484.7 PYLL per 100,000 people, whereas North Africa Global Alcohol-Attributable Cancer, Liver and the Middle East had the lowest mortality burden, with Cirrhosis, and injury Mortality As Part of 2.0 deaths and 117.2 PYLL per 100,000 people. overall Mortality In 1990, alcohol-attributable injuries were responsible for 485,100 deaths (54,700 deaths among women and In 2010, alcohol-attributable cancer, liver cirrhosis, and injury 430,400 deaths among men) and 21,934,800 PYLL (2,409,100 caused 1,500,000 deaths (319,500 deaths among women PYLL among women and 19,525,700 PYLL among men), and 1,180,500 deaths among men). This represents 2.8 percent equal to 9.2 deaths (2.1 deaths per 100,000 women and of all deaths (1.3 percent of all deaths among women and 16.1 deaths per 100,000 men) and 413.8 PYLL per 100,000 4.1 percent of all deaths among men), or 21.8 deaths per people (91.6 PYLL per 100,000 women and 731.3 PYLL 100,000 people (9.4 deaths per 100,000 women and 34.0 per 100,000 men). The absolute number of alcohol- deaths per 100,000 men). In 1990, alcohol-attributable cancer, attributable injury deaths and PYLL and the number of liver cirrhosis, and injury caused 1,101,400 deaths (250,800 alcohol-attributable injury deaths and PYLL per 100,000 deaths among women and 850,600 deaths among men), people each increased from 1990 to 2010. representing 20.8 deaths per 100,000 people (9.5 deaths per Appendix 1 presents the number and percentage of 100,000 women and 31.9 deaths per 100,000 men). The table outlines the mortality burden (measured in deaths and alcohol-attributable cancer, liver cirrhosis, and injury deaths PYLL) of alcohol-attributable cancer, liver cirrhosis, and and PYLL by GBD study region for 1990 and 2010. Appendix injury for 1990 and 2010 by age and by sex. Compared with 2 presents the number of alcohol-attributable cancer, liver the mortality burden in 1990, the absolute number of alcohol- cirrhosis, and injury deaths per 100,000 people. Unlike fig- attributable deaths from cancer, liver cirrhosis, and injury in ures 1, 2, and 3, the figures in Appendix 2 use the same scale 2010 is higher, and the rate of deaths per 100,000 also for each cause of death. increased for men but decreased slightly for women in 2010.

Figure 4 alcohol-attributable injury deaths per 100,000 people in 2010 by global-burden-of-disease region.

notE: more detailed information can be obtained from the author.

Global Alcohol-Attributable Deaths 179

UseChronic Diseases and Conditions Related to Alcohol Use

The burden of mortality from alcohol-attributable cancer, Measurement Limitations liver cirrhosis, and injury led to 51,898,400 PYLL (9,214,300 PYLL among women and 42,684,100 PYLL among men) in The methods used to estimate the number of alcohol- 2010 and 39,246,800 PYLL (7,424,600 PYLL among attributable cancer, liver cirrhosis, and injury deaths and women and 31,822,100 PYLL among men) in 1990. This PYLL have limitations as a result of the available data on mortality burden represents 3.0 percent (1.3 percent for mortality and the measurement of alcohol consumption and women and 4.3 percent for men) of all PYLL in 2010 and RRs. Most low- and middle-income countries do not have 2.0 percent (0.9 percent for women and 2.9 percent for reliable mortality data and measurement of adult mortality men) of all PYLL in 1990. In 2010, alcohol-attributable in these countries (through verbal autopsies or surveys) is cancer, liver cirrhosis, and injury led to 753.4 PYLL per infrequent. Therefore, estimates of mortality and PYLL have 100,000 people (269.8 PYLL per 100,000 women and a large degree of uncertainty (Wang et al. 2012). Additionally, 1,228.9 PYLL per 100,000 men) and to 740.4 PYLL per for high-income countries, information concerning the cause 100,000 people (282.2 PYLL per 100,000 women and of death has long been acknowledged as containing inaccu- 1,191.9 per 100,000 men) in 1990. Again, the overall rates racies (James et al. 1955), and more recent studies have of PYLL per 100,000 people increased, but this effect was confirmed considerable degrees of error in this information attributed to increases for men, coupled with slight decreases (Nashelsky and Lawrence 2003; Shojania et al. 2003). To for women. adjust for inaccuracies and inconsistencies in mortality data,

table 1 Deaths and years of life lost (yll) From Cancer, liver Cirrhosis, and injuries attributable to alcohol Consumption in 1990 and 2010 Year Gender Age (Years) Deaths % of All Deaths YLL % of All YLL

1990 Women 0 to 15 4,000 0.1 324,400 0.1 15 to 34 22,300 1.5 1,349,500 1.5 35 to 64 128,700 2.9 4,437,000 3.0 65+ 95,800 1.0 1,313,800 1.1 total 250,800 1.2 7,424,600 0.9

men 0 to 15 6,700 0.1 540,400 0.1 15 to 34 174,400 8.4 10,547,900 8.4 35 to 64 502,600 7.4 18,167,100 7.8 65+ 166,800 1.8 2,566,700 2.0 total 850,600 3.4 31,822,100 2.9

total total 1,101,400 2.4 39,246,800 2.0

2010 Women 0 to 15 3,800 0.1 313,800 0.1 15 to 34 28,800 1.7 1,741,700 1.7 35 to 64 162,000 3.1 5,570,800 3.1 65+ 124,800 0.9 1,587,900 1.1 total 319,500 1.3 9,214,300 1.3

men 0 to 15 6,100 0.1 492,400 0.1 15 to 34 214,900 8.5 12,972,300 8.5 35 to 64 709,200 7.9 25,549,800 8.2 65+ 250,300 1.9 3,669,500 2.2 total 1,180,500 4.1 42,684,100 4.3

total total 1,500,000 2.8 51,898,400 3.0

notE: more detailed information can be obtained from the author.

180 Alcohol Research: Current Reviews the 2010 GBD study modelled the number of deaths math- the risk of injury is associated with blood alcohol content ematically (Wang et al. 2012). (Taylor and Rehm 2012; see also Cherpitel 2013). Survey data measuring alcohol consumption, patterns of alcohol consumption, and the prevalence of lifetime abstainers, former drinkers, and current drinkers also are susceptible implications of Alcohol-Attributable Mortality to numerous biases (Shield and Rehm 2012). To correct for the undercoverage that is observed when alcohol consump- In 1990 and in 2010, alcohol consumption had a huge tion is measured by population surveys (as compared with impact on mortality. Regions such as Europe (especially per capita consumption of alcohol), alcohol consumption Eastern Europe) and parts of Sub-Saharan Africa (especially was modelled by triangulating per capita and survey data south Sub-Saharan Africa) that have a high per capita con- (see above). Total alcohol consumption was set to 80 percent sumption of alcohol and detrimental drinking patterns are of per capita consumption in order to account for alcohol more affected by alcohol consumption compared with other produced and/or sold, but not consumed, and to account regions. It is important to note that the alcohol-attributable for the undercoverage of the alcohol consumption typically mortality burden is composed of two elements: AAF and the observed in studies that calculate the alcohol RRs (Rehm et overall mortality burden of the respective disease. Accordingly, al. 2010b). Additionally, although alcohol was measured the observed overall increase from 1990 to 2010 in alcohol- using adult per capita consumption and most people 14 attributable cancer, liver cirrhosis, and injury deaths and years and younger do not consume alcohol or binge regularly, in PYLL can be attributed to two different sources: (1) an some adolescents 10 to 14 years of age report previously trying increase in the number of cancer, liver cirrhosis, and injury alcohol and previously being intoxicated (Windle et al. 2008). deaths (mainly attributed to increases of these deaths in low- The CRA was based on alcohol RR functions that typically to middle-income countries) (Lozano et al. 2012) and (2) an were differentiated by sex and adjusted for age, smoking status, increase in alcohol consumption in low- to middle-income and other potentially confounding factors. The use of adjusted countries (Shield et al. 2013b). RR functions may introduce bias into the estimated number Low- and middle-income countries have higher rates of deaths and PYLL that would not have occurred if no one of alcohol-attributable mortality per 100,000 people, even had ever consumed alcohol (Flegal et al. 2006; Korn and though these countries typically have lower AAF (as their Graubard 1999; Rockhill and Newman 1998). However, overall burden of mortality is higher). Economic wealth is most of the published literature on alcohol-as-a-risk-factor– correlated with overall mortality (Lozano et al. 2012), and, only reports adjusted RRs, and, thus, the use of unadjusted thus, the mortality burden per litre of alcohol consumed is alcohol RRs for the CRA study would have led to imprecise highest in low-income countries, followed by middle-income estimates as a result of leaving out most of the studies. The countries (Rehm et al. 2009; Schmidt et al. 2010). It follows, bias of using adjusted RRs is likely to be small, as most analyses therefore, that increases in the alcohol-attributable mortality of the estimated RRs show no marked differences after burden in low- and middle-income countries attributed to adjustment for the potentially confounding factors and effect economic growth may be able to be reduced or controlled measure modifiers. Future CRA studies may require more for by implementing alcohol control policies such as taxation complex modelling techniques for alcohol if other dimensions (Shield et al. 2011; Sornpaisarn et al. 2012a, b, 2013), bans of alcohol consumption, such as irregular heavy-drinking occasions, impact RR estimates. on advertising, and restrictions on availability (Anderson et Finally, this analysis did not account for a lag time for al. 2009; World Health Organization 2011b) preferably the calculation methods used in this paper. This is especially while maintaining the relatively high levels of abstention in a problem for cancer, which has a lag time of 15 to 20 years these countries. (Holmes et al. 2012; Rehm et al. 2007). In other words, the The typical causes of death associated with alcohol use alcohol-attributable deaths and PYLL in 2010 actually are disorders are liver cirrhosis and injuries, (i.e., exactly the cat- based on consumption patterns from 1990 to 1995, but in egories described in this article). Liver cirrhosis and injuries, this paper were estimated based on consumption in 1990 and to a lesser degree cancer, may primarily be responsible and 2010. Although liver cirrhosis also is a chronic disease for the high proportion of alcohol-attributable mortality that develops over time like cancer (Rehm et al. 2013a), the explained by alcohol use disorders (Rehm et al. 2013b); impact of population-level consumption on liver cirrhosis however, additional research is required to empirically confirm deaths can be quite abrupt. For example, Gorbachev’s anti- this hypothesis. By increasing the rate of treatment for alcohol alcohol campaign was reflected in a clear reduction in Russia’s use disorders (Rehm et al. 2013b), the mortality burden of liver cirrhosis mortality (Leon 1997). Likewise, the German alcohol-attributable diseases also can be reduced. Recent seizure of alcohol in France in World War II led to reduced research has shown that the mortality burden associated with cirrhosis mortality (Zatonski et al. 2010). Most of the effect alcohol use disorders, albeit high, has been underestimated of alcohol consumption on liver cirrhosis probably is cap- (see Harris and Barraclough 1998 for the first meta-analysis; tured within 1 year (Holmes et al. 2012). For injury, with and Callaghan et al. 2012; Campos et al. 2011; Guitart et al. the exception of suicide, there is no noticeable lag time as 2011; Hayes et al., 2011; Saieva et al. 2012; Tikkanen et al.

Global Alcohol-Attributable Deaths 181 2009 for recent papers that observed a markedly higher mor- haRRis, E. C. anD baRRaClough, b. Excess mortality of mental disorder. British Journal of tality risk than in the first meta-analysis). ■ Psychiatry 173:11–53, 1998. PmiD: 9850203 hayEs, R.D.; Chang, C.K.; FERnanDEs, a., Et al. associations between substance use disor- der sub-groups, life expectancy and all-cause mortality in a large british specialist men- Acknowledgements tal healthcare service. Drug and Alcohol Dependence 118:56–61, 2011. PmiD: 21440382

We thank all who helped to realize the Comparative Risk holmEs, J.; mEiER, P. s.; booth, a., Et al. the temporal relationship between per capita Assessment for alcohol within the 2010 Global Burden of alcohol consumption and harm: a systematic review of time lag specifications in aggre- Disease and Injury study, and to especially thank Gerrit gate time series analyses. Drug and Alcohol Dependence 123:7–14, 2012. PmiD: 22197480 Gmel and Tara Kehoe-Chan, who helped program the alcohol- attributable fractions. international agency for Research on Cancer. Alcohol Drinking. lyon, France: international agency for Research on Cancer, 1988.

international agency for Research on Cancer Alcohol Consumption and Ethyl Financial Disclosure Carbamate. monograph 96 on the Evaluation of Carcinogenic Risks to humans. lyon, France: international agency for Research on Cancer, 2010.

The authors declare that they have no competing financial international agency for Research on Cancer. A Review of Human Carcinogens: interests. Personal Habits and Indoor Combustions. Vol. 100E. lyon, France: international agency for Research on Cancer, 2012.

JamEs, g.; Patton, R. E. anD hEslin, a.s. accuracy of cause-of-death statements on death References certificates. Public Health Reports 70:39–51, 1955. PmiD: 13224799

KEhoE, t.; gmEl, g., JR.; shiElD, K.D., anD REhm, J. Determining the best population-level anDERson, P.; Chisholm, D. anD FuhR, D.C. Effectiveness and cost-effectiveness of policies alcohol consumption model and its impact on estimates of alcohol-attributable harms. and programmes to reduce the harm caused by alcohol. Lancet 373:2234–2246, 2009. PmiD: 19560605 Population Health Metrics 10:6, 2012. PmiD: 22490226 KoRn, E.l., anD gRaubaRD, b.i. Analysis of Health Surveys. new york, ny: John Wiley & baan, R.; stRaiF, K.; gRossE, y., Et al. Carcinogenicity of alcoholic beverages. Lancet Oncology 8:292–293, 2007. PmiD: 17431955 sons, 1999.

laChEnmEiER, D.W.; PRzybylsKi, m.C. anD REhm, J. Comparative risk assessment of carcino- bagnaRDi, V.; Rota, m.; bottERi, E., Et al. light alcohol drinking and cancer: a meta-analysis. Annals of Oncology 24:301–308, 2013. PmiD: 22910838 gens in alcoholic beverages using the margin of exposure approach. International Journal of Cancer 131:E995–E1003, 2012. PmiD: 22447328 baliunas, D.o.; tayloR, b.J.; iRVing, h., Et al. alcohol as a risk factor for type 2 diabetes— a systematic review and meta-analysis. Diabetes Care 32:2123–2132, 2009. PmiD: lEon, D.a.; ChEnEt, l.; shKolniKoV, V.m.; Et al. huge variation in Russian mortality rates 19875607 1984-1994: artefact, alcohol, or what? Lancet 350:383–388, 1997. PmiD: 9259651

bEniChou, J. a review of adjusted estimators of attributable risk. Statistical Methods in lim, s.s.; Vos, t.; FlaXman, a.D.; Et al. a comparative risk assessment of burden of dis- Medical Research 10:195–216, 2001. PmiD: 11446148 ease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the global burden of Disease study 2010 Lancet Callaghan, R. C.; Cunningham, J. K.; VERDiChEVsKi, m., Et al. all-cause mortality among indi- 380:2224–2260, 2012. PmiD: 23245609 viduals with disorders related to the use of methamphetamine: a comparative cohort study. Drug and Alcohol Dependence 125:290–294, 2012. PmiD: 22503689 lozano, R.; naghaVi, m.; FoREman, K.; Et al. global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the global CamPos, J.; RoCa, l.; guDE, F., Et al. long-term mortality of patients admitted to the hos- burden of Disease study 2010. Lancet 380: 2095–2128, 2012. PmiD: 23245604 pital with alcohol withdrawal syndrome. Alcoholism: Clinical and Experimental Research 35:1180–1186, 2011. PmiD: 21352245 nashElsKy, m.b., anD laWREnCE, C.h. accuracy of cause of death determination without forensic autopsy examination. American Journal of Forensic Medicine and Pathology ChERPitEl, C. J. Focus on: trauma and emergency outcomes. Alcohol Research: Current 24:313–319, 2003. PmiD: 14634467 Reviews 35(2):150–154, 2013. nEuFElD, m., anD REhm, J. alcohol consumption and mortality in Russia since 2000 : are CoRRao, g.; bagnaRDi, V.; zambon, a., anD la VECChia, C. a meta-analysis of alcohol con- there any changes following the alcohol policy changes starting in 2006? Alcohol and sumption and the risk of 15 diseases. Preventive Medicine 38:613–619, 2004. PmiD: Alcoholism 48:222-230, 2013. PmiD: 23299570 15066364 REhm, J. the risks associated with alcohol use and alcoholism. Alcohol Research & ECKaRDt, m.J.; FilE, s.E.; gEssa, g.l., Et al. Effects of moderate alcohol consumption on Health 34:135–143, 2011. PmiD: 22330211 the central nervous system. Alcoholism: Clinical and Experimental Research 22:998–1040, 1998. PmiD: 9726269 REhm, J.; baliunas, D.; boRgEs, g.l.; Et al. the relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction 105: 817–843, Ezzati, m.; hooRn, s. V.; loPEz, a. D., Et al. Comparative quantification of mortality and 2010a. PmiD: 20331573 burden of disease attributable to selected risk factors. in: lopez, a.D.; mathers, C.D.; Ezzati, m.; Et al., Eds. Global Burden of Disease and Risk Factors, pp. 241–268. REhm, J.; gmEl, g.; sEmPos, C.t.; anD tREVisan, m. alcohol-related morbidity and mortality. Washington, DC: World bank, 2006. Alcohol Research & Health 27:39–51, 2003a. PmiD: 15301399

FlEgal, K.m.; Williamson, D.F.; anD gRaubaRD, b. i. using adjusted relative risks to calculate REhm, J.; KEhoE, t.; gmEl, g.; Et al. statistical modeling of volume of alcohol exposure for attributable fractions. American Journal of Public Health 96:398, 2006. PmiD: epidemiological studies of population health: The U S example. Population Health 16449574 Metrics 8:3, 2010b. PmiD: 20202213

guitaRt, a.m.; EsPElt, a.; CastEllano, y., Et al. impact of alcohol use disorder on mortality: REhm, J.; mathERs, C.; PoPoVa, s.; Et al. global burden of disease and injury and econom- are there age and gender differences? [article in spanish]. Gaceta Sanitaria ic cost attributable to alcohol use and alcohol use disorders. Lancet 373: 2223–2233, 25:385–390, 2011. PmiD: 21802796 2009. PmiD: 19560604

182 Alcohol Research: Current Reviews REhm, J.; PatRa, J.; anD PoPoVa, s. alcohol drinking cessation and its effect on shiElD, K.D.; RylEtt, m.; gmEl, g.; Et al. global alcohol exposure estimates by country, esophageal and head and neck cancers: a pooled analysis. International Journal of territory and region for 2005: a contribution to the Comparative Risk assessment for the Cancer 121:1132–1137, 2007. PmiD: 17487833 2010 global burden of Disease study. Addiction 108:912-922, 2013b. PmiD: 23347092

REhm, J.; REhn, n.; Room, R.; Et al. the global distribution of average volume of alcohol shiElD, K.D.; gmEl, g., JR.; PatRa, J.; anD REhm, J. global burden of injuries attributable to consumption and patterns of drinking. European Addiction Research 9:147–156, alcohol consumption in 2004: a novel way of calculating the burden of injuries attributable 2003b. PmiD: 12970583 to alcohol consumption. Population Health Metrics 10:9, 2012. PmiD: 22607112

REhm, J.; samoKhValoV, a.V.; anD shiElD, K.D. global burden of alcoholic liver diseases. shiElD, K. D.; PaRRy, C.; anD REhm, J. Chronic diseases and conditions related to alcohol Journal of Hepatology 59:160–168, 2013. PmiD: 23511777 use Alcohol Research: Current Reviews 35(2):155–173, 2013c.

shoJania, K.g.; buRton, E.C.; mCDonalD, K.m.; anD golDman, l. Changes in rates of autop- REhm, J., anD shiElD, K.D. alcohol drinking. in: international agency for Research on Cancer, ed. World Cancer Report 2013. lyon, France: international agency for sy-detected diagnostic errors over time: a systematic review. JAMA: Journal of the American Medical Association 289: 2849–2856, 2003. PmiD: 12783916 Research on Cancer, in Press. soRnPaisaRn, b.; shiElD, K.D.; CohEn, J.; Et al. Elasticity of alcohol consumption, alcohol- REhm, J.; shiElD, K.D.; gmEl,g.; Et al. modeling the impact of alcohol dependence on related harms, and drinking initiation in low- and middle-income countries: a systemat- mortality burden and the effect of available treatment interventions in the European ic review and meta-analysis. International Journal of Drug and Alcohol Research 2:45– union. European Neuropsychopharmacology 23:89–97, 2013b. PmiD: 22920734 58, 2013. REhm, J.; shiElD, K.D.; REhm, m.X.; Et al. Alcohol Consumption, Alcohol Dependence, and soRnPaisaRn, b.; shiElD, K.D.; anD REhm, J. alcohol taxation policy in thailand: implications Attributable Burden of Disease in Europe: Potential Gains from Effective Interventions for for other low- to middle-income countries. Addiction 107:1372–1384, 2012b. PmiD: Alcohol Dependence. toronto, Canada: Center for addiction and mental health, 2012. 22324742

Ehm ayloR ohaPatRa Et al R , J.; t , b.; m , s.; . alcohol as a risk factor for liver cirrhosis: a soRnPaisaRn, b.; shiElD, K.D.; anD REhm, J. two-chosen-one taxation: Examining its poten- systematic review and meta-analysis. Drug and Alcohol Review 29:437–445, 2010c. tial effectiveness to reduce drinking initiation and heavy alcohol consumption in low- to PmiD: 20636661 middle-income countries. Addiction 107:1389–1390, 2012b. PmiD: 22779414

RoCKhill, b.; nEWman, b.; anD WEinbERg, C. use and misuse of population attributable frac- tayloR, b.; iRVing, h. m.; KantEREs, F., Et al. the more you drink, the harder you fall: a sys- tions. American Journal of Public Health 88:15–19, 1998. PmiD: 9584027 tematic review and meta-analysis of how acute alcohol consumption and injury or colli- sion risk increase together. Drug and Alcohol Dependence 110:108–116, 2010. PmiD: RoERECKE, m., anD REhm, J. irregular heavy drinking occasions and risk of ischemic heart 20236774 disease: a systematic review and meta-analysis. American Journal of Epidemiology 171:633–644, 2010. PmiD: 20142394 tayloR, b., anD REhm, J. the relationship between alcohol consumption and fatal motor vehicle injury: high risk at low alcohol levels. Alcoholism: Clinical and Experimental RoERECKE, m., anD REhm, J. alcohol intake revisited: Risks and benefits. Current Research 36:1827–1834, 2012. PmiD: 22563862 Atherosclerosis Reports 14:556–562, 2012a. PmiD: 22864603 tiKKanEn, R.; holi, m.; linDbERg, n.; Et al. Recidivistic offending and mortality in alcoholic RoERECKE, m., anD REhm, J. the cardioprotective association of average alcohol consump- violent offenders: a prospective follow-up study. Psychiatry Research 168: 18–25, tion and ischaemic heart disease: a systematic review and meta-analysis. Addiction 2009. PmiD: 19467714 107:1246–1260, 2012b. PmiD: 22229788 WaltER, s.D. the estimation and interpretation of attributable risk in health research. Rothman, K.J.; gREEnlanD, s.; anD lash, t.l. modern Epidemiology. 3rd ed. Philadelphia, Biometrics 32:829–849, 1976. PmiD: 1009228 Pa: lippincott Williams & Wilkins, 2008. WaltER, s.D. Prevention of multifactorial diseases. American Journal of Epidemiology Rush, b. An Inquiry Into the Effects of Ardent Spirits Upon the Human Dody and Mind: 112:409–416, 1980. PmiD: 7424889 With an Account of the Means of Preventing, and of the Remedies for Curing Them. 8th Wang, h.; DWyER-linDgREn, l.; loFgREn, K.t.; Et al. age-specific and sex-specific mortality ed. Reprint. Exeter, nh: Richardson, 1785. in 187 countries, 1970-2010: a systematic analysis for the global burden of Disease saiEVa, C.; baRDazzi, g.; masala, g.; Et al. general and cancer mortality in a large cohort study 2010. Lancet 380:2071–2094, 2012. PmiD: 23245603 of italian alcoholics. alcoholism: Clinical and Experimental Research 36:342–350, WinDlE, m.; sPEaR, l.P.; Fuligni, a.J.; Et al. transitions into underage and problem drink- 2012. PmiD: 22085221 ing: Developmental processes and mechanisms between 10 and 15 years of age. sChmiDt, l.a.; mäKElä, P.; REhm, J.; Et al. alcohol: equity and social determinants. in: blas, Pediatrics 121(suppl 4:s273–s289, 2008. PmiD: 18381494 E., anD KuRuP, a.s. eds. Equity, Social Determinants and Public Health Programmes. World health organization. International Classification of Diseases and Related Health geneva, switzerland: World health organization, 2010. Problems, 10th Revision. geneva, switzerland: World health organization, 2007. shiElD, K.D.; gmEl, g.; KEhoE-Chan, t.; Et al. mortality and potential years of life lost World health organization. Global Status Report on Alcohol and Health. geneva, attributable to alcohol consumption by race and sex in the united states in 2005. PLoS switzerland: World health organization, 2011a. ONE 8:e51923, 2013a. PmiD: 23300957 World health organization. Global Status Report on Noncommunicable Diseases 2010. shiElD, K.D., anD REhm, J. Difficulties with telephone-based surveys on alcohol in high- Description of the Global Burden of NCDs, Their Risk Factors and Determinants. income countries: the Canadian example. International Journal of Methods in geneva, switzerland: World health organization, 2011b. Psychiatric Research 21:17–28, 2012. PmiD: 22337654 zatonsKi, W.a.; sulKoWsKa, u.; manCzuK, m.; Et al. liver cirrhosis mortality in Europe, with shiElD, K.D.; REhm, m.; PatRa, J.; Et al. global and country specific adult per capita con- special attention to Central and Eastern Europe. European Addiction Research sumption of alcohol, 2008. sucht 57:99–117, 2011. 16:193–201, 2010. PmiD: 20606444

Global Alcohol-Attributable Deaths 183 APiS: the NiAAA Alcohol Policy information System

Michael Hilton, Ph.D.

Michael Hilton, Ph.D. is the he Alcohol Policy Information and a “control” State that did not deputy director of the Division System (APIS) is an NIAAA- change its policy. As described in tsponsored Web site that provides other articles in this issue, measures of Epidemiology & Prevention detailed information on alcohol-related Research, National Institute on of policy outcomes can be taken public policies at both the State and from a number of existing data series Alcohol Abuse and Alcoholism, Federal levels. Updated annually, the that report rates of mortality, mor- Bethesda, Maryland. APIS information can be used to bidity, accidents, crimes, and the identify policy changes in 33 policy areas. Up to two-thirds of these policies like. APIS provides an accurate and can be tracked back to 1998, and consistently maintained source of the data on the remaining one-third are “independent variable” of interest in available since 2003. these studies, showing changes in APIS provides a tool for conducting codified statutes and regulations. research on the effectiveness of alcohol As the lead Federal agency policies in reducing alcohol-related responsible for research into alcohol harm. Natural experiments on the use and its related consequences, impact of a policy are possible when- NIAAA believes that providing this ever a State makes a change in some research tool is critically important. A aspect of alcohol policy. To conduct vigorous program of research on the such an experiment, scientists obtain a series of regularly collected data 1 it should be noted that enforcement efforts affect the effec- tiveness of any given policy. however, there is, at present, no on an outcome of interest, compare method of measuring state-by-state enforcement efforts that problem levels before and after a policy are both broadly recognized as valid by policy researchers or change, and report this comparison economically feasible to gather. although the institute encour- ages the development of better methods to measure enforce- for both the “experimental” State ment, aPis currently is not able to provide this information.

Figure the aPis Web site (www.alcohol policy.niaaa.nih.gov) features three basic displays

184 Alcohol Research: Current Reviews effects of various drinking-and-driving staff in such a way that consistent update (the default value) or as policies has been a key factor in decisions can be made across all of a specific date selected by the reducing alcohol-related traffic fatalities States, consistent decisions can viewer. Any given APIS policy to drivers, passengers, and pedestrians, be made across time as policies will vary across States in a number ages 16 to 20 by 50 percent since change, and the basis for the decisions of specific features. Understanding 1982. By providing information on can be clearly communicated to these details is important. other aspects of alcohol policy, APIS research users. can lead to similar advances in In addition to strict documenta- research knowledge and improve- tion, a hallmark of APIS has been care- • “Changes Over Time” contains ments in public health. ful attention to quality control. Each entries whenever a State changes APIS development began in new law or regulation is reviewed by a its policy. There is one row show- September of 2001, and the Web site minimum of two research lawyers and ing the policy before the change and another showing the policy went public in June of 2003. The 33 a senior social scientist. Any conflicting afterward. Should a State make existing policy topics cover underage interpretations are then resolved. On additional policy changes, even drinking, blood alcohol concentra- occasion, new developments in some more rows are added. tion limits, transportation, taxation, detail of a policy reveal a gap or an retail sales, alcohol control systems, ambiguity in the coding procedures that were used previously. In such • “Timeline of Policy Changes” pregnancy and alcohol, and health displays calendar years on the care services and financing. cases, that coding is reviewed and, if necessary, changes are made. Likewise, horizontal axis and States on the New topics are added periodically vertical axis. Icons indicate the when errors are encountered in APIS’s and are selected on the basis of the year (and quarter) when a policy legal descriptions, those corrections are likelihood that they will stimulate change went into effect. Note that made and documented in a “Change new research on policy outcomes. for all APIS information, policies Log.” Fortunately, the number of such Candidate policy topics should cover are categorized according to the instances has remained remarkably low. areas where recent policymaking date when they took effect rather Policy changes are reviewed each activity has occurred or where current than when they were enacted. discussions imply that impending year for each topic and for every change is likely. Additional criteria State. Shortly after the beginning of There are a number of other useful include (1) a scientifically grounded each calendar year, legal tools become Web site features and displays. The expectation that policy change in the available that make it possible to “Maps and Charts” feature shows particular candidate area is likely to identify all of the codified statutory graphic depictions of selected infor- produce measurable changes in some and regulatory activity that has mation. An “About This Policy” public health outcome; (2) sufficient occurred in a given State over the section gives useful background data are available to measure these preceding year. APIS legal researchers information on each policy. “State policy outcomes; (3) the legal research sift through these policies to see if Profiles of Underage Drinking Laws” involved in locating relevant statutes changes have occurred in any of the contains brief State-by-State summaries and regulations and in constructing 32 policy areas tracked by APIS. of policy in 12 topic areas specifically precise and consistent operational Typically, the most difficult aspect related to underage drinking. definitions that categorize meaningful of this work is “demonstrating the “Citations” are given so that the features of the laws is a manageable negative,” that is, verifying that there statutory or regulatory text of each task that falls within the resources has not been a change in some spe- policy covered by the Web site can available; and (4) case law or admin- cific aspect of a policy. This work is be located. Finally, there are links to istrative decision making that is not costly and has become the largest NIAAA’s current funding opportunity captured in statutory or regulatory part of the APIS contract budget. announcements, inviting researchers texts plays a limited role in interpreting The APIS Web site (www.alco- to submit proposals to study policy the policy. hol policy.niaaa.nih.gov) features impacts. NIAAA encourages all Documenting the decision rules three basic displays (see figure): researchers who are interested in for categorizing the features of State the effect of public policy on alcohol- policies is a critical concern. This • “Data on a Specific Date” shows related health outcomes to visit the documentation must capture the the features of a given policy as APIS Web site and to consider decisions made by the APIS research of the date of the last system-wide applying for research support.

APiS: the NiAAA Alcohol Policy information System 185 the Burden of Alcohol Use Factors Limiting Estimation of Alcohol Burden the Absence of Recent national Surveillance data Chief among the factors inhibiting the estimation of alcohol Focus on Children and burden in children and preadolescents is the absence of ongoing national surveillance data. The prevalence of child Preadolescents alcohol use can theoretically be estimated from either adoles- cents’ retrospective recall of their alcohol use in childhood or from survey research with children. Retrospective reports of the age at first drink, however, are not very reliable for this life stage. Typically, reported age of John E. Donovan, Ph.D. onset of alcohol use increases as a function of the age of the adolescents questioned (Bailey et al. 1992; Engels et al. 1997; the study of alcohol use by children ages 12 and younger Johnson et al. 1998; Labouvie et al. 1997; Parra et al. 2003). has been very limited. this article summarizes information For example, in the most recent national data from the 2009 from u.s. national and statewide surveys on the prevalence of Youth Risk Behavior Survey (YRBS), 28.1 percent of 9th alcohol use among children in grades 6 and lower, data on graders reported that they drank alcohol before age 13, com- health conditions wholly attributable to alcohol, the pared with 14.2 percent of 12th graders (Eaton et al. 2010). prevalence of children’s treatment admissions for alcohol These are not cohort effects but rather evidence of “forward abuse, and their rates of presentation at emergency telescoping,” as shown by the fact that although the percentages departments for acute alcohol intoxication. Factors at all grades have declined over time, a similar pattern can be hampering the estimation of alcohol burden in this seen in each of the previous YRBS surveys (1991–2007). population include the lack of ongoing national surveys of This pattern also is evident in the 1993–2010 national data alcohol use and problems in children, the hand-me-down from the Monitoring the Future (MTF) surveys (see figures nature of alcohol assessments in this population, and the 6 to 20 in Johnston et al. 2011): in every year, less than one- lack of studies to establish whether there is a causal half as many 12th graders as 8th graders report alcohol use relationship between childhood-onset drinking and morbidity initiation by grade 6. Thus, estimates based on retrospective and mortality in adolescence and later in life that would recall are problematic as a summary of the prevalence of permit determination of alcohol-attributable fractions. this childhood drinking. article concludes that although the alcohol burden in Direct surveys of children constitute a more appropriate childhood is low, it may be augmented by both referred approach for capturing normative data on child drinking. alcohol burden through parental drinking and alcohol abuse However, of the three major ongoing Federally sponsored and by deferred alcohol burden from longer-term national surveys in the United States—the annual MTF consequences of early use. kEY WoRDS: Alcohol consumption; survey, the biennial YRBS, and the annual National Survey alcohol use, abuse, and dependence; age of alcohol and on Drug Use and Health (NSDUH)—only the NSDUH other drug use onset; prevalence; alcohol burden; alcohol- includes children who are age 12, and none includes children attributable fractions; alcohol-related problems; alcohol younger than 12. According to the 2010 NSDUH results intoxication; alcohol poisoning; childhood; child; (Tables 2.15B and 2.16B in Substance Abuse and Mental preadolescent; youth; elementary school student; mortality; Health Services Administration [SAMHSA] 2011), 7.1 per- morbidity; survey; national surveillance data; health and cent of 12-year-olds had ever had a drink of alcohol (i.e., a disease; emergency care; treatment; underage drinking can of beer, a glass of wine, or a shot of liquor) in their life, 4.4 percent had a drink in the past year, 1.6 percent had a drink in the past month, and 0.4 percent had consumed five or more drinks on the same occasion. Despite the absence of children in these ongoing Federal surveillance studies, preliminary information on the prevalence he burden of alcohol use usually is expressed as a function of alcohol use in children has nevertheless been compiled of the contribution of alcohol use in a population to tmorbidity and mortality in that population (Rehm et al. 2010). It is difficult to calculate the burden of alcohol use for middle-school and high-school adolescents (see John E. Donovan, Ph.D., is associate professor of psychiatry Patrick and Schulembery, p. 193 in this issue) and nearly and epidemiology, Western Psychiatric Institute and Clinic, impossible to do so for children and preadolescents. There University of Pittsburgh Medical Center, Pittsburgh, are a number of reasons for this, most of which reflect the Pennsylvania. early stage of development of the research literature on alcohol use in this young population.

186 Alcohol Research: Current Reviews through a comprehensive search of internet sources to locate There are few current Nationwide data sources on the Nationwide and Statewide surveys of children in grades 6 prevalence of children’s experience of problems attributed and below (see Donovan 2007). Based on this review of the to alcohol use that could inform estimates of their wholly four Nationwide and seven Statewide datasets located, it is alcohol-attributable health conditions (i.e., alcohol dependence clear that a substantial number of children have had some and acute intoxication). Several community-level studies exposure to alcohol. Data from the cross-national Health suggest that rates of alcohol use disorders are close to zero Behaviour of School-Aged Children Study (Nic Gabhainn prior to adolescence (Cohen et al. 1993; Giaconia et al. and François 2000) indicate that in a 1998 national sample 1994; Sung et al. 2004). The low number of Nationwide of 11-year-old U.S. students, 62 percent of boys and 58 per- admissions for treatment of alcohol abuse at ages 10–12 cent of girls had ever tasted alcohol, 8 percent of boys and 7 bears this out (see figures 14 and 15 in SAMHSA 2008). percent of girls had consumed alcohol at least weekly, and 3 Patients under the age of 15 constitute just 0.5 percent of percent of both boys and girls had ever been drunk twice or those admitted for treatment of alcohol abuse alone and 0.7 more. According to the 1999 Partnership Attitude Tracking percent of those admitted primarily for treatment of alcohol Study (sponsored by the Partnership for a Drug-Free America), abuse who also had abused another drug (Table 3.2a in which surveyed a national probability sample of nearly 2,400 SAMHSA 2008). U.S. elementary-school students, 9.8 percent of 4th graders, Likewise, in contrast to adolescents, children rarely pre- 16.1 percent of 5th graders, and 29.4 percent of 6th graders sent at hospital emergency departments for acute intoxica- tion (alcohol poisoning). In 2009, the rate of visits to emer- had had more than just a sip of alcohol in their life. In gency departments for acute alcohol intoxication was 5.6 per 2000–2001, the National Survey of Parents and Youth 100,000 for U.S. children ages 0–5 and 1.0 per 100,000 for (NSPY) collected alcohol use information on 1,560 9- to children ages 6–11 versus 310.8 per 100,000 for adolescents 12-year-olds and found that 6.2 percent of 9-year-olds, 5.5 ages 12–17 (Drug Abuse Warning Network 2010). Of all percent of 10-year-olds, 9.2 percent of 11-year-olds, and calls to poison-control centers in the United States in 2009 15.5 percent of 12-year-olds had had more than a few sips involving children ages 5 or younger, 2.12 percent of cases of alcohol in their life. Data on alcohol use in the past year involved ingestion of alcohol (Bronstein et al. 2010). This (rather than lifetime) are reported annually by Pride Surveys probably is an underestimate, as many children ingested prod- (see www.pridesurveys.com): according to the 2009–2010 ucts such as cold medicines, cologne, perfume, aftershave, summary of school-district surveys performed across the and mouthwash that contain ethanol (see Vogel et al. 1995). United States, 4.0 percent of 4th graders, 4.8 percent of 5th In summary, there are few surveillance studies of alcohol graders, and 8.3 percent of 6th graders had drunk alcohol in use and alcohol-related problems among children and pread- the past year. Both the Nationwide and Statewide datasets olescents. The extant data indicate that although the rates of examined showed a decline in the prevalence of child drinking alcohol use are low in this population, substantial numbers over the past 10 years or so. The datasets located for this of children do have experience with alcohol and the rates of review, however, generally are either outdated or nonrepre- wholly alcohol-attributable health conditions are very low in sentative, and their limitations must be recognized in any this population. No evidence has been generated regarding attempt to estimate the burden of alcohol use in this population. the influence of child drinking on other diseases or injuries The absence of any recent national survey of alcohol use within childhood. among children argues for the need to institute ongoing Nationwide surveillance of this population. Problems of Measurement It is nevertheless evident, however, that the percentage A second major limitation for estimating alcohol burden in of children who have experience with alcohol decreases as this population is the widespread use of “hand-me-down” the intensity of alcohol involvement increases (from a sip or measures for the assessment of children’s alcohol use. taste to more than a few sips ever to use in the past year, past Measures originally developed for use with adults have been month, or past week), and that it differs as a function of modified for use with college students; then modified for grade, gender, and ethnicity (see Donovan 2007). Alcohol use with adolescents; and, finally, modified for assessment use rates increased with age, doubling between grades 4 and of children. Reliance on such hand-me-down assessments 6, with the largest jump in prevalence between grades 5 and has resulted, for instance, in only limited research into sipping 6. At each grade level, boys were more likely to have used and tasting of alcohol despite the fact that this is the most alcohol than girls. African-American children were nearly common form of children’s experience with alcohol (see as likely as white and Hispanic children to have used alcohol. Casswell 1996; Casswell et al. 1991; Donovan and Molina About one-third as many children reported having had more 2008; Johnson et al. 1997). than a sip of alcohol as reported having had only a sip. In The hand-me-down nature of child and adolescent general, around one-third of children who had ever used assessments is nowhere more evident than in the case of alcohol reported its use in the past year as well, and use in heavy episodic (binge) drinking, a major contributor to adult the past month occurred in only about one-third of those morbidity and mortality. In adults, binge drinking has been children who reported use in the past year. operationally defined as five or more drinks per occasion for

Children and Preadolescents 187 men and as four of more drinks per occasion for women (Wechsler et al. 1995); these levels of intake result in blood table 1 Recommended Cut Points (number of Drinks) for alcohol concentrations (BAC) of 0.08 percent (the legal defi- Developmentally appropriate Definition of binge Drinking in Children nition of intoxication) if consumed within a 2-hour window. and adolescents (Donovan 2009) Using these definitions for children and adolescents is inap- Age Boys Girls propriate, however, because they weigh less and thus have smaller volumes of total body water than adults. A recent 9–13 3+ 3+ analysis (Donovan 2009) modified the Widmark equation for estimating BAC so it would be more developmentally 14–15 4+ 3+ appropriate. This was done by incorporating formulas for estimating total body water that were derived from children 16–17 5+ 3+ and adolescents and by using ethanol elimination rates derived from child and adolescent presentations for acute alcohol intoxication at emergency departments. BAC estimates Barriers to Collecting Child data were calculated for intake of from one to five standard drinks for boys and girls separately at each age from 9 to 17 to Although monitoring the Nationwide prevalence of children’s determine how many drinks were required to result in alcohol use would constitute a step in the right direction, an estimated mean BAC of 0.08 percent or higher. Data increased research also is needed. It is possible that so few from more than 4,700 children and adolescents from the studies have been conducted in this area because of the per- 1999–2002 National Health and Nutrition Examination ception of several barriers to such research (see Donovan Survey were analyzed. Girls and boys ages 9–13 had mean 2007). One perceived barrier is that few children drink, so estimated BACs of 0.08 percent or higher at three or more there is little variation to explain. A second is the difficulty drinks, as did girls ages 14–17. Boys ages 14 and 15 had of gaining school-district approval to access elementary mean estimated BACs of 0.08 percent or higher at four or school populations, necessitating the use of targeted-age more drinks, and boys ages 16 and 17 reached this level at directory sampling or household enumeration sampling five or more drinks. Table 1 summarizes the resulting recom- methods. A third barrier sometimes raised is the misappre- mendations for defining binge drinking for children and hension that parents will be reluctant to consent to their adolescents by age and gender. Only boys ages 16 or 17 met children’s participation in alcohol research. the adult definition. In addition to the concern over hand-me-down assessments, there is a lack of consensus on the definition of the various Referred Childhood Alcohol Burden through levels of alcohol involvement for both children and adolescents. Parent Drinking As is evident in the summary of survey studies above, drinker status was defined variously as consumption of more than Parents contribute to the alcohol burden of their children a sip, more than a few sips, or a whole drink. This severely in a variety of ways. First, they model drinking behavior for hinders the performance of meta-analyses across studies and good or for ill. National surveys show that the majority the description of trends over time. Bacon (1976) noted a (87.9 percent) of adults in the United States ages 26 and similar lack of consensus 35 years ago. older have ever drank, 69.0 percent drank in the past year, In general, evidence from both test–retest examinations and 54.9 percent drank in the past month (Table 2.37B and collateral reports suggests that children’s self-reports of in SAMHSA 2011). Children learn about alcohol and its their alcohol use are as valid as adolescent self-reports (Dielman effects and usages from observing their parents drinking et al. 1995; Donovan et al. 2004). Given their typically low or from hearing their parents talk about their drinking, as levels of intake and the opportunistic nature of their drinking, well as from their exposure to drinking in the larger social misreporting in child reports of their alcohol involvement is environment (e.g., relatives, peers and their families, neigh- unlikely to reflect cognitive overload. More likely, difficulties borhood events, alcohol commercials on TV and radio, stem from a lack of familiarity with alcohol beverage types magazine ads, Internet Web sites, social media, and drinking (beer versus liquor, for example) and with estimation of in movies and even in animated feature films) (see Zucker drink volumes consumed. At least one recently developed et al. 2008, 2009). Children whose parents drink are more inventory uses pictorial images to assess alcohol and drug use likely to initiate early use (Donovan and Molina 2008, and their risk factors (see Andrews et al. 2003; Ridenour et 2011; Hawkins et al. 1997). al. 2009, 2011). Second, parents actively teach their children about alcohol. In addition to making child alcohol assessments more Children are first introduced to alcohol use by parents or developmentally appropriate and user friendly, surveillance other relatives in a family context (see Jackson 1997; Jahoda studies of child alcohol use need to be expanded to include and Cramond 1972; Johnson et al. 1997). Such precocious questions on the intensity and patterning of their current socialization into alcohol use can reflect either Old World alcohol use (e.g., frequency of use, usual and greatest intake, cultural beliefs regarding the role of alcohol as food or as a frequency of binge drinking, and contexts of use). necessary adjunct for celebrations or the belief that introducing

188 Alcohol Research: Current Reviews children to alcohol use as part of family dinners or events adulthood. Early onset of alcohol use predicts involvement serves to inoculate them from later involvement in problem in alcohol problems, alcohol abuse, and alcohol dependence drinking. Research has not yet established, however, whether in adolescence (Gruber et al. 1996; Hawkins et al. 1997; learning to drink in a family context actually serves to protect Horton, 2007; McGue and Iacono, 2005; Pederson and children from developing later alcohol problems. The relevant Skrondal, 1998; Warner et al. 2007). Early-onset drinking longitudinal research (Dielman et al. 1989; McMorris et al. also relates to a variety of other problematic outcomes in 2011; van der Vorst et al. 2010) suggests that this is not the adolescence, including absences from school, delinquent case: prior supervised drinking increases the likelihood of behavior, drinking and driving, smoking, marijuana and unsupervised drinking and more negative consequences later other illicit drug use, sexual intercourse, and pregnancy on. In addition, children who were permitted to drink alcohol (Ellickson et al. 2001; Gruber et al. 1996; McCluskey et al. at home have been found to show increased alcohol involve- 2002; Stueve and O’Donnell, 2005). ment and drunkenness over time (Jackson et al. 1999; Komro There also is evidence that early initiation of alcohol use et al. 2007). Research also shows that European adolescents, affects a number of outcomes in young adulthood as well. who are more often introduced to alcohol in family contexts, These young-adult outcomes include not only alcohol use typically are more likely to be involved in binge drinking disorder (e.g., Hingson et al. 2006; King and Chassin, 2007) and intoxication than U.S. adolescents of the same age but also prescription drug misuse (Hermos et al. 2008), (Currie et al. 2008; Friese and Grube 2010; Grube 2009). substance use disorders (Hingson et al. 2008; King and Third, the home environment is the most popular Chassin, 2007), employment problems (Ellickson et al. source of alcohol for children. Among 6th-grade students 2003), unintentional injuries (Hingson and Zha 2009; who had ever had alcohol, the largest percentage (32.7 percent) Hingson et al. 2000), and risky driving and drinking and obtained the alcohol from a parent or guardian the last time driving (Hingson et al. 2002; Zakrajsek and Shope 2006). they drank (Hearst et al. 2007). Other adults become a more Retrospective data from adults also have shown a relation- important source of alcohol than parents as children move into ship between earlier onset of drinking and lifetime experi- adolescence. Greater access to alcohol in the home and greater ence of an alcohol use disorders (e.g., DeWit et al. 2000; parental provision of alcohol are associated with greater alcohol Grant and Dawson 1997). Research currently is lacking, intake and problems later on (Komro et al. 2007; van den however, on whether early-onset drinking relates to psy- Eijnden et al. 2011). chosocial functioning in other young-adult life areas, such In addition to their direct impact on child drinking, as educational, occupational, marital, social, political, and parental drinking and alcohol abuse may increase child mor- community functioning, and relationship with parents. bidity and mortality through other means as well. Children As yet, there are few studies of the mechanisms linking also may be placed at increased risk through prenatal exposure early-onset drinking to young-adult alcohol problems and to maternal drinking (Jacobson and Jacobson 2002; Mattson other negative outcomes. McGue and Iacono (2008) see et al. 2001; Rasmussen 2005; Richardson et al. 2002; Streissguth this linkage as emanating from the interrelations between et al. 1999); through genetic inheritance of liabilities to early drinking and other problem behaviors in adolescence alcohol abuse and related addictive behaviors (Schuckit (Donovan and Jessor 1985) and the stability of this syndrome 1994; Sher 1991; Zucker et al. 2003); through alcohol- into young adulthood (Jessor et al. 1991), which is seen as impaired parenting, abuse, and neglect (Bijur et al. 1992; reflecting both inherited vulnerability and the influence of Dube et al. 2001; Kelleher et al. 1994); and through their early problem behavior on the selection of risky social envi- adoption of parent-socialized alcohol-specific intentions, ronments. Identification of such underlying mediating attitudes, and expectancies (e.g., Donovan et al. 2009; mechanisms is an important component of establishing any Handley and Chassin 2009; Tildesley and Andrews 2008), causal linkage between early-onset drinking and these later leading to both short-term and longer-term consequences. outcomes that would inform estimation of their alcohol- In addition, children are at risk of injury or death through attributable fractions (Rehm et al. 2010). The greater the riding in cars driven by an alcohol-impaired parent: in 2009 role of mediating variables in this pathway, the smaller the alone, 14 percent of the children ages 14 and younger killed alcohol-attributable fraction is likely to be. in traffic crashes were killed in alcohol-impaired driving crashes, and one-half of these children were passengers in vehicles driven by a driver with a BAC of 0.08 percent or Conclusions higher (U.S. Department of Transportation 2011). In summary, there are few surveillance studies of alcohol use and alcohol-related problems among children and preadoles- Deferred Childhood Alcohol Burden through cents, a situation that makes estimation of alcohol burden Long-term Consequences in this population problematic. The available data indicate that whereas the rates of alcohol use are relatively low in this The measurable burdens of child and preadolescent drinking population, substantial numbers of children do in fact have are for the most part postponed into adolescence and young experience with alcohol. With respect to wholly alcohol-

Children and Preadolescents 189 attributable health conditions, the available data suggest very DiElman, t.E.; lEECh, s.l.; anD loVElanD-ChERRy, C. Parents’ and children’s reports of parent- low levels of alcohol abuse and acute intoxication among ing practices and parent and child alcohol use. Drugs & Society 8(3/4):83–101, 1995. children. The scattered and inaccessible nature of much of DiElman, t.E.; shoPE, J.t.; lEECh, s.l.; anD butChaRt, a.t. Differential effectiveness of an this available data highlights the need for better ongoing elementary school-based alcohol misuse prevention program. Journal of School Health surveillance of this population. Although these direct assess- 59(6): 255–263, 1989. PmiD: 2770251 ments imply that alcohol burden in children is relatively low, DonoVan, J.E. Really underage drinkers: the epidemiology of children’s alcohol use in their alcohol burden is increased through the alcohol use and the united states. Prevention Science 8(3):192–205, 2007. PmiD: 17629790 abuse of their parents, and through the increased likelihood DonoVan, J.E. Estimated blood alcohol concentrations for child and adolescent drinking among early drinkers of alcohol problems and other negative and their implications for screening instruments. Pediatrics 123(6):e975–e981, 2009. outcomes in adolescence and young adulthood. ■ PmiD: 19482748 DonoVan, J.E., anD JEssoR, R. structure of problem behavior in adolescence and young adulthood. Journal of Consulting and Clinical Psychology 53(6):890–904, 1985. PmiD: Acknowledgements 4086689 DonoVan, J.E.; lEECh, s.l.; zuCKER, R.a., Et al. Really underage drinkers: alcohol use This research was supported by Grant No. AA–012342 from among elementary students. Alcoholism: Clinical and Experimental Research the National Institute on Alcohol Abuse and Alcoholism. 28(2):341–349, 2004. PmiD: 15112942

The author thanks Drs. Joel Grube and Duncan Clark DonoVan, J.E., anD molina, b.s. Childhood risk factors for early-onset drinking. Journal of for their help in locating several reports. Studies on Alcohol and Drugs 72(5):741–751, 2011. PmiD: 21906502

DonoVan, J.E.; molina, b.s.; anD KElly, t.m. alcohol outcome expectancies as socially shared and socialized beliefs. Psychology of Addictive Behaviors 23(2):248–259, Financial Disclosure 2009. PmiD: 19586141

DonoVan, J.E., anD molina, b.s.g. Children’s introduction to alcohol use: sips and tastes. The author declares that he has no competing financial Alcoholism: Clinical and Experimental Research 32(1):108–119, 2008. PmiD: 18070249 interests. Drug abuse Warning network. National Estimates of Drug-Related Emergency Department Visits, 2004–2009. Rockville, mD: Center for behavioral health statistics and Quality, substance abuse and mental health services administration, 2010. References available at: http://www.samhsa.gov/data/DaWn.aspx. accessed may 14, 2012.

anDREWs, J.a.; tilDEslEy, E.; hoPs, h.; Et al. Elementary school age children’s future inten- DubE, s.R.; anDa, R.F.; FElitti, V.J.; Et al. growing up with parental alcohol abuse: tions and use of substances. Journal of Clinical Child and Adolescent Psychology Exposure to childhood abuse, neglect, and household dysfunction. Child Abuse & 32(4):556–567, 2003. PmiD: 14710464 Neglect 25(12):1627–1640, 2001. PmiD: 11814159

baCon, s.D. Defining adolescent alcohol use: implications for a definition of adolescent Eaton, D.K.; Kann, l.; KinChEn, s.; Et al. youth risk behavior surveillance—united states, alcoholism. Journal of Studies on Alcohol 37(7):1014–1019, 1976. 2009. Morbidity and Mortality Weekly Report: Surveillance Summaries 59 (5):1–142, 2010. PmiD: 20520591 bailEy, s.l.; FlEWElling, R.l.; anD RaChal, J.V. the characterization of inconsistencies in self-reports of alcohol and marijuana use in a longitudinal study of adolescents. ElliCKson, P.l.; tuCKER, J.s.; anD KlEin, D.J. ten-year prospective study of public health Journal of Studies on Alcohol 53(6):636–647, 1992. PmiD: 1434637 problems associated with early drinking. Pediatrics 111(5 Pt. 1):949–955, 2003. PmiD: 12728070 biJuR, P.E.; KuRzon, m.; oVERPECK, m.D.; anD sChEiDt, P.C. Parental alcohol use, problem drinking, and children’s injuries. JAMA: Journal of the American Medical Association ElliCKson, P.l.; tuCKER, J.s.; KlEin, D.J.; anD mCguigan, K.a. Prospective risk factors for 267(23):3166–3171, 1992. PmiD: 1593737 alcohol misuse in late adolescence. Journal of Studies on Alcohol 62(6):773–782, 2001. PmiD: 11838914 bRonstEin, a.C., sPyKER, D.a., CantilEna, l.R., JR.; Et al. 2009 annual Report of the american association of Poison Control Centers’ national Poison Data system (nPDs): EngEls, R.C.; KnibbE, R.a.; anD DRoP, m.J. inconsistencies in adolescents’ self-reports of 27th annual Report. Clinical Toxicology 48(10):979–1178, 2010. PmiD: 21192756 initiation of alcohol and tobacco use. Addictive Behaviors 22(5):613–623, 1997. PmiD: 9347063 CassWEll, s. alcohol use: growing up and learning about drinking: Children in Dunedin in the 1980s. in silva, P.a. and stanton, W.R., Eds. From Child to Adult: The Dunedin FRiEsE, b.m., anD gRubE, J.W. Youth Drinking Rates and Problems: A Comparison of Multidisciplinary Health and Development Study. auckland, new zealand: oxford European Countries and the United States. Rockville, mD: office of Juvenile Justice and university Press, 1996, pp. 206–224. Delinquency Prevention, office of Justice Programs, u.s. Department of Justice, 2010.

CassWEll, s.; stEWaRt, J.; Connolly, g.; anD silVa, P. a longitudinal study of new zealand available at: http://www.udetc.org/documents/youthDrinkingRatesandProblems.pdf. children’s experience with alcohol. British Journal of Addiction 86(3):277–285, 1991. accessed may 14, 2012.

PmiD: 2025690 giaConia, R.m.; REinhERz, h.z.; silVERman, a.b.; Et al. age of onset of psychiatric disorders

CohEn, P.; CohEn, J.; KasEn, s.; Et al. an epidemiological study of disorders in late child- in a community population of older adolescents. Journal of the American Academy of hood and adolescence: i. age- and gender-specific prevalence. Journal of Child Child and Adolescent Psychiatry 33(5):706–717, 1994. PmiD: 8056734 Psychology and Psychiatry 34(6):851–867, 1993. PmiD: 8408371 gRant, b.F., anD DaWson, D.a. age at onset of alcohol use and its association with Dsm- CuRRiE, C.; niC gabhainn, s.; goDEau, E.; Et al., EDs. Inequalities in Young People’s Health: iV alcohol abuse and dependence: Results from the national longitudinal alcohol Health Behaviour in School-aged Children International Report from the 2005/2006 Epidemiologic survey. Journal of Substance Abuse 9(2):103–110, 1997. PmiD: Survey. Copenhagen, Denmark: Who Regional office for Europe, 2008, pp. 131–134. 9494942 available at: http://www.hbsc.org/publications/international/. accessed may 14, 2012. gRubE, J.W. Environmental approaches to preventing drinking and drinking problems DEWit, D.J.; aDlaF, E.m.; oFFoRD, D.R.; anD ogboRnE, a.C. age of first alcohol use: a risk among youth. in scheier, l., Ed. Handbook of Drug Use Etiology: Theory, Methods, and factor for the development of alcohol disorders. American Journal of Psychiatry Empirical Findings. Washington, DC: american Psychological association, 2009, pp. 157(5):745–750, 2000. PmiD: 10784467 493–509.

190 Alcohol Research: Current Reviews gRubER, E.; DiClEmEntE, R.J.; anDERson, m.m.; anD loDiCo, m. Early drinking onset and its Students. ann arbor, mi: institute for social Research, 2011. available at: association with alcohol use and problem behavior in late adolescence. Preventive http://www.monitoringthefuture.org. accessed may 14, 2012. Medicine 25(3):293–300, 1996. PmiD: 8781007 KEllEhER, K.; ChaFFin, m.; hollEnbERg, J.; anD FisChER, E. alcohol and drug disorders among hanDlEy, E.D., anD Chassin, l. intergenerational transmission of alcohol expectancies in a physically abusive and neglectful parents in a community-based sample. American high-risk sample. Journal of Studies on Alcohol and Drugs 70(5):675–682, 2009. Journal of Public Health 84(10):1586–1590, 1994. PmiD: 7943475 PmiD: 19737491 King, K.m., anD Chassin, l. a prospective study of the effects of age of initiation of alco- haWKins, J.D.; gRaham, J.W.; maguin, E.; Et al. Exploring the effects of age of alcohol use hol and drug use on young adult substance dependence. Journal of Studies on Alcohol initiation and psychosocial risk factors on subsequent alcohol misuse. Journal of and Drugs 68(2):256–265, 2007. PmiD: 17286344 Studies on Alcohol 58(3):280–290, 1997. PmiD: 9130220 KomRo, K.a.; malDonaDo-molina, m.m.; toblER, a.l.; Et al. Effects of home access and hEaRst, m.o.; FulKERson, J.a.; malDonaDo-molina, m.m.; Et al. Who needs liquor stores availability of alcohol on young adolescents’ alcohol use. Addiction when parents will do? the importance of social sources of alcohol among young urban 102(10):1597–1608, 2007. PmiD: 17854336 teens. Preventive Medicine 44(6):471–476, 2007. PmiD: 17428525 labouViE, E.; batEs, m.E.; anD PanDina, R.J. age of first use: its reliability and predictive hERmos, J.a.; WintER, m.R.; hEEREn, t.C.; anD hingson, R.W. Early age-of-onset drinking utility. Journal of Studies on Alcohol 58(6):638–643, 1997. PmiD: 9391924 predicts prescription drug misuse among teenagers and young adults: Results from a national survey. Journal of Addiction Medicine 2(1):22–30, 2008. PmiD: 21768969 mattson, s.n.; sChoEnFiElD, a.m.; anD RilEy, E.P. teratogenic effects of alcohol on brain and behavior. Alcohol Research & Health 25(3):185–191, 2001. PmiD: 11810956 hingson, R.W., anD zha, W. age of drinking onset, alcohol use disorders, frequent heavy drinking, and unintentionally injuring oneself and others after drinking. Pediatrics mCClusKEy, C. P.; KRohn, m.D.; lizottE, a.J.; anD RoDRiQuEz, m.l. Early substance use and 123(6):1477–1484, 2009. PmiD: 19482757 school achievement: an examination of latino, White, and african american youth. Journal of Drug Issues 32(2):921–944, 2002. hingson, h.W.; hEEREn, t.; anD EDWaRDs, E.m. age at drinking onset, alcohol dependence, and their relation to drug use and dependence, driving under the influence of drugs, mCguE, m., anD iaCono, W.g. the association of early adolescent problem behavior with and motor-vehicle crash involvement because of drugs. Journal of Studies on Alcohol adult psychopathology. American Journal of Psychiatry 162(6):1118–1124, 2005. and Drugs 69(2):192–201, 2008. PmiD: 18299759 PmiD: 15930060 hingson, R.; hEEREn, t.; JamanKa, a.; anD hoWlanD, J. age of drinking onset and uninten- mCguE, m., anD iaCono, W.g. the adolescent origins of substance use disorders. tional injury involvement after drinking. JAMA: Journal of the American Medical International Journal of Methods in Psychiatric Research 17(suppl. 1):s30–s38, 2008. Association 284(12):1527–1533, 2000. PmiD: 11000646 PmiD: 18543360 hingson, R.; hEEREn, t.; lEVEnson, s.; Et al. age of drinking onset, driving after drinking, mCmoRRis, b.J.; Catalano, R.F.; Kim, m.J.; Et al. influence of family factors and supervised and involvement in alcohol-related motor-vehicle crashes. Accident: Analysis and alcohol use on adolescent alcohol use and harms: similarities between youth in differ- Prevention 34(1):85–92, 2002. PmiD: 11789578 ent alcohol policy contexts. Journal of Studies on Alcohol and Drugs 72(3): 418–428, 2011. PmiD: 21513678 hingson, R.W.; hEEREn, t.; anD WintER, m.R. age at drinking onset and alcohol depen- dence: age at onset, duration, severity. Archives of Pediatrics & Adolescent Medicine niC gabhainn, s., anD FRançois, y. substance use. in: Currie, C., hurrelmann, K., 160(7):739–746, 2006. PmiD: 16818840 settertobulte, W., et al., Eds. Health Behaviour in School-Aged Children: A WHO Cross- National Study (HBSC) International Report. Copenhagen, Denmark: World health hingson, R.; hEEREn, t.; anD zaKoCs, R. age of drinking onset and involvement in physical organization Regional office for Europe, 2000. available at hwwp://www.hbsc.org/pub- fights after drinking. Pediatrics 108(4):872–877, 2001. PmiD: 11581438 lications/international/. accessed may 14, 2012.

hoRton, E.g. Racial differences in the effects of age of onset on alcohol consumption olson, h.C.; samPson, P.D.; baRR, h.; Et al. Prenatal exposure to alcohol and school prob- and development of alcohol-related problems among males from mid-adolescence to lems in late childhood: a longitudinal prospective study. Development and young adulthood. Journal of Ethnicity in Substance Abuse 6(1):1–13, 2007. PmiD: Psychopathology 4(3):341–359, 1992. 17430813 PaRRa, g.R.; o’nEill, s.E.; anD shER, K.J. Reliability of self-reported age of substance JaCKson, C. initial and experimental stages of tobacco and alcohol use during late child- involvement onset. Psychology of Addictive Behaviors 17(3):211–218, 2003. PmiD: hood: Relation to peer, parent, and personal risk factors. Addictive Behaviors 14498815 22(5):685–698, 1997. PmiD: 9347070 PEDERsEn, W., anD sKRonDal, a. alcohol consumption debut: Predictors and conse- JaCKson, C.; hEnRiKsEn, l.; anD DiCKinson, D. alcohol-specific socialization, parenting quences. Journal of Studies on Alcohol 59(1):32–42, 1998. PmiD: 9498313 behaviors and alcohol use by children. Journal of Studies on Alcohol 60(3):362–367, 1999. PmiD: 10371264 RasmussEn, C. Executive functioning and working memory in fetal alcohol spectrum dis- order. Alcoholism: Clinical and Experimental Research 29(8):1359–1367, 2005. PmiD: JaCobson, J.l., anD JaCobson, s.W. Effects of prenatal alcohol exposure on child develop- 16131842 ment. Alcohol Research & Health 26(4):282–286, 2002. PmiD: 12875038 REhm, J; baliunas, D; boRgEs, g.l.; Et al. the relation between different dimensions of JahoDa, g., anD CRamonD, J. Children and Alcohol: A Developmental Study in Glasgow. alcohol consumption and burden of disease: an overview. Addiction 105(5):817–843, london: her majesty’s stationery office, 1972. 2010. PmiD: 20331573 JEssoR, R.; DonoVan, J.E.; anD Costa, F.m. Beyond Adolescence: Problem Behavior and RiChaRDson, g.a., Ryan, C., WillFoRD, J., Et al. Prenatal alcohol and marijuana exposure: Young Adult Development. new york: Cambridge university Press, 1991. Effects on neuropsychological outcomes at 10 years. Neurotoxicology and Teratology Johnson, C.C.; gREEnlunD, K.J.; WEbbER, l.s.; anD bEREnson, g.s. alcohol first use and atti- 24(3):309–320, 2002. PmiD: 12009486 tudes among young children. Journal of Child and Family Studies 6(3):359–372, RiDEnouR, t.a.; ClaRK, D.b.; anD CottlER, l.b. the illustration-based assessment of liability 1997. and EXposure to substance use and antisocial behavior© for children. American Johnson, R.a.; gERstEin, D.R.; anD RasinsKi, K.a. adjusting survey estimates for response Journal of Drug and Alcohol Abuse 35(4):242–252, 2009. PmiD: 20180677 bias: an application to trends in alcohol and marijuana use. Public Opinion Quarterly RiDEnouR, t.a.; minnEs, s.; malDonaDo-molina, m.m.; Et al. Psychometrics and cross-cultur- 62(3):354–377, 1998. al comparisons of the illustration-based assessment of liability and Exposure to Johnston, l.D.; o’malley, P.m.; bachman, J.g.; and schulenberg, J.E. Monitoring the substance use and antisocial behavior© for children. Open Family Studies Journal Future National Survey Results on Drug Use, 1975–2010: Volume I. Secondary School 4(suppl. 1–m2):17–26, 2011.

Children and Preadolescents 191 sChuCKit, m.a. a clinical model of genetic influences in alcohol dependence. Journal of Van DEn EiJnDEn, R.; Van DE mhEEn, D.; VEt, R.; anD VERmulst, a. alcohol-specific parenting Studies on Alcohol 55(1):5–17, 1994. PmiD: 8189726 and adolescents’ alcohol-related problems: the interacting roles of alcohol availability and parental rules. Journal of Studies on Alcohol and Drugs 72(3):408–417. 2011. shER, K.J. Children of Alcoholics: A Critical Appraisal of Theory and Research. Chicago, PmiD: 21513677 il: university of Chicago Press, 1991. Van DER VoRst, h.; EngEls, R.C.; anD buRK, W.J. Do parents and best friends influence the stREissguth, a.P.; baRR, h.m.; booKstEin, F.l.; Et al. the long-term neurocognitive conse- normative increase in adolescents’ alcohol use at home and outside the home? quences of prenatal alcohol exposure: a 14-year study. Psychological Science Journal of Studies on Alcohol and Drugs 71(1):105–114, 2010. PmiD: 20105420 10(3):186–190, 1999. VogEl, C.; CaRaCCio, t.; moFEnson, h.; anD haRt, s. alcohol intoxication in young children. stuEVE, a., anD o’DonnEll, l.n. Early alcohol initiation and subsequent sexual and alco- Journal of Toxicology. Clinical Toxicology 33(1):25–33, 1995. PmiD: 7837310 hol risk behaviors among urban youths. American Journal of Public Health WaRnER, l.a.; WhitE, h.R.; anD Johnson, V. alcohol initiation experiences and family histo- 95(5):887–893, 2005. PmiD: 15855470 ry of alcoholism as predictors of problem-drinking trajectories. Journal of Studies on substance abuse and mental health services administration (samhsa). Results from Alcohol and Drugs 68(1):56–65, 2007. PmiD: 17149518 the 2010 National Survey on Drug Use and Health: Summary of National Findings WEChslER, h.; DoWDall, g.W.; DaVEnPoRt, a.; anD Rimm, E.b. a gender-specific measure of (nsDuh series h–41, hhs Publication no. sma 11–4658). Rockville, mD: Center for binge drinking among college students. American Journal of Public Health behavioral health statistics and Quality, samhsa, 2011. 85(7):982–985, 1995. PmiD: 7604925

samhsa. Treatment Episode Data Set (TEDS): 1996–2006. National Admissions to zaKRaJsEK, J.s.; anD shoPE, J.t. longitudinal examination of underage drinking and sub- Substance Abuse Treatment Services (Dasis series: s–43, Dhhs Publication no. sma sequent drinking and risky driving. Journal of Safety Research 37(5):443–451, 2006. 08–4347). Rockville, mD: office of applied studies, samhsa, 2008. PmiD: 17123546

zuCKER, R.a.; DonoVan, J.E.; mastEn, a.s.; Et al. Early developmental processes and the sung, m.; ERKanli, a.; angolD, a.; anD CostEllo, E.J. Effects of age at first substance use continuity of risk for underage drinking and problem drinking. Pediatrics 121(suppl. and psychiatric comorbidity on the development of substance use disorders. Drug and 4):s252–s272, 2008. PmiD: 18381493 Alcohol Dependence 75(3):287–299, 2004. PmiD: 15283950 zuCKER, R.a.; DonoVan, J.E.; mastEn, a.s.; Et al. Developmental processes and mecha- tilDEslEy, E.a.; anD anDREWs, J.a. the development of children’s intentions to use alco- nisms: ages 0–10. Alcohol Research & Health 32(1):16–29, 2009. hol: Direct and indirect effects of parent alcohol use and parenting behaviors. Psychology of Addictive Behaviors 22(3):326–339, 2008. PmiD: 18778126 zuCKER, R.a.; Wong, m.m.; PuttlER, l.i.; anD FitzgERalD, h.E. Resilience and vulnerability among sons of alcoholics: Relationship to developmental outcomes between early u.s. Department of transportation. Traffic Safety Facts, 2009 Data: Children. Washington, childhood and adolescence. in luthar, s., Ed. Resilience and Vulnerability: Adaptation DC: national highway traffic safety administration, 2011. Dot hs 811 387. available at: in the Context of Childhood Adversities. new york: Cambridge university Press, 2003, http://www-nrd.nhtsa.dot.gov/Pubs/811387.pdf. accessed may 14, 2012. pp. 76–103.

192 Alcohol Research: Current Reviews the Prevalence of Drinking and Historical Prevalence and Predictors Changes As is true for adults, alcohol is the most commonly abused of Adolescent Alcohol Use substance among American youth. The MTF study has been documenting the prevalence and trends in alcohol use and Binge Drinking in the frequency and binge drinking (i.e., consumption of five or more drinks in a row) for the past several decades in annual, national samples of American 8th-, 10th-, and 12th-grade United States students. Using these data, Patrick and Schulenberg (2010) found that very few 8th- and 10th-grade students who reported having ever used alcohol had not used alcohol in the past year, suggesting that most of the alcohol use reported is relatively recent. Therefore, this article focuses Megan E. Patrick, Ph.D., and John E. Schulenberg, Ph.D. on alcohol use in the past 12 months and the past 30 days, as well as self-reported drunkenness in the past 30 days and because alcohol use typically is initiated during adolescence binge drinking in the past 2 weeks. The prevalence figures and young adulthood and may have long-term consequences, for these variables for 2011 are summarized in table 1, the monitoring the Future (mtF) study annually assesses various broken down by grade level, gender, and racial/ethnic measures of alcohol use among 8th-, 10th-, and 12th-grade subgroups (for more information, see Johnston et al. 2012). students. these analyses have found that although alcohol use In 2011, 27 percent of 8th graders, 50 percent of 10th among these age groups overall has been declining since graders, and 64 percent of 12th graders reported having used 1975, levels remain high. thus, in 2011 about one-quarter of alcohol in the past 12 months. The corresponding rates for 8th graders, one-half of 10th graders, and almost two-thirds of alcohol use in the past 30 days were 13 percent, 27 percent, 12th graders reported drinking alcohol in the month preceding and 40 percent, respectively. Furthermore, 4 percent of 8th the interview. binge drinking (i.e., consumption of five or more graders, 14 percent of 10th graders, and 25 percent of 12th drinks in a row) was also prevalent. specific rates of drinking, graders reported having been drunk within the past month. binge drinking, and getting drunk varied among different Finally, binge drinking in the past 2 weeks was reported by student subgroups based on gender and race/ethnicity. the mtF 6 percent of 8th graders, 15 percent of 10th graders, and 22 study has also identified numerous factors that influence the risk percent of 12th graders. of alcohol use among adolescents, including parents and Interestingly, it is more common for students to report peers, school and work, religiosity and community attachment, binge drinking 2 or more times in the past 2 weeks than to exercise and sports participation, externalizing behavior and report binge drinking only once in the past 2 weeks; thus, other drug use, risk taking and sensation seeking, well-being, 61 percent of 8th graders and 62 percent of 10th graders and drinking attitudes and reasons for alcohol use. Drinking who had engaged in binge drinking in the previous 2 weeks during adolescence can have long-term effects on a person’s did so on multiple occasions (Patrick and Schulenberg 2010). life trajectory. therefore, these findings have broad implications This observation suggests a fast shift to frequent heavier for prevention and intervention efforts with this population. drinking for many young people. In addition, the surveys kEY WoRDS: Underage drinking; binge drinking; adolescent; indicate that extreme binge drinking (i.e., consumption high school student; young adult; prevalence; predictors; of 10 or more or 15 or more drinks in a row) is a problem causes of alcohol and other drug use; risk factors; school risk among 12th graders (this variable was not assessed among factors; environmental risk factors; family risk factors; peer 8th and 10th graders). Thus, 10.5 percent of high school risk factors; gender differences; racial/ethnic differences; seniors reported consuming 10 or more drinks in a row, and Monitoring the Future (MtF) Study; United States 5.6 percent reported consuming 15 or more drinks in a row in the past 2 weeks (Patrick et al. 2013). Alcohol use differs not only by age but also by demo- graphic subgroups, including gender and race/ethnicity n the United States, alcohol use typically begins and escalates during adolescence and young adulthood. To describe the ihistorical and developmental trends in substance use in this age group, the Monitoring the Future (MTF) study Megan E. Patrick, Ph.D., is a research assistant professor (Johnston et al. 2012) was designed in 1975. Since then, at the Institute for Social Research, and John E. Schulenberg, this ongoing national-cohort sequential longitudinal study Ph.D., is professor in the Department of Psychology and assessing the epidemiology and etiology of substance use among adolescents and adults has been funded by the research professor at the Institute for Social Research, National Institute on Drug Abuse (NIDA). This article University of Michigan, Ann Arbor, Michigan. summarizes findings from the MTF study regarding the prevalence and predictors of alcohol use during adolescence.

Prevalence and Predictors of Adolescent Alcohol Use and Binge Drinking in the United States 193 (see table 1). In 8th grade, girls tend to have somewhat higher White and African American youth. By 12th grade, however, rates of alcohol use (i.e., 13 percent in the past 30 days) than White adolescents have the highest prevalence levels of do boys (12 percent). Among older students, however, this the three racial/ethnic groups on all alcohol use measures, ratio is reversed, with 38 percent of female and 42 percent of African American adolescents have the lowest levels, and male 12th graders reporting alcohol use in the past 30 days. Hispanics have intermediate levels. For example, for binge- This gender difference continues into adulthood, with men drinking, prevalence rates among 12th graders are 26 percent consistently using alcohol at higher rates compared with for Whites, 11 percent for African Americans, and 21 women (Johnston et al. 2012; Wilsnack et al. 2000). A similar percent for Hispanics. Some, but not all, of these race/ interaction seems to exist between grade and race/ethnicity ethnicity differences in alcohol use among 12th graders are (Wallace et al. 2003). Thus, among 8th graders, Hispanic attributable to differential high-school dropout rates among youth tend to report a greater prevalence of alcohol consumption the different groups. Thus, dropout rates tend to be higher in the last 12 months (36 percent) or last 30 days (18 percent), among racial/ethnic minority youth, and alcohol and other as well as of being drunk in the last 30 days (6 percent) and drug (AOD) use tends to be higher among school dropouts binge drinking in the past 2 weeks (10 percent) than do both than among those staying in school (Bachman et al. 2008).

table 1 Prevalence of alcohol use (%) by Demographic subgroups in 8th, 10th, and 12th graders, 2011

Any Use in Any Use in Been Drunk in 5+ Drinks in a Row Past 12 Months Past 30 Days Past 30 Days in P ast 2 Weeks

8th Graders total 26.9 12.7 4.4 6.4 gender boys 26.2 12.1 4.4 6.1 girls 27.1 12.8 4.2 6.5 Race/Ethnicity* White 26.2 12.3 4.7 6.2 african american 26.2 11.6 2.9 5.1 hispanic 36.0 18.0 5.6 10.4

10th Graders total 49.8 27.2 13.7 14.7 gender boys 49.1 28.2 14.9 16.5 girls 50.3 26.0 12.4 12.7 Race/Ethnicity White 52.1 29.1 15.6 16.1 african american 43.6 20.8 8.3 9.4 hispanic 54.8 31.8 13.8 19.7

12th Graders total 63.5 40.0 25.0 21.6 gender boys 63.3 42.1 27.5 25.5 girls 63.5 37.5 22.0 17.6 Race/Ethnicity White 66.8 43.8 29.9 25.9 african american 55.2 30.1 14.2 11.3 hispanic 65.3 39.7 20.0 20.8

*to derive percentages for each racial subgroup, data for the specified year and the previous year were combined to increase subgroup sample sizes and thus provide more stable estimates. notE: For 8th graders, the approximate weighted N is 16,000. For 10th graders, the approximate weighted N is 14,900. For 12th graders, the approximate weighted N is 14,100. souRCE: the Monitoring the Future study, the university of michigan.

194 Alcohol Research: Current Reviews Overall, alcohol use among adolescents and young adults and race/ethnicity (e.g., Brown et al. 2001; Donovan et al. has been declining to the lowest levels in recent decades, as 1999; Patrick and Schulenberg 2010). Like many other shown by the trends in self-reported alcohol use in the past large-scale studies on adolescent AOD use, the MTF study 12 months and drunkenness in the past 30 days (see figure has cast a wide net in terms of risk and protective factors, 1) (Johnston et al. 2012; Patrick and Schulenberg 2010). correlates, and consequences of substance use. Not only is Similar trends have been observed for alcohol use in the this approach well suited to placing alcohol use within the past 30 days and binge drinking in the past 2 weeks. These larger context of adolescent development, it makes good historical shifts in AOD use can be attributed to multiple use of the MTF large-scale survey approach that emphasizes influences. For example, changes in the minimum legal drinking age (e.g., Wagenaar et al. 2001) as well as in perceived breadth of measurement. Conceptually, the analyses drew social norms (e.g., Keyes et al. 2012) have been shown to from broad multidomain models when examining causes, contribute to changes in alcohol use. Of particular interest correlates, and outcomes of adolescent alcohol use (e.g., are historical shifts that relate to changes in developmental Brown et al. 2009; Chassin et al. 2009; Maggs and Schulenberg trajectories. Latent growth modeling analyses with multicohort 2005). This section summarizes MTF study findings con- data have demonstrated that, compared with earlier cohorts, cerning several domains of predictors of AOD use during more recent cohorts exhibit lower initial levels of binge adolescence, after considering methodological issues when drinking but more rapid increases from age 18 to young examining causes and consequences of adolescent alcohol use. adulthood (Jager et al. 2013). This acceleration of alcohol use helps explain the findings that use among adolescents Methodological Issues in understanding Risk Factors has been decreasing at faster rates than among young adults for and Consequences of Adolescent Alcohol use in recent decades. When considering the correlates of AOD use, any attempt Predictors of Alcohol Use Among Adolescents to discern whether these correlates are causes or conse- quences of substance use is hampered by three factors: Despite the changes in alcohol use that have occurred over the past three decades, the relevant risk and protective factors • Firm conclusions about causal connections are difficult tend to remain very stable across historic time, age, gender, without randomly controlled experiments.

100

90

80 8th grade: past 12 months

70 10th grade: past 12 months

60 12th grade: past 12 months

50 8th grade: been drunk in past 30 days Percent 40 10th grade: been drunk in past 30 days

30 12th grade: been drunk in past 30 days

20

10

0 ‘91 ‘93 ‘95 ‘97 ‘99 ‘01 ‘03 ‘05 ‘07 ‘09 ‘11 Year

Figure trends in alcohol use in the past 12 months and in having been drunk in the past 30 days for 8th, 10th, and 12th graders, 1991–2011.

Prevalence and Predictors of Adolescent Alcohol Use and Binge Drinking in the United States 195 • Alcohol use during adolescence typically is reciprocally predictors (Dever et al. 2012; Pilgrim et al. 2006). Of particular related to risk factors across development, such that prob- importance, this effect was equally important (i.e., invariant) lems that contribute to alcohol use may get worse with across gender and race/ethnicity (Pilgrim et al. 2006). Further - continued alcohol use (e.g., Cairns and Cairns, 1994; more, parental monitoring was especially protective against Dodge et al. 2009; Schulenberg and Maslowsky 2009). substance use for high-risk–taking adolescents (Dever et al. 2012). • Factors that are identified as causes or as consequences of The literature for decades has indicated that peer use alcohol use during adolescence in the total sample likely is one of the strongest correlates of AOD use. This was do not apply to all young people, given the heterogeneity confirmed in the MTF; thus, in an analysis of multiple in developmental course (Schulenberg 2006). predictors of binge drinking among 8th and 10th graders from 1991 to 2007, having friends who get drunk was the Cross-sectional studies, in which each individual is strongest risk factor, regardless of the grade level or cohort evaluated only once, typically provide little leverage for analyzed (Patrick and Schulenberg 2010). Moreover, friends’ concluding whether a given construct is a cause, correlate, alcohol use in high school predicted both concurrent binge or consequence of alcohol use, emphasizing the importance drinking and future trajectories of binge drinking (Schulenberg of conceptual guidance, logic, and statistical controls. et al. 1996). Overall, the frequency of evenings out with Furthermore, when adolescents report using multiple substances, friends (unsupervised by adults) was associated with more it is difficult to determine whether they are using the drugs AOD use (Bachman et al. 2008; Brown et al. 2001; Patrick simultaneously or whether use of one substance leads to use and Schulenberg 2010). Of course, a central issue when eval- of another. Longitudinal panel studies, in which the same uating the role of peer use as a correlate and predictor of individuals are followed over time, provide more leverage but alcohol use is the extent to which friends actually influence still leave room for alternative interpretations. For example, an individual or the individual select friends who, like them, these studies may suffer from selection effects—that is, a already drink. During adolescence and the transition to construct excluded from the analysis actually “causes” both adulthood, both of these processes typically play a role (e.g., drug use and assumed consequence of drug use, rendering Patrick et al. 2012b). the relationship between cause and consequence spurious. Some recent analytic strategies that have been used with Influence of School and Work longitudinal data, such as propensity score analyses (Bachman et al. 2011) and fixed effects analysis (Patrick et al. 2012a; The broad domain of education also significantly relates to Staff et al. 2009), allow for greater control of selection AOD use during adolescence (Crosnoe 2011). Studies con- effects and thus better leverage on likely causal connections. sistently have found that grades, educational expectations, Nevertheless, despite such statistical advances, experiments and school bonding are negatively correlated with AOD use, in which participants are randomly assigned to experimental whereas school disengagement, school failure, school misbe- groups remain the gold standard for demonstrating causal havior, and skipping school are positively correlated with connections. AOD use (Bachman et al. 2008; Bryant et al. 2003; Dever Finally, the use of self-report data may limit the useful- et al. 2012; Patrick and Schulenberg 2010; Pilgrim et al. ness of study findings because such data rely on participants 2006; Schulenberg et al. 1994). For example, in a longitudinal to remember and accurately perceive their own level of sub- analysis examining 8th-grade predictors of concurrent and stance use. Nevertheless, most studies like the MTF study subsequent AOD use, school misbehavior and peer encour- rely on these measures, because they have been found to be agement of misbehavior were positively associated with valid and reliable (Bachman et al. 2011; O’Malley et al. 1983) concurrent substance use and increased substance use over and because it is very expensive and burdensome to collect time. Conversely, school bonding, interest, and effort were physiological data (e.g., blood, urine, or hair) and/or infor- negatively associated with concurrent and increased sub- mation from multiple reporters (e.g., parents or peers) in stance use, as were academic achievement and parental help large-scale studies. with school (Bryant et al. 2003). Positive school attitudes were of particular importance and were especially influential as protective factors against substance use for low-achieving Influence of Parents and Peers adolescents. The relationship between educational factors One developmental transition characteristic of adolescence and AOD use is bidirectional, and it is clear that AOD use is the movement away from parents and increasing involve- can contribute to educational difficulties. In general, however, ment with peers. Nonetheless, parents still play a pivotal role it seems that based on MTF study longitudinal data and in adolescent experiences and in fact can sometimes counter careful consideration of selection factors, the more common the effects of other risk factors for AOD use. Like many direction of influence is that school difficulties contribute other reports in the literature (e.g., Dishion and McMahon to AOD use during adolescence (Bachman et al. 2008). 1998; Kiesner et al. 2009), the MTF study found that parental By the time they leave high school, most adolescents supervision and monitoring relate to lower AOD use among have worked part time during the school year. Although it 8th and 10th graders and together are one of the strongest has long been recognized that hours of work during adoles-

196 Alcohol Research: Current Reviews cence are positively related to use of AODs, conclusions Risk taking and Sensation Seeking about causal connections have remained elusive (Staff et The willingness to take risks and high levels of sensation al. 2009). Analysis of MTF study data found that when seeking also both correlate with higher levels of AOD use sociodemographic and educational characteristics are con- (Dever et al. 2012; Patrick and Schulenberg 2010; Pilgrim et trolled for, the positive relationship between hours of work al. 2006; Schulenberg et al. 1996). Among 8th graders and and AOD use diminishes, suggesting that selection effects 10th graders, the impact of risk taking on substance use exist. In other words, long hours of work and substance use (including alcohol) was partly mediated through school have a common set of causes, particularly disengagement bonding (which negatively affected AOD use) and time with from school (Bachman and Schulenberg 1993; Bachman friends (which positively affected AOD use); these effects et al. 2011). The influence of selection effects is further sup- were largely invariant across race/ethnicity and gender ported by findings that simply wanting to work long hours (Pilgrim et al. 2006). is associated with heavier AOD use. This is true regardless of actual hours spent working, and especially among those who Well-Being do not work (Bachman et al. 2003; Staff et al. 2010). Self-esteem tends to be negatively correlated with AOD use and, correspondingly, self-derogation and depressive affect Religiosity and Community Attachment tend to be positively correlated with AOD use during ado- Numerous studies found that religiosity tends to be nega- lescence (Maslowsky and Schulenberg, in press; Patrick and tively correlated with AOD use during adolescence (Brown Schulenberg 2010; Schulenberg et al. 1996). When examin- et al. 2001; Wallace et al. 2003, 2007; Wray-Lake et al. ing the relative contributions of conduct problems, depres- 2012). This is true for both African American and White sive affect, and the interaction of conduct problems and youth. In fact, religiosity does not explain race differences depressive affect on AOD use, depressive affect is not as in substance use (Wallace et al. 2003). Religiosity tends powerful a predictor as are conduct problems. However, to operate at both the individual and contextual levels, the interaction of the two variables (i.e., high levels of both) because highly religious adolescents attending highly reli- is a relatively powerful predictor of alcohol use, especially for gious schools have lower alcohol use compared with highly younger adolescents (Maslowsky and Schulenberg, in press). religious adolescents attending non–highly religious schools (Wallace et al. 2007). More broadly, community attach- drinking Attitudes and Reasons for using Alcohol ments, including religiosity as well as social trust and social Attitudes regarding alcohol use and reasons for use are pow- responsibility, tend to be negatively correlated with AOD erful correlates and predictors of drinking behavior. Indeed, use during adolescence (Wray-Lake et al. 2012). disapproval of binge drinking is one of the strongest protec- tive factors against heavy drinking (Patrick and Schulenberg Exercise and Sports Participation 2010). A long-standing focus of the MTF study has been to show how, at the population level, changes in perceptions of Whereas exercise correlates negatively with alcohol use, risk about and disapproval of substance use precede changes participating in team sports correlates positively with alcohol in substance use (Bachman et al. 1998; Johnston et al. 2012; use during high school (Terry-McElrath et al. 2011). This is Keyes et al. 2011). A recent analysis assessed the effects of especially true for males (Dever et al. 2012). age, period (i.e., the year in which data were obtained), and cohort effects of population-based social norms regarding Externalizing Behaviors and other drug use heavy alcohol use (i.e., level of disapproval of heavy use) on As part of a broader set of problem behaviors, it is not individual-level heavy drinking during adolescence. The study found that cohort effects predominated, indicating surprising that alcohol use is associated with externalizing that being part of a birth cohort that reported higher disap- behaviors as well as cigarette smoking and illicit drug use proval of heavy drinking set the stage for lower alcohol use during adolescence. In the MTF study, externalizing behaviors (Keyes et al. 2012). overall, and aggressive behavior and theft/property damage Motivations or reasons for drinking also are associated in particular, correlated with AOD use during adolescence with alcohol use behaviors and may serve as a marker for the (Bachman et al. 2008; Brown et al. 2001; Maslowsky and development of problematic behavioral patterns. The reasons Schulenberg, in press; Patrick and Schulenberg 2010). for alcohol use typically change across adolescence and into Disentangling causal connections is difficult, however, and it adulthood. MTF study investigators have assessed reasons is likely that alcohol use both contributes to and is caused by for drinking using MTF study panel data following high- externalizing behaviors (Osgood et al. 1988), particularly if school seniors into young adulthood. (MTF survey questions these behaviors involve spending unsupervised time with regarding motivations are not included in the 8th- and 10th- peers (Osgood et al. 1996). Cigarette smoking and other grade surveys.) Of particular interest here, 12th-grade illicit drug use also tend to be highly correlated with alcohol adolescents tend to report higher motivation for drinking use during adolescence (Patrick and Schulenberg 2010). to get drunk (as well as other social and coping reasons for

Prevalence and Predictors of Adolescent Alcohol Use and Binge Drinking in the United States 197 drinking) than do young adults. Conversely, 12th graders drinking multiple times within a given 2-week period (Patrick report lower motivations to use alcohol to relax, to sleep, and and Schulenberg 2010) highlights the importance of prevent- because it tastes good, all of which increase across the transi- ing early initiation as well as early escalation of AOD use. tion to adulthood (Patrick and Schulenberg 2011; Patrick Levels of alcohol use have been declining in recent et al. 2011). It is important to understand the reasons for decades, suggesting that past interventions, such as increas- alcohol use among adolescents, because the reasons for use ing the minimum legal drinking age to 21, have been effec- reported in 12th grade, when adolescents are about 18, show tive. However, although it is worth recognizing that most long-term longitudinal associations with alcohol use and adolescents manage to avoid heavy alcohol use and that such symptoms of alcohol use disorders decades later (Patrick et use is not an inevitable developmental progression, alcohol al. 2011; Schulenberg et al. 1996). remains the most commonly used substance among adoles- cents, and its use is a leading cause of death and injury (U.S. Department of Health and Human Services 2007). To Long-term Consequences of Alcohol Use design effective programs and target prevention efforts toward students most likely to develop problematic levels of Attempting to discern long-term consequences of adolescent alcohol use, it is essential to identify characteristics of indi- AOD use is fraught with conceptual and methodological viduals at greatest risk. This effort is aided by the fact that complexities (e.g., Schulenberg et al. 2003), yet it is critical the importance of risk and protective factors tends to remain for understanding the development (i.e., etiology) of adult very stable over time. As summarized above, demographic alcohol use disorders. Numerous studies have demonstrated differences in drinking behavior point to important sub- that alcohol use in middle school and high school may groups that should be targeted, including young men and be an important indicator of later problems. For example, White and Hispanic adolescents. Finally, the findings although most students mature out of their heavy alcohol described here point to several risk and protective factors use (Schulenberg and Maggs 2002; Schulenberg and Patrick to consider when designing prevention and intervention 2012; Schulenberg et al. 1996), substance use in high school programs, including parental involvement, peer influences, is one of the strongest predictors of substance use in adult- academic success, religiosity, externalizing and internalizing hood. Specifically, binge drinking in 12th grade predicts behaviors, alcohol attitudes, and self-reported reasons for symptoms of alcohol use disorders 17 years later, at age 35 drinking. ■ (Merline et al. 2004, 2008; Patrick et al. 2011). Further- more, trajectories of binge drinking are predictive of alcohol use disorders during middle adulthood (Schulenberg and Acknowledgements Patrick 2012), and continued substance use into young adulthood is associated with HIV-related risk behaviors Data collection and manuscript preparation were funded by (Patrick et al. 2012). Finally, binge drinking in high school R01–DA001411 and R01–DA016575. The content here is predicts subsequent dropping out of college, although an solely the responsibility of the authors and does not necessar- increase in binge drinking during college is related to not ily represent the official views of the sponsors. dropping out (Schulenberg and Patrick 2012).

Financial Disclosure implications for Prevention and intervention The authors declare that they have no competing financial Studies on the etiology and epidemiology of alcohol use interests. ought to go hand in hand in order to combine the broader approach of epidemiology with the more in-depth emphasis of etiology. As the discussion in this article has shown, there References are both historical and developmental predictors related to adolescent AOD use that are changing over time. Under - baChman, J.g., anD sChulEnbERg, J. how part-time work intensity relates to drug use, prob- standing the scope of alcohol use during the middle-school lem behavior, time use, and satisfaction among high school seniors: are these conse- and high-school years, and associated long-term problems, is quences, or merely correlates? Developmental Psychology 29:220–235, 1993. an important step toward effectively intervening to reduce baChman, J.g.; Johnston, l.D.; anD o’mallEy, P.m. Explaining the recent increases in stu- dents’ marijuana use: impacts of perceived risks and disapproval, 1976 through 1996. high-risk drinking and its negative consequences. The scope American Journal of Public Health 88(6):887–892, 1998. PmiD: 9618614 of the problem is underscored by the findings that more than one in five American high-school seniors in the class of baChman, J.g.; Johnston, l.D.; o’mallEy, P.m.; anD sChulEnbERg, J.E. The Monitoring the Future Project After Thirty-Seven Years: Design and Procedures. (monitoring the Future 2011 reported binge drinking in the previous 2 weeks. The occasional Paper no. 76.) ann arbor, mi: institute for social Research, 2011. documented developmental increases in alcohol use across adolescence and young adulthood make this a particularly baChman, J.g.; o’mallEy, P.m.; sChulEnbERg, J.E.; Et al. The Education–Drug Use Connection: How Successes and Failures in School Relate to Adolescent Smoking, important time for intervention. In particular, the fast escala- Drinking, Drug Use, and Delinquency. mahwah, nJ: lawrence Erlbaum associates, tion among adolescents from binge drinking once to binge 2008.

198 Alcohol Research: Current Reviews baChman, J.g.; saFRon, D.J.; sy, s.R.; anD sChulEnbERg, J.E. Wishing to work: new per- mERlinE, a.C.; o’mallEy, P.m.; sChulEnbERg, J.E.; Et al. substance use among adults 35 spectives on how adolescents’ part-time work intensity is linked to educational disen- years of age: Prevalence, adulthood predictors, and impact of adolescent substance gagement, substance use, and other problem behaviours. International Journal of use. American Journal of Public Health 94(1):96–102, 2004. PmiD: 14713705 Behavioral Development 27(4):301–315, 2003. mERlinE, a.; JagER, J.; anD sChulEnbERg, J.E. adolescent risk factors for adult alcohol use baChman, J.g.; staFF, J.; o’mallEy, P.m.; Et al. twelfth-grade student work intensity linked and abuse: stability and change of predictive value across early and middle adulthood. to later educational attainment and substance use: new longitudinal evidence. Addiction 103(suppl. 1):84–99, 2008. PmiD: 18426542 Developmental Psychology 47(2):344–363, 2011. PmiD: 21171751 o’mallEy, P.m.; baChman, J.g.; anD Johnston, l.D. Reliability and consistency in self- bRoWn, s.a.; mCguE, m.; maggs, J.l.; Et al. underage alcohol use: summary of develop- reports of drug use. International Journal of the Addictions 18(6):805–824, 1983. mental processes and mechanisms, ages 16–20. Alcohol Research & Health PmiD: 6605313 32(1):41–52, 2009. PmiD: 23104446 osgooD, D.W.; Johnston, l.D.; o’mallEy, P.m.; anD baChman, J.g. the generality of bRoWn, t.n.; sChulEnbERg, J.; baChman, J.g.; Et al. are risk and protective factors for sub- deviance in late adolescence and early adulthood. American Sociological Review stance use consistent across historical time? national data from the high school classes 53:81–93, 1988. of 1976 through 1997. Prevention Science 2(1):29–43, 2001. PmiD: 11519373 osgooD, D.W.; Wilson, J.K.; o’mallEy, P.m.; Et al. Routine activities and individual bRyant, a.l.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. how academic achievement, attitudes, deviant behaviors. American Sociological Review 61:635–655, 1996.

and behaviors relate to the course of substance use during adolescence: a 6-year, multi- PatRiCK, m.E., anD sChulEnbERg, J.E. alcohol use and heavy episodic drinking prevalence wave national longitudinal study. Journal of Research on Adolescence 13(3):361–397, and predictors among national samples of american eighth- and tenth-grade students. 2003. Journal of Studies on Alcohol and Drugs 71(1):41–45, 2010. PmiD: 20105412

CaiRns, R.b., anD CaiRns, b.D. Lifelines and Risks: Pathways of Youth in Our Time. new PatRiCK, m.E., anD sChulEnbERg, J.E. how trajectories of reasons for alcohol use relate to york: Cambridge university Press, 1994. trajectories of binge drinking: national panel data spanning late adolescence to early adulthood. Developmental Psychology 47(2):311–317, 2011. PmiD: 21219061 Chassin, l.; hussong, a.; anD bEltRan, i. adolescent substance use. in: lerner, R.m., and steinberg, l., Eds. Handbook of Adolescent Psychology, 3rd Edition. hoboken, nJ: PatRiCK, m.E.; o’mallEy, P.m.; Johnston, l.D.; Et al. hiV/aiDs risk behaviors and substance Wiley, 2009. use by young adults in the united states. Prevention Science 13(5):532–538, 2012a. PmiD: 22886042 CRosnoE, R. Fitting In, Standing Out: Navigating the Social Challenges of High School to Get an Education. new york: Cambridge university Press, 2011. PatRiCK, m.E.; sChulEnbERg, J.E.; maggs, J.l.; anD masloWsKy, J. substance use and peers during adolescence and the transition to adulthood: selection, socialization, and devel- DEVER, b.V.; sChulEnbERg, J.E.; DWoRKin, J.b.; Et al. Predicting risk-taking with and without opment. in: sher, K., Ed. Handbook of Substance Use Disorders. new york: oxford substance use: the effects of parental monitoring, school bonding, and sports partici- university Press, 2012b. pation. Prevention Science 13(6):605–615, 2012. PmiD: 22960940 PatRiCK, m.E.; sChulEnbERg, J.E.; maRtz, m.E.; Et al. Extreme binge drinking among 12th- Dishion, t.J., anD mCmahon, R.J. Parental monitoring and the prevention of child and grade students in the united states: Prevalence and predictors. JAMA Pediatrics, adolescent problem behavior: a conceptual and empirical formulation. Clinical Child 167(11):1019–1025, 2013 PmiD: 24042318. and Family Psychology Review 1(1):61–75, 1998. PmiD: 11324078 PatRiCK, m.E.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. adolescents’ reported reasons for DoDgE, K.a.; malonE, P.s.; lansFoRD, J.E.; Et al. a dynamic cascade model of the develop- alcohol and marijuana use as predictors of substance use and problems in adulthood. ment of substance use onset. Monographs of the Society for Research in Child Journal of Studies on Alcohol and Drugs 72(1):106–116, 2011a. PmiD: 21138717 Development 74(3):1–119, 2009. PmiD: 19930521 PatRiCK, m.E.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. age-related changes in reasons for DonoVan, J.E.; JEssoR, R.; anD Costa, F.m. adolescent problem drinking: stability of psy- using alcohol and marijuana from ages 18 to 30 in a national sample. Psychology of chosocial and behavioral correlates across a generation. Journal of Studies on Alcohol Addictive Behaviors 25(2):330–339, 2011b. PmiD: 21417516 60(3):352–361, 1999. PmiD: 10371263 PilgRim, C.C.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. mediators and moderators of JagER, J.; sChulEnbERg, J.E.; o‘mallEy, P.m.; anD baChman, J.g. historical variation in drug parental involvement on substance use: a national study of adolescents. Prevention use trajectories across the transition to adulthood: the trend towards lower intercepts Science 7(1):75–89, 2006. PmiD: 16572302 and steeper, ascending slopes. Development and Psychopathology 25(2):527–543, 2013. PmiD: 23627961 sChulEnbERg, J.E. understanding the multiple contexts of adolescent risky behavior and positive development: advances and future directions. Applied Developmental Science Johnston, l.D.; o’mallEy, P.m.; baChman, J.g.; anD sChulEnbERg, J.E. Monitoring the 10(2):107–113, 2006. Future National Survey Results on Drug Use, 1975–2011. Volume I: Secondary School Students. ann arbor, mi: institute for social Research, university of michigan, 2012. sChulEnbERg, J.E., anD maggs, J.l. a developmental perspective on alcohol use and heavy drinking during adolescence and the transition to young adulthood. Journal of KEyEs, K.m.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. the social norms of birth cohorts Studies on Alcohol (suppl. 14):54–70, 2002. PmiD: 12022730 and adolescent marijuana use in the united states, 1976–2007. Addiction 106(10): 1790–1800, 2011. PmiD: 21545669 sChulEnbERg, J.E., anD masloWsKy, J. taking substance use and development seriously: Developmentally distal and proximal influences on adolescent drug use. Monographs of KEyEs, K.m.; sChulEnbERg, J.E.; o’mallEy, P.m.; Et al. birth cohort effects on adolescent the Society for Research in Child Development 74(3):121–130, 2009. PmiD: 19930536 alcohol use: the influence of social norms from 1976 to 2007. Archives of General Psychiatry 69(12):1304–1313, 2012. PmiD: 22868751 sChulEnbERg, J.E., anD PatRiCK, m.E. historical and developmental patterns of alcohol and drug use among college students: Framing the problem. in: White, h.R., and Rabiner, KiEsnER, J.; Dishion, t.J.; Poulin, F.; anD PastoRE, m. temporal dynamics linking aspects D., Eds. College Drinking and Drug Use. new york: guilford, 2012, pp. 13–35. of parent monitoring with early adolescent antisocial behavior. Social Development 18(4):765–784, 19966920 sChulEnbERg, J.; baChman, J.g.; o’mallEy, P.m.; anD Johnston, l.D. high school education- al success and subsequent substance use: a panel analysis following adolescents into maggs, J.l., anD sChulEnbERg, J.E. trajectories of alcohol use during the transition to young adulthood. Journal of Health and Social Behavior 35(1):45–62, 1994. PmiD: adulthood. Alcohol Research and Health 28:195–211, 2005. 8014429 masloWsKy, J., anD sChulEnbERg, J.E. interaction matters: Quantifying conduct problems by sChulEnbERg, J.E.; maggs, J.m.; anD o’mallEy, P.m. how and why the understanding of depressive symptoms interaction and its association with adolescent alcohol, cigarette, developmental continuity and discontinuity is important: the sample case of long-term and marijuana use in a national sample. Development and Psychopathology, consequences of adolescent substance use. in: mortimer, J.t., and shanahan, m.J., 25:1029–1043, 2013 PmiD: 24229547 Eds. Handbook of the Life Course. new york: Plenum Publishers, 2003, pp. 413–436.

Prevalence and Predictors of Adolescent Alcohol Use and Binge Drinking in the United States 199 sChulEnbERg, J.; o’mallEy, P.m.; baChman, J.g.; Et al. getting drunk and growing up: WagEnaaR, a.C.; o’mallEy, P.m.; anD laFonD, C. lowered legal blood alcohol limits for trajectories of frequent binge drinking during the transition to young adulthood. young drivers: Effects on drinking, driving, and driving-after-drinking behaviors in 30 Journal of Studies on Alcohol 57(3):289–304, 1996. PmiD: 8709588 states. American Journal of Public Health 91(5):801–804, 2001. PmiD: 11344892 WallaCE, J.m., JR.; baChman J.g.; o’mallEy, P.m.; Et al. gender and ethnic differences in staFF, J.; mEssERsmith, E.E.; anD sChulEnbERg, J.E. adolescents and the world of work. in: smoking, drinking, and illicit drug use among american 8th, 10th and 12th grade stu- lerner, R., and steinberg, b., Eds. Handbook of Adolescent Psychology, 3rd Edition. dents, 1976–2000. Addiction 98(2):225–234, 2003a. PmiD: 12534428 new york: John Wiley and sons, 2009, pp. 270–313. WallaCE, J.m., JR.; bRoWn, t.n.; baChman J.g.; anD laVEist, t.a. the influence of race and staFF, J.; sChulEnbERg, J.E.; masloWsKy, J.; Et al. substance use changes and social role religion on abstinence from alcohol, cigarettes and marijuana among adolescents. transitions: Proximal developmental effects on ongoing trajectories from late adoles- Journal of Studies on Alcohol 64(6):843–848, 2003b. PmiD: 14743948

cence through early adulthood. Development and Psychopathology 22(4):917–932, WallaCE, J.m., JR.; yamaguChi, R.; baChman, J.g.; Et al. Religiosity and adolescent sub- 2010. PmiD: 20883590 stance use: the role of individual and contextual influences. Social Problems 54(2):308–327, 2007. tERRy-mCElRath, y.m.; o’mallEy, P.m.; anD Johnston, l.D. Exercise and substance use among american youth, 1991–2009. American Journal of Preventive Medicine WilsnaCK, R.W.; VogEltanz, n.D.; WilsnaCK, s.C.; anD haRRis, t.R. gender differences in alcohol consumption and adverse drinking consequences: Cross-cultural patterns. 40(5):530–540, 2011. PmiD: 21496752 Addiction 95(2):251–265, 2000. u.s. Department of health and human services (Dhhs). The Surgeon General’s Call to WRay-laKE, l.; maggs, J.l.; Johnston, l.D.; Et al. associations between community attach- Action to Prevent and Reduce Underage Drinking. Washington, DC: office of the ments and adolescent substance use in nationally representative samples. Journal of surgeon general, Dhhs, 2007. Adolescent Health 51(4):325–331, 2012. PmiD: 22999832

new Alcohol Alert

Number 86 Epigenetics—A New Frontier for Alcohol Research

In today’s modern age, with the entire human genetic makeup (the genome) decoded and genetic testing more and more common, most people understand that changes, or mutations, in genes can lead to death, disease, and disorder. But there is another way our genes can impact our health. Increasingly, researchers and physicians are exploring the concept of “epigenetics” – a complex process that determines which genes in the body are active at any given time. This rapidly emerging area of research suggests that age, the environment, and exposure to drugs and other chemicals, including alcohol, directly affect epigenetics, altering normal genetic patterns and leading to abnormal expression or silencing of essential genes. The research also suggests that these epigenetic changes can, in some instances, be passed from one generation to the next. This Alert describes how alcohol influences epigenetics and how those influences may be associated with illness and disorders, including fetal alcohol spectrum disorders (FASD), cancer, liver disease and other gastrointestinal disorders, brain development, the body’s internal clock, and its ability to fight disease. This issue also explores therapies that may target those changes caused by epigenetics.

http://pubs.niaaa.nih.gov/publications/

200 Alcohol Research: Current Reviews We have learned that predisposing factors include an indi- the Burden of Alcohol Use vidual’s genetic susceptibility to the positive and negative effects of alcohol, alcohol use during high school, campus norms related to drinking, expectations regarding the benefits and detrimental effects of drinking, penalties for underage Excessive Alcohol Consumption drinking, parental attitudes about drinking while at college, whether one is member of a Greek organization or involved in athletics, and conditions within the larger community and Related Consequences Among that determine how accessible and affordable alcohol is. Consequences include missed classes and lower grades, injuries, College Students sexual assaults, overdoses, memory blackouts, changes in brain function, lingering cognitive deficits, and death. This article reviews recent research findings about alcohol consumption by today’s college students and the outcomes that follow. It examines what we know about the causes and consequences of excessive drinking among college students Aaron White, Ph.D., and Ralph Hingson, Sc.D. relative to their non-college peers and many of the strategies used to collect and analyze relevant data, as well as the inherent hurdles and limitations of such strategies. Research shows that multiple factors influence college drinking, from an individual’s genetic susceptibility to the positive and negative effects of alcohol, alcohol use during high school, Excessive Drinking At College campus norms related to drinking, expectations regarding the benefits and detrimental effects of drinking, penalties for Currently, only two active national survey studies are able to underage drinking, parental attitudes about drinking while at characterize the drinking habits of college students in the college, whether one is member of a greek organization or United States. The National Survey on Drug Use and Health involved in athletics, and conditions within the larger community (NSDUH), an annual survey sponsored by the Substance that determine how accessible and affordable alcohol is. Abuse and Mental Health Services Administration (SAMHSA), Consequences of college drinking include missed classes and involves face-to-face interviews with approximately 67,500 lower grades, injuries, sexual assaults, overdoses, memory persons ages 12 and older each year regarding use of alcohol blackouts, changes in brain function, lingering cognitive deficits, and other drugs. Monitoring the Future (MTF) is an annual, and death. this article examines recent findings about the paper-and-pencil national survey of alcohol and other drug causes and consequences of excessive drinking among college use with a sample comprising nearly 50,000 students in 8th, students relative to their non-college peers and many of the l0th, and 12th grades drawn from roughly 420 public and strategies used to collect and analyze relevant data, as well as private schools. Approximately 2,400 graduating seniors are the inherent hurdles and limitations of such strategies. kEY resurveyed in subsequent years, allowing for the monitoring WoRDS: Alcohol consumption; alcohol use, abuse, and of trends in college drinking. dependence; alcohol burden; alcohol effects and In addition, two prior surveys yielded data on college consequences; harmful drinking; underage drinking; binge drinking that remain valuable and relevant. The National drinking; college student; risk factors; genetic factors; Epidemiologic Survey on Alcohol and Related Conditions environmental factors; social norms; parental attitude; Greek (NESARC), sponsored by NIAAA, collected data on alcohol organization; athletes; community environment; academic and other drug use from a sample of roughly 46,500 citizens performance; injury; sexual assault; overdose; memory 18 and older using face-to-face computer-assisted interviews. blackout; brain function; cognitive deficits; death; Two waves of data (2001 and 2004) were collected from the accessibility; availability; affordability; survey; data same sample, and data from an independent sample are collection; data analysis. scheduled to be collected in 2013. The Harvard College Alcohol Study (CAS), although no longer active, was a land-

Aaron White, Ph.D., is program director, College ince 1976, when the National Institute on Alcohol Abuse and Alcoholism (NIAAA) issued its first report and Underage Drinking Prevention Research; and Son alcohol misuse by college students, research Ralph Hingson, Sc.D., is director, Division of Epidemiology advances have transformed our understanding of excessive and Prevention Research, both at the National Institute on drinking on college campuses and the negative outcomes Alcohol Abuse and Alcoholism, Bethesda, Maryland. that follow from it. For instance, we now know that a broad array of factors influence whether a particular college student will choose to drink, the types of consequences they suffer from drinking, and how they respond to those consequences.

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 201 mark paper-and-pencil survey that provided national data but specifies a 30-day time period (Chen et al. 2011). NIAAA (years 1993, 1997, 1999, and 2001) from roughly 15,000 uses the four or more/five or more gender-specific measure but students on more than 100 college campuses each year specifies a time frame of 2 hours for consumption, as this would (Wechsler and Nelson 2008). Data from both NESARC generate blood alcohol levels of roughly 0.08 percent, the legal and Harvard CAS remain useful for examining associations limit for driving, for drinkers of average weight (NIAAA 2004). between patterns of drinking at college and the frequency According to NSDUH, the percentage of 18- to 22-year- and prevalence of alcohol-related consequences for both old college students who reported drinking five or more drinks drinkers and nondrinkers. on an occasion in the previous 30 days remained relatively Data from NSDUH and MTF suggest that roughly 65 stable from 2002 (44 percent) to 2010 (44 percent) (SAMHSA percent of college students drink alcohol in a given month 2011). Among 18- to 22-year-olds not enrolled in college, the (see figure 1 for data from MTF), and Harvard CAS all sug- percentage who engaged in binge drinking decreased significantly gest that a large percentage of college students who drink do from 2002 (39 percent) to 2010 (36 percent) (see figure 2). so to excess. Excessive, or “binge,” drinking is defined in Looking at a longer time period, data from MTF suggest NSDUH, MTF, and NESARC as consuming five or more that there have been significant declines in the percentage of drinks in an evening, although the instruments vary in the college students consuming five or more drinks in the previ- specified time frames given (i.e., once or more in the past ous 2 weeks, from 44 percent in 1980 to 36 percent in 2011 month for NSDUH, past 2 weeks for MTF, and multiple (Johnston et al. 2012) (see figure 3). This time frame includes time periods for NESARC) (Johnston et al. 2001a; SAMHSA the passage of the National Minimum Drinking Age Act of 2011). The Harvard CAS was the first national study of col- 1984, which effectively increased the drinking age from 18 lege students to utilize a gender-specific definition of binge to 21 in the United States. drinking (i.e., four or more drinks in an evening for females Across the four waves of data collection in the Harvard or five or more for males in the past 2 weeks) to equate the CAS (1993, 1997, 1999, and 2001), rates of binge drinking risk of alcohol-related harms (Wechsler et al. 1995). The remained relatively stable (44, 43, 45, and 44 percent, respec- Centers for Disease Control and Prevention (CDC) utilizes tively) (Wechsler et al. 2002) (see figure 4). However, the the same four or more/five or more gender-specific measures number of non–binge drinkers decreased, whereas the number Percent

Years of Administration

Figure 1 alcohol: trends in 30-day prevalence among college students vs. others 1 to 4 years beyond high school (twelfth graders included for comparision).

souRCE: the monitoring the Future study, the university of michigan. notE: others refers to high school graduates 1 to 4 years beyond high school not currently enrolled full time in college.

202 Alcohol Research: Current Reviews of frequent binge drinkers (three or more binge-drinking of male and 32 percent of female college students crossed the episodes in a 2-week period) increased. Wechsler and colleagues binge threshold in a given 2-week period. Further, 40 percent (2002) reported that binge drinkers consumed 91 percent of students—more males (44 percent) than females (37 of all the alcohol consumed by college students during the percent)—reported getting drunk in a given month. Research study period. Frequent binge drinkers, a group comprising suggests that gender differences in alcohol use by college only 1 in 5 college students, accounted for 68 percent of all students have narrowed considerably over the years. In their alcohol consumed (Wechsler and Nelson 2008). landmark 1953 report on college drinking, Yale researchers Straus and Bacon indicated that, based on survey data from more than 15,000 students on 27 college campuses, 80 individual and Environmental Contributors to percent of males and 49 percent of females reported having Excessive Drinking been drunk at some point. Nearly 60 years later, in 2011, data from MTF indicated that 68 percent of males and 68 Survey data indicate that males outpace females with regard percent females reported having been drunk. These new, to binge drinking. According to MTF, in 2011, 43 percent higher levels of drinking among females seem to be ingrained Percent Using in Past Month

Figure 2 binge alcohol use among adults aged 18 to 22, by college enrollment: 2002–2011. survey years are shown on the horizontal axis, and the percentage using in the past month is shown on the vertical axis. For each college enrollment status (enrolled full time in college and not enrolled full time in college), there is a line showing use over the years 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, and 2011. tests of statistical significance at the .05 level were performed between 2011 and each of the previous years listed; significant results are indicated where appropriate. among adults aged 18 to 22 enrolled full time in college, 44.4 percent were past-month binge alcohol users in 2002, 43.5 percent in 2003, 43.4 percent in 2004, 44.8 percent in 2005, 45.6 percent in 2006, 43.6 percent in 2007, 40.7 percent in 2008, 43.6 percent in 2009, 42.2 percent in 2010, and 39.1 percent in 2011. the differences between the 2011 estimate and the 2002, 2003, 2004, 2005, 2006, 2007, 2009, and 2010 estimates were statistically significant. among adults aged 18 to 22 not enrolled full time in college, 38.9 percent were past-month binge alcohol users in 2002, 38.7 percent in 2003, 39.4 percent in 2004, 38.3 percent in 2005, 38.5 percent in 2006, 38.6 percent in 2007, 38.2 percent in 2008, 38.0 percent in 2009, 35.4 percent in 2010, and 35.4 percent in 2011. the differences between the 2011 estimate and the 2002 through 2009 esti- mates were statistically significant.

souRCE: substance abuse and mental health services administration. Results From the 2011 National Survey on Drug Use and Health: Summary of National Findings, nsDuh series h–44, hhs Publication no. (sma) 12–4713. Rockville, mD: substance abuse and mental health services administration, 2012.

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 203 in the youth drinking culture. Whereas binge-drinking rates excessive drinking at college and experiencing alcohol-related declined significantly among high-school seniors over the consequences (Varvil-Weld et al. 2013; White et al. 2002). last decade, the effect was driven by a decline among males only. Binge-drinking rates among females remained relatively stable (Johnston et al. 2012) (see figure 5). Strengths and Weaknesses of Binge-Drinking Beyond gender, survey studies of college drinking reveal Measures a range of characteristics of both individual students and campus environments that influence the likelihood of binge Several studies indicate that crossing commonly used drinking. Data from the Harvard CAS and other studies binge-drinking thresholds increases a college student’s risk reveal that males, Caucasians, members of Greek organizations, of experiencing negative alcohol-related consequences. For students on campuses with lower percentages of minority instance, data from the Harvard CAS indicate that students and older students, athletes, students coping with psychological who binge one or two times during a 2-week period are distress, those on campuses near a high density of alcohol outlets, students with access to cheap drink specials, a will- roughly three times as likely as non–binge drinkers to get ingness to endure the consequences of alcohol misuse, and behind in school work, do something regretful while drinking, drinking at off-campus parties and bars all contribute to experience a memory blackout, have unplanned sex, fail to excessive drinking (Mallett et al. 2013; Wechsler and Kuo use birth control during sex, damage property, get in trouble 2003; Yusko et al. 2008). Further, students living off campus with police, drive after drinking, or get injured (Wechsler et and/or in Greek housing, those who drink to try to fit it, al. 2000). The more often a student binges, the greater the students with inflated beliefs about the proportion of other risk of negative outcomes. Further, the more binge drinking students who binge drink, and those with positive expecta- that occurs on a campus, the more likely non–binge drinkers tions about the results of drinking are more likely to drink and abstainers are to experience secondhand consequences excessively (Scott-Sheldon et al. 2012; Wechsler and Nelson of alcohol use, such as having studying or sleep disrupted, 2008). Importantly, excessive drinking prior to college rela- being a victim of sexual assault, and having property damaged tive to other college-bound students is predictive of both (Wechsler and Nelson 2008). Percent

Years of Administration

Figure 3 alcohol: trends in 2-week prevalence of consuming five or more drinks in a row among college students vs. others 1 to 4 years beyond high school (12th graders included for comparision).

souRCE: the monitoring the Future study, the university of michigan. notE: others refers to high school graduates 1 to 4 years beyond high school not currently enrolled full time in college.

204 Alcohol Research: Current Reviews Because of the increased risk of consequences to self and in a 2-week period, compared with 7 percent of females and others that occurs when a person drinks at or beyond the 18 percent of males not in college. Further, 2 percent of all binge threshold, a great deal of emphasis is placed on tracking college females surveyed and 10 percent of college males the percentage of college students that cross binge thresholds. consumed 15 or more drinks in a 2-week period. Rates among Although this has proven extremely valuable, as Wechsler non-college peers were similar, at 2 percent among females and Nelson (2001, p. 289) state, “Alcohol use is a complex and 9 percent among males (Johnston et al. 2012). For a behavior. No single measure will capture all the relevant 140-pound female, consuming 15 drinks over a 6-hour period aspects of alcohol use.” One limitation of using a single would produce an estimated blood alcohol level above 0.4 threshold is that it removes data regarding just how heavily percent, a level known to have claimed, directly, several lives students actually drink (Alexander and Bowen 2004; Read on college campuses in recent years. For a 160-pound male, et al. 2008) and assigns the same level of risk to all students drinking in this way would lead to a blood alcohol level who cross the thresholds regardless of how far beyond the above 0.3 percent, a potentially lethal level associated with threshold they go. This is an important consideration as memory blackouts and injury deaths. recent studies suggest that plenty of college students who Data from the Harvard CAS suggested that students cross the binge threshold when they drink go far beyond it. who binge drink frequently (three or more times in a 2-week In a study of 10,424 first-semester college freshmen, period) are at particularly high risk of negative alcohol-related more than one-half of all males and one-third of all females outcomes. Compared with students who binge drink one categorized as binge drinkers drank at levels two or more or two times in a 2-week period, those who binge three or times the binge threshold (8 or more drinks for women and more times are twice as likely to experience alcohol-induced 10 or more drinks for men) at least once in the 2 weeks memory losses (27 percent vs. 54 percent, respectively), not before the survey. Indeed, one in four binge-drinking males use protection during sex (10 percent vs. 20 percent, respec- consumed 15 or more drinks at a time during that period tively), engage in unplanned sex (22 percent vs. 42 percent, (White et al. 2006). Naimi and colleagues (2010) reported respectively), and get hurt or injured (11 percent vs. 27 that 18- 24-year-olds in the United States drink an average percent, respectively), and are equally likely to need medical of 9.5 drinks per binge episode, nearly twice the standard treatment for an overdose (1 percent vs. 1 percent). Whereas binge threshold. Data from MTF also reveal that both col- binge frequency is associated with an increased risk of nega- lege students and their non-college peers often drink at levels tive outcomes, additional research indicates that there is a that exceed the binge threshold. On average, between 2005 relationship between how often a student binges and the and 2011, 7 percent of college females surveyed and 24 percent peak number of drinks he or she consumes. White and col- of college males consumed 10 or more drinks at least once leagues (2006) reported that 19 percent of frequent binge Percent of Students

Figure 4 Drinking habits of college students from harvard Cas.

souRCE: Johnston, l.D.; o'malley, P.m.; bachman, J.g.; and schulenberg, J.E. Monitoring the Future National Survey Results on Drug Use, 1975–2011: Volume I: Secondary School Students. ann arbor, mi: institute for social Research, the university of michigan.

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 205 drinkers consume three or more times the binge threshold although sorting students into binge drinking categories fails (12 or more drinks for females and 15 or more for males) at to capture high peak levels of consumption among students, least once in a 2-week period compared with only 5 percent a large proportion of harms actually occurs at or near the of infrequent binge drinkers. As a result of the association standard four or more/five or more threshold. between frequency of binge drinking and peak levels of con- sumption, it is difficult to determine if the increase in risk that comes with frequent bingeing is a result of the number Do Students know How to Define Standard of binge episodes, per se, or the number of drinks consumed Servings? in an episode. Importantly, although evidence suggests that many students Despite concerns about the accuracy of self-report data for drink at levels far beyond the binge threshold, additional assessing levels of alcohol use among college students and the research suggests that the majority of alcohol-related harms general population, such surveys remain the most common on college campuses result from drinking at levels near the tool for assessing alcohol use. One major concern is whether standard four/five-drink measure. This is related to the well- students and other young adults are aware of what consti- known prevention paradox in which the majority of health tutes a single serving of alcohol. Research shows that college problems, such as alcohol-related consequences, tend to occur students and the general public tend to define and pour single among those considered to be at lower risk (Rose 1985). servings of alcohol that are significantly larger than standard For a particular individual, the odds of experiencing alcohol- drinks, suggesting they might underestimate their true levels related harms increase as the level of consumption increases of consumption on surveys (Devos-Comby and Lange 2008; (Wechsler and Nelson 2001). However, at the population Kerr and Stockwell 2012). For instance, White and col- level, far fewer people drink in this manner. As a result, more leagues (2003, 2005) asked students to pour single servings total consequences occur among those who drink at relatively of different types of alcohol beverages into cups of various lower risk levels. For instance, based on data from roughly sizes. Overall, students poured drinks that were too large. 9,000 college-student drinkers across 14 college campuses in When asked to simply define standard drinks in terms of California, Gruenewald and colleagues (2010) estimated that fluid ounces, students tended to overstate the number of more than one-half of all alcohol-related consequences resulted ounces in a standard drink. The average number of ounces from drinking occasions in which four or fewer drinks were of liquor in student-defined mixed drinks was 4.5 ounces consumed. Similarly, using national data from nearly 50,000 rather than the 1.5 ounces in actual standard drinks (White students surveyed across the four waves of the Harvard CAS, et al. 2005). When students were provided with feedback Weitzman and Nelson (2004) observed that roughly one- regarding discrepancies between their definitions of single quarter to one-third of alcohol-related consequences, including servings and the actual sizes of standard drinks, they tended getting injured, vandalizing property, having unprotected sex, and falling behind in school, occurred among students to revise their self-reported levels of consumption upward, who usually consume three or four drinks per occasion. Such leading to a significant increase in the number of students findings raise the possibility that a reduction in high peak categorized as binge drinkers (White et al. 2005). Such findings levels of consumption might not necessarily result in large suggest that students underreport their levels of consumption overall reductions in alcohol-related consequences on a campus. on surveys, raising the possibility that more students drink However, a reduction in high peak levels of drinking would excessively than survey data indicate. certainly help save the lives of students who drink at these Although a lack of knowledge regarding standard serving high levels. sizes could lead students to underestimate, and thus under- In summary, while binge-drinking thresholds are useful report, how much they drink, field research suggests that the for sorting students into categories based on levels of risk, a discrepancy between self-reported and actual levels of con- single threshold cannot adequately characterize the drinking sumption might be smaller than expected from lab studies. habits of college students or the risks associated with alcohol For instance, Northcote and Livingston (2011) conducted use on college campuses (Read et al. 2008). It is not uncom- a study in which they monitored the number of drinks con- mon for college students to far exceed standard binge thresh- sumed by research participants in bars and then asked them olds. Presently, only MTF tracks and reports the incidence to report their consumption a few days later. Reports by of drinking beyond the binge threshold on college campuses. study participants were consistent with the observations Such data are important as they allow for better tracking of made by researchers for participants who had consumed changes in the drinking habits of students. For instance, it is less than eight total drinks. Only those who consumed eight possible that the number of students who drink at extreme drinks or more tended to underestimate their consumption. levels could increase, whereas the overall percentage of stu- When comparing estimated blood alcohol concentrations dents who binge drink declines or remains stable. Such a (BAC) based on self-report to actual BAC readings in college phenomenon might help explain why some consequences of students returning to campus from bars, actual BAC levels excessive alcohol use, like overdoses requiring hospitalization, tended to be lower, rather than higher, than levels calculated seem to be on the rise despite relatively stable levels of binge using self-reported consumption (Kraus et al. 2005). Similarly, drinking on college campuses across several decades. Finally, when actual BAC levels are compared with estimated BAC

206 Alcohol Research: Current Reviews levels in bar patrons, estimates are spread evenly between (2006) observed similar outcomes of self-reported alcohol accurate, underestimates, and overestimates (Clapp et al. 2009). consumption in a paper-and-pencil survey and an electronic In short, although self-reported drinking data might not survey. However, other studies suggest that students actually be perfect, and college students lack awareness of how stan- feel more comfortable answering personal questions truth- dard drink sizes are defined, research does not suggest that fully when completing questionnaires electronically (Turner the discrepancies between self-reported and actual drinking et al. 1998), which can lead to higher levels of self-reported levels are large enough to question the general findings of substance use and other risky behaviors. Both Lygidakis and college drinking surveys. colleagues (2010) and Wang and colleagues (2005) indicate that adolescents completing electronic surveys reported higher levels of alcohol and other drug use compared with Paper-and-Pencil, Face-to-Face, and Electronic those completing paper-and-pencil versions. Surveys: Does it Make a Difference? Response rate is an important consideration, with higher response rates increasing the representativeness of the National surveys of college drinking often utilize paper-and- sample and limiting the likelihood that response biases will pencil questionnaires (e.g., MTF and Harvard CAS) or face- influence the outcomes. Two national paper-and-pencil to-face computer-assisted personal interviews (e.g., NSDUH surveys mentioned above, MTF and Harvard CAS, report and NESARC). It now is possible to collect survey data response rates for college students of approximately 59 electronically via the Internet and also using handheld percent. For MTF, this response rate represents a retention devices, such as smartphones and personal digital assistants. rate, as the participants were followed up after high school. This raises questions about the comparability between tradi- Response rates for the in-person computer-assisted personal tional survey methods and electronic data collection. interviews, NSDUH and NESARC, which assess college Several studies comparing traditional (e.g., paper and student drinking but are not limited to college students, are pencil) and electronic means of data collection suggest that roughly 77 percent and 81 percent, respectively. Currently, the approaches yield generally similar results from survey there is no basis for assessing response rates for national participants (Boyer et al. 2002; Jones and Pitt 1999; LaBrie Web-based assessments of college drinking. However, smaller et al. 2006; Lygidakis et al. 2010). For instance, in a compar- studies suggest that response rates might be comparable, if ison of Web-based and paper-and-pencil survey approaches, not higher, than other approaches. McCabe and colleagues Knapp and Kirk (2003) found no differences in outcomes, (2002) reported that, among 7,000 undergraduate students, suggesting that Web-based surveys do not diminish the accu- one-half of whom were surveyed about alcohol and other racy or honesty of responses. Similarly, LaBrie and colleagues drug use via the Internet and half surveyed via paper-and- Percent

Males and females who binge once or more per 2 weeks

Figure 5 Percent of 12th-grade male and female students who reported drinking at least once in the prior 2 weeks.

souRCE: Wechsler, h.; lee, J.E.; Kuo, m., et al. trends in college binge drinking during a period of increased prevention efforts: Findings from 4 harvard school of Public health College alcohol study surveys: 1993–2001. Journal of American College Health 50(5):203–217, 2002. PmiD: 11990979

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 207 pencil surveys delivered through the mail, the response rates to 24-year-olds who were enrolled full time in 4-year colleges were 63 percent for the Web survey and 40 percent for the (33 percent in 2005, the most recent year analyzed) (Hingson paper-and-pencil survey. Further, response rates for Web-based et al. 2009). Because college students drink more heavily surveys can be improved by sending reminders via e-mail than their non-college peers, it is possible this approach (van Gelder et al. 2010). underestimated the magnitude of alcohol-related consequences In summary, in recent years an increasing number of on college campuses. Hingson and colleagues (2002, 2005, researchers have utilized electronic survey methods to collect 2009) also used the percentage of college students who college-drinking data. At present, evidence suggests that reported various alcohol-related behaviors (e.g., being assaulted these methods can yield results quite similar to those obtained by another drinking college student) in national surveys from traditional survey methods and that response rates to derive national estimates of the total numbers of college might actually be higher. students who experienced these consequences. Based on the above strategies along with other sources of data, researchers have estimated the following rates and Alcohol-Related Consequences Among College prevalence of alcohol-related harms involving college students: Students • Death: It is possible that more than 1,800 college students Drinking to intoxication leads to widespread impairments between the ages of 18 and 24 die each year from alcohol- in cognitive abilities, including decisionmaking and impulse related unintentional injuries, including motor-vehicle control, and impairments in motor skills, such as balance crashes (Hingson et al. 2009). and hand-eye coordination, thereby increasing the risk of injuries and various other harms. Indeed, research suggests • Injury: An estimated 599,000 students between the ages that students who report “getting drunk” even just once in of 18 and 24 are unintentionally injured each year under a typical week have a higher likelihood of being injured, the influence of alcohol (Hingson et al. 2009). experiencing falls that require medical treatment, causing injury in traffic crashes, being taken advantage of sexually, • Physical Assault: Approximately 646,000 students between and injuring others in various ways (O’Brien et al. 2006). the ages of 18 and 24 are assaulted each year by another Students who drink with the objective of getting drunk are student who has been drinking (Hingson et al. 2009). far more likely to experience a range of consequences, from hangovers to blackouts, than other students who drink • Sexual Assault: Perhaps greater than 97,000 students (Boekeloo et al. 2011). between the ages of 18 and 24 are victims of alcohol- National estimates suggest that thousands of college related sexual assault or date rape each year (Hingson et students are injured, killed, or suffer other significant conse- al. 2009). quences each year as a result of drinking. However, researchers have questioned the manner in which such national estimates • Unsafe Sex: An estimated 400,000 students between the are calculated. In many cases, the lack of college identifiers ages of 18 and 24 had unprotected sex and nearly 110,000 in datasets means that the actual amount of annual alcohol- students between the ages of 18 and 24 report having attributable harm that occurs among college students is been too intoxicated to know if they consented to having unknown. Although the Harvard CAS collected data regarding sex (Hingson et al. 2002). the consequences of drinking, its final year of administration was 2001. Currently, assessing the damage done, on a • Health Problems: More than 150,000 students develop national level, by college drinking requires estimating rates an alcohol-related health problem each year (Hingson et of consequences using a variety of data sources. Such assess- al. 2002). ments are complicated by the fact that outcomes considered to be negative consequences by researchers (e.g., blackouts • Suicide Attempts: Between 1.2 and 1.5 percent of college and hangovers) are not always perceived as negative by students students indicate that they tried to commit suicide within (Mallett et al. 2013). Further, college students often drink the past year as a result of drinking or drug use (Presley et off campus, such as during spring breaks and summer vacations, al. 1998). meaning that many alcohol-related consequences experienced by college students are not necessarily associated with college • Drunk Driving: Roughly 2.7 million college students itself. As such, our understanding of alcohol-related conse- between the ages of 18 and 24 drive under the influence quences among college students remains somewhat cloudy. of alcohol each year (Hingson et al. 2009). In one set of estimates, Hingson and colleagues (2002, 2005, 2009) utilized census data and national datasets • Memory Loss: National estimates suggest that 10 percent regarding traffic crashes and other injury deaths to estimate of non–binge drinkers, 27 percent of occasional binge the prevalence of various alcohol-related harms among all drinkers, and 54 percent of frequent binge drinkers young people aged 18–24. Next, they attributed an amount reported at least one incident in the past year of blacking of harm to college students equal to the proportion of all 18- out, defined as having forgotten where they were or what

208 Alcohol Research: Current Reviews they did while drinking (Wechsler et al. 2000; White Williams 2003; Wood et al. 1997), studies linking binge 2003). drinking to poorer academic performance outnumber the former studies two to one. Presley and Pimentel (2006) • Property Damage: More than 25 percent of administra- reported that in a national survey of college students, those tors from schools with relatively low drinking levels and who engaged in binge drinking and drank at least three more than 50 percent from schools with high drinking times per week were 5.9 times more likely than those who levels say their campuses have a “moderate” or “major” drank but never binged to perform poorly on a test or pro- problem with alcohol-related property damage (Wechsler ject as a result of drinking (40.2 vs. 6.8 percent), 5.4 times et al. 1995). more likely to have missed a class (64.4 vs. 11.9 percent), and 4.2 times more likely to have had memory loss (64.2 • Police Involvement: Approximately 5 percent of 4-year vs. 15.3 percent) (Thombs et al. 2009). Singleton and college students are involved with the police or campus colleagues (2007, 2009), in separate prospective studies, security as a result of their drinking (Wechsler et al. 2002) found negative associations between heavy alcohol use and and an estimated 110,000 students between the ages of grade point average. Jennison (2004), based on a national 18 and 24 are arrested for an alcohol-related violation prospective study, reported binge drinkers in college were such as public drunkenness or driving under the influence more likely to drop out of college, work in less prestigious (Hingson et al. 2002). A more recent national study jobs, and experience alcohol dependence 10 years later. reported that 8.5 percent of students were arrested or had Wechsler and colleagues (2000) and Powell and colleagues other trouble with the police because of drinking (Presley (2004), based on the Harvard CAS, found frequent binge and Pimentel 2006). drinkers were six times more likely than non–binge drinkers to miss class and five times more likely to fall behind in • Alcohol Abuse and Dependence: Roughly 20 percent of school. White and colleagues (2002) observed that the number college students meet the criteria for an alcohol use disorder of blackouts, a consequence of heavy drinking, was negatively in a given year (8 percent alcohol abuse, 13 percent alcohol associated with grade point average (GPA). It is important dependence). Rates among age mates not in college are to note that although data regarding GPA often are collected comparable (17 percent any alcohol use disorder, 7 percent via self-report, the negative association between alcohol con- alcohol abuse, 10 percent alcohol dependence) (Blanco et sumption and GPA holds even when official records are al. 2008). obtained (Singleton 2007). Collectively, the existing research suggests that heavy drinking is associated with poorer academic With regard to assessing the number of college students success in college. who die from alcohol each year, in addition to the lack of college identifiers in datasets, another barrier is the fact that levels of alcohol often are not measured in nontraffic fatalities. Alcohol Blackouts As such, attributable fractions, based on analyses of existing reports in which alcohol levels were measured postmortem, Excessive drinking can lead to a form of memory impairment are used to estimate the number of deaths by various means known as a blackout. Blackouts are periods of amnesia during that likely involved alcohol. The CDC often uses attributable which a person actively engages in behaviors (e.g., walking, fractions calculated by Smith and colleagues (1999) based talking) but the brain is unable to create memories for the upon a review of 331 medical-examiner studies. An updated events. Blackouts are different from passing out, which approach is needed. The combination of including college means either falling asleep or becoming unconscious from identifiers in medical records and measuring alcohol levels excessive drinking. During blackouts, people are capable of in all deaths would allow for accurate assessments of the participating in events ranging from the mundane, such as role of alcohol in the deaths of college students and their eating food, to the emotionally charged, such as fights and non-college peers. even sexual intercourse, with little or no recall (Goodwin 1995). Like milder alcohol–induced short-term memory impairments caused by one or two drinks, blackouts primarily Academic Performance are anterograde, meaning they involve problems with the formation and storage of new memories rather than problems About 25 percent of college students report academic conse- recalling memories formed prior to intoxication. Further, quences of their drinking, including missing class, falling short-term memory often is left partially intact. As such, behind in class, doing poorly on exams or papers, and receiving during a blackout, an intoxicated person is able to discuss lower grades overall (Engs et al. 1996; Presley et al. 1996a, b; events that happened prior to the onset of the blackout and Wechsler et al. 2002). Although some published research to hold new information in short-term storage long enough studies have not found a statistically significant association to have detailed conversations. They will not, however, be between binge drinking and academic performance (Gill able to transfer new information into long-term storage, 2002; Howland et al. 2010; Paschall and Freisthler 2003; leaving holes in their memory. Because of the nature of

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 209 blackouts, it can be difficult or impossible to know when a injury, whereas those with six or more blackouts were 2.5 drinker in the midst of one (Goodwin 1995). times more likely. In a follow-up report based on the same There are two general types of blackouts based on the sample, Mundt and Zakletskaia (2012) estimated that severity of the memory impairments. Fragmentary black- among study participants, one in eight emergency-department outs, sometimes referred to as gray outs or brown outs, are a (ED) visits for alcohol-related injuries involved a blackout. form of amnesia in which memory for events is spotty but On a campus of 40,000 students, this would translate into not completely absent. This form is the most common. roughly $500,000 in annual costs related to blackout-associated En bloc blackouts, on the other hand, represent complete ED visits. amnesia for events (Goodwin 1995). In the study of 50 students with blackout histories by Blackouts surprisingly are common among college students White and colleagues (2004), estimated peak BACs during who drink alcohol. White and colleagues (2002) reported the night of the last blackout generally were similar for males that one-half (51 percent) of roughly 800 college students (0.30 percent) and females (0.35 percent), although it is who had ever consumed alcohol at any point in their lives unlikely that self-reported alcohol consumption during reported experiencing at least one alcohol-induced blackout, nights in which blackouts occur is highly accurate. A study defined as awakening in the morning not able to recall of amnesia in people arrested for either public intoxication, things one did or places one went while under the influence. driving under the influence, or underage drinking found The average number of total blackouts in those who experi- that the probability of a fragmentary or en bloc blackout enced them was six. Of those who had consumed alcohol was 50/50 at a BAC of 0.22 percent and the probability of during the 2 weeks before the survey was administered, 9 an en bloc blackout, specifically, was 50/50 at a BAC of 0.31 percent reported blacking out. Based on data from 4,539 percent, based on breath alcohol readings (Perry et al. 2006). inbound college students during the summer between high- In their study of blackouts in college students, Hartzler and school graduation and the start of the freshmen year, 12 Fromme (2003a) noted a steep increase in the likelihood percent of males and females who drank in the previous 2 of blackouts above a BAC of 0.25 percent, calculated from weeks experienced a blackout during that time (White and self-reported consumption. Thus, from existing research, it Swartzwelder 2009). Data from the Harvard CAS indicate seems that the odds of blacking out increase as BAC levels that blackouts were experienced in a 30-day period by 25 climb and that blackouts become quite common at BAC percent of students in 1993 and 27 percent of students in levels approaching or exceeding 0.30 percent. As such, the 1997, 1999, and 2001 (Wechsler et al. 2002). A small study high prevalence of blackouts in college students points to by White and colleagues (2004), in which 50 students with the magnitude of excessive consumption that occurs in the histories of blackouts were interviewed, suggests that frag- college environment. It should be noted, however, that BAC mentary blackouts are far more common than en bloc black- levels calculated based on self-reported consumption are outs. Roughly 80 percent of students described their last unlikely to be accurate given the presence of partial or com- blackout as fragmentary. plete amnesia during the drinking occasion. Blackouts tend to occur following consumption of rela- It seems that some people are more sensitive to the tively large doses of alcohol and are more likely if one drinks effects of alcohol on memory than others and are therefore quickly and on an empty stomach, both of which cause a at increased risk of experiencing blackouts. Wetherill and rapid rise and high peak in BAC (Goodwin 1995; Perry et Fromme (2011) examined the effects of alcohol on contex- al. 2006). For this reason, pregaming, or prepartying, which tual memory in college students with and without a history typically involves fast-paced drinking prior to going out of blackouts. Performance on a task was similar while the to an event, increases the risk of blacking out. Labrie and groups were sober, but students with a history of blackouts colleagues (2011) reported that 25 percent of 2,546 students performed more poorly when intoxicated than those without who engaged in prepartying experienced at least one black- a history of blackouts. Similarly, Hartzler and Fromme out in the previous month. Playing drinking games and (2003b) reported that when mildly intoxicated, study partic- drinking shots were risk factors. Further, skipping meals ipants with a history of blackouts performed more poorly on to restrict calories on drinking days is associated with an a narrative recall task than those without a history of black- increased risk of blackouts and other consequences (Giles outs. When performing a memory task while sober, brain et al. 2009). activity measured with functional magnetic resonance imag- Because blackouts typically follow high peak levels of ing is similar in people with a history of blackouts and those drinking, it is not surprising that they are predictive of other without such a history (Wetherill et al. 2012). However, alcohol-related consequences. Mundt and colleagues (2012) when intoxicated, those with a history of blackouts exhibit examined past-year blackouts in a sample of more than 900 lower levels of activity in several regions of the frontal lobes students in a randomized trial of a screening and brief inter- compared with subjects without a history of blackouts. vention for problem alcohol use and found that blackouts Thus, studies suggest that there are differences in the predicted alcohol-related injuries over a subsequent 2-year effects of alcohol on memory and brain function between period. Compared with students who had no history of those who experience blackouts and those who do not. blackouts, those who reported one to two blackouts at baseline Research by Nelson and colleagues (2004), using data from were 1.5 times more likely to experience an alcohol-related monozygotic twins, suggests that there could be a significant

210 Alcohol Research: Current Reviews genetic component to these differences. Controlling for fre- In total, nearly 30,000 young people in this age-group, more quency of intoxication, the researchers found that if one males (19,847) than females (9,525) were hospitalized for twin experienced blackouts, the other was more likely than alcohol overdoses with no other drugs involved in 2008. chance to experience them as well. Further, Asian-American Hospitalizations for overdoses involving other drugs but students with the aldehyde dehydrogenase ALDH2*2 allele1 not alcohol increased 55 percent over the same time period, are less likely to experience blackouts than those without it, while those involving alcohol and drugs in combination rose even after adjusting for maximum number of drinks con- 76 percent. In total, there were 59,000 hospitalizations in sumed in a day (Lucsak et al. 2006). 2008 among 18- to 24-year-olds for alcohol overdoses only Several challenges hinder the assessment of blackouts or in combination with other drugs. Given that 33 percent and the events that transpire during them. Blackouts repre- of people in this age-group were full-time college students sent periods of amnesia. As such, it is difficult to imagine at 4-year colleges in 2008, a conservative estimate would that self-reported drinking levels are highly accurate for suggest approximately 20,000 hospitalizations for alcohol nights when blackouts occur. Further, in order for a person overdoses alone or in combination with other drugs involved to know what transpired during a blackout, and sometimes college students, although the exact number is not known. to be aware that a blackout occurred at all, they need to be Data from the Drug Abuse Warning Network (DAWN) told by other individuals. Often, the information provided indicate that ED visits for alcohol-related events increased in by these other individuals is unreliable as they were intoxi- a similar fashion as those observed for inpatient hospitaliza- cated themselves (Nash and Takarangi 2011). Thus, it is tions. Among those ages 18 to 20, ED visits for alcohol-related quite likely that self-reported rates and frequencies of black- events with no other drugs increased 19 percent, from 67,382 outs, drinking levels during nights in which blackouts occur, cases in 2005 to 82,786 cases in 2009. Visits related to com- and the rates of various types of consequences that occur bined use of alcohol and other drugs increased 27 percent, during them, are underestimated. from 27,784 cases in 2005 to 38,067 cases in 2009. In 2009, 12 percent of ED visits related to alcohol involved use of alcohol in combination with other drugs (SAMHSA 2011). Alcohol overdoses Alcohol interacts with a wide variety of illicit and pre- scription drugs, including opioids and related narcotic anal- When consumed in large quantities during a single occasion, gesics, sedatives, and tranquilizers (NIAAA 2007a; Tanaka such as a binge episode, alcohol can cause death directly by 2002). Importantly, BAC required for fatal overdoses are suppressing brain stem nuclei that control vital reflexes, like lower when alcohol is combined with prescription drugs. An breathing and gagging to clear the airway (Miller and Gold analysis of 1,006 fatal poisonings attributed to alcohol alone 1991). Even a single session of binge drinking causes inflam- or in combination with other drugs revealed that the median mation and transient damage to the heart (Zagrosek et al. postmortem BACs in those who overdosed on alcohol alone 2010). The acute toxic effects of alcohol in the body can was 0.33 percent, compared with 0.13 percent to 0.17 percent manifest in symptoms of alcohol poisoning, which include among those who overdosed on a combination of alcohol vomiting, slow and irregular breathing, hypothermia, mental and prescription drugs (Koski et al. 2003, 2005). The combined confusion, stupor, and death (NIAAA 2007b; Oster-Aaland use of alcohol and other drugs peaks in the 18- to 24-year- et al. 2009). Using data from the Global Burden of Disease old age range (McCabe et al. 2006), suggesting that college- Study, the World Health Organization (WHO) estimated aged young adults are at particularly high risk of suffering that, in 2002, alcohol poisoning caused 65,700 deaths consequences from alcohol-and-other-drug combinations. worldwide, with 2,700 poisoning deaths occurring in the The above findings reflect the fact that heavy consump- United States (WHO 2009). New stories about alcohol tion of alcohol quickly can become a medical emergency. overdoses among college students and their non-college One does not need to get behind the wheel of a car after peers have become increasingly common, a fact that is per- drinking or jump off a balcony into a swimming pool on a haps not surprising given the tendency toward excessive dare to risk serious harm. Simply drinking too much alcohol drinking in this age-group. is enough to require hospitalization and potentially cause To investigate the prevalence of hospitalizations for death. Further, combining alcohol with other drugs can alcohol overdoses—which stem from excessive intoxication increase the risk of requiring medical intervention substantially. or poisoning—among college-aged young people in the Thus, efforts to minimize the consequences of alcohol- United States, White and colleagues (2011) examined rates related harms on college campuses should not lose sight of inpatient hospitalizations for 18- to 24-year-olds between of the fact that alcohol often is combined with other drugs 1999 and 2008 using data from the Nationwide Inpatient and, when this is the case, the risks can be greater than when Sample, which contains hospital discharge records from alcohol or drugs are used alone. roughly 20 percent of all hospitals in the country. Hospitalizations for alcohol overdoses without any other drugs involved 1 the alDh2*2 allele results in decreased action by the enzyme acetaldehyde dehydrogenase, which is responsible for the breakdown of acetaldehyde. the accumulation of acetaldehyde after drinking increased 25 percent among 18- to 24-year-olds from 1999 alcohol leads to symptoms of acetaldehyde poisoning, such as facial flushing and increased heart to 2008, highlighting the risks involved in heavy drinking. and respiration rates.

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 211 Measuring the true scope of medical treatment for alcohol It is widely held that sexual assaults, with and without overdoses among college students is difficult for several rea- alcohol involvement, are underreported on college campuses. sons. First, in datasets such as the Nationwide Emergency Title IX of the Education Amendments Act of 1972, a Federal Department Sample (NEDS) and the Nationwide Inpatient civil rights law, requires universities to address sexual harass- Sample (NIS), no college identifiers are included to indicate ment and sexual violence. However, universities vary with whether a young person treated for an alcohol overdose is regard to how they handle such cases, and a student’s percep- enrolled in college. Many schools do not track or report the tion of safety and protection can influence the likelihood number of students treated for an alcohol overdose, and of reporting a sexual assault. Indeed, many universities have many students drink excessively when away from campus. indicated changes in rates of reports of assaults consistent Further, schools that implement Good Samaritan or Amnesty with changes in campus policies regarding how such cases policies, which allow students to get help for overly intoxi- are handled. As such, although it is clear that alcohol often cated peers without fear of sanctions, could create the false is involved in sexual assaults on college campuses, questions impression that overdoses are on the rise. For instance, after about the frequency and nature of such assaults remain. Cornell University implemented an amnesty policy, they witnessed an increase in calls to residence assistants and 911 for help dealing with an intoxicated friend (Lewis and Spring Break and 21st Birthday Celebrations— Marchell 2006). Given the dangerous nature of alcohol over- Event-Specific Drinking occasions doses, with or without other drugs involved, it is important to improve the tracking of these events at colleges and in the More college students drink, and drink more heavily, during larger community. specific celebratory events, such as spring break and 21st birthday celebrations, than during a typical week. Spring break is a roughly weeklong recess from school that takes Sexual Assault place in the spring at colleges throughout the United States. While some students continue to work, travel home, or simply Sexual assault is a pervasive problem on college campuses, relax, others use the opportunity to travel to beaches and and alcohol plays a central role in it. A study of roughly other party destinations. During spring break, approximately 5,500 college females on two campuses revealed that nearly 42 percent of students get drunk on at least 1 day, 11 percent 20 percent experienced some form of sexual assault while at drink to the point of blacking out or passing out, 32 percent college (Krebs et al. 2009). Data from the Harvard CAS sug- report hangovers, and 2 percent get into trouble with the gested that 5 percent of women surveyed were raped while police (Litt et al. 2013). Students with a history of binge at college (Mohler-Kuo et al. 2004). In a national sample of drinking and those intending to get drunk tend to drink students who completed the Core Alcohol and Drug Survey the heaviest (Patrick et al. 2013), suggesting that prevention in 2005, 82 percent of students who experienced unwanted efforts aimed at altering students’ intentions to get drunk sexual intercourse were intoxicated at the time. Similarly, while on spring break might lead to a reduction in peak nearly three-quarters (72 percent) of respondents to the drinking and the consequences that follow (Mallett et al. Harvard CAS study who reported being raped were intoxicated 2013). Interestingly, students who typically are light drinkers at the time. In many cases, rape victims are incapacitated by are more likely than those who typically are binge drinkers alcohol. In one study, 3.4 percent of rape victims reported to experience consequences from excessive drinking during being so intoxicated they were unable to consent (Mohler- spring break (Lee et al. 2009). Kuo et al. 2004). In a study of 1,238 college students on In addition to spring break, 21st birthday celebrations three campuses over a 3-year period, 6 percent of students are another event-specific opportunity for students to drink reported being raped while incapacitated by alcohol (Kaysen excessively. An estimated 4 out of 5 college students drink et al. 2006). alcohol to celebrate their 21st birthdays (Rutledge et al., Research suggests that the involvement of alcohol 2008) and many students drink more than they plan. Of increases the risk of being victimized in several ways, such 150 male and female college students surveyed about their as by impairing perceptions that one is in danger and by intentions to drink during their upcoming 21st birthday reducing the ability to respond effectively to sexual aggres- celebrations, 68 percent consumed more than they antici- sion (Abbey 2002; McCauley et al. 2010; Testa and Livingston pated while only 21 percent drank less and 11 percent were 2009). Further, alcohol might increase the chances that a accurate. On average, males intended to consume 8.5 drinks male will commit a sexual assault by leading them to misin- but consumed 12.5, while females expected to drink 7 but terpret a female’s friendly gestures or flirtation as interest in had 9 (Brister et al., 2010). As with spring-break drinking, sex and by increasing sexual aggression (Abbey 2002). When students with a history of binge drinking and those who asked to read a story about a potential date rape involving intended to drink heavily on their 21st birthday consumed intoxicated college students, both male and female subjects the most (Brister et al., 2011). In one study, roughly 12 who are intoxicated were more likely to view the female as percent of students reported consuming 21 or more drinks sexually aroused and the male as acting appropriately (Abbey while celebrating, and one-third of females (35 percent) and et al. 2003). nearly half of males (49 percent) reached estimated BACs

212 Alcohol Research: Current Reviews above 0.25 percent (Rutledge et al., 2008). Such high levels References of consumption substantially increase the odds of sexual assaults, fights, injuries, and death (Mallett et al., 2013). abbEy, a. alcohol-related sexual assault: a common problem among college students. Journal of Studies on Alcohol (suppl 14):118–128, 2002. PmiD: 12022717 Research indicates that brief interventions conducted in the week leading up to the 21st birthday celebration can reduce abbEy, a.; buCK, P.o.; zaWaCKi, t.; anD saEnz, C. alcohol’s effects on perceptions of a levels of consumption and associated consequences, suggesting potential date rape. Journal of Studies on Alcohol 64(5):669–677, 2003. PmiD: 14572189 that the risks of experiencing alcohol related consequences alEXanDER, E.n., anD boWEn, a.m. Excessive drinking in college: behavioral outcome, not stemming from 21st birthday celebrations could be partially binge, as a basis for prevention. Addictive Behaviors 29(6):1199–1205, 2004. PmiD: mitigated through specifically timed prevention efforts 15236823 (Neighbors et al. 2009, 2012). blanCo, C.; oKuDa, m.; WRight, C. Et al. mental health of college students and their non- college-attending peers: Results from the national Epidemiologic study on alcohol and Related Conditions. Archives of General Psychiatry 65(12):1429–1437, 2008. PmiD: 19047530 Summary boEKEloo, b.o.; noViK, m.g.; anD bush, E. Drinking to get drunk among incoming fresh- We have learned a considerable amount about the drinking men college students. American Journal of Health Education 42(2):88–95, 2011. habits of college students and the consequences that follow PmiD: 23440674 since NIAAA first reported on the matter in 1976. Surprisingly, boyER, K.K.; olson, J.R.; CalatonE, R.J.; anD JaCKson, E.C. Print versus electronic surveys: drinking levels have remained relatively stable on and around a comparison of two data collection methodologies. Journal of Operations Management 20(4):357–373, 2002. college campuses over the last 30 years, with roughly two out of five male and female students engaging in excessive, or bRistER, h.a.; WEthERill, R.R.; anD FRommE, K. anticipated versus actual alcohol consump- tion during 21st birthday celebrations. Journal of Studies on Alcohol and Drugs binge, drinking. Excessive drinking results in a wide range of 71(2):180–183, 2010. PmiD: 20230714 consequences, including injuries, assaults, car crashes, mem- ory blackouts, lower grades, sexual assaults, overdoses and bRoWn, s.a.; taPERt, s.F.; gRanholm, E.; anD DElis, D.C. neurocognitive functioning of adolescents: Effects of protracted alcohol use. Alcoholism: Clinical and Experimental death. Further, secondhand effects from excessive drinking Research 24(2):164–171, 2000. PmiD: 10698367 place non–binge-drinking students at higher risk of injury, sexual assaults, and having their studying disrupted. butlER, l.h., anD CoRREia, C.J. brief alcohol intervention with college student drinkers: Face-to face versus computerized feedback. Psychology of Addictive Behaviors Estimates of the rates of alcohol use and related conse- 23(1):163–167, 2009. PmiD: 19290702. quences are imperfect. Lack of knowledge of standard drink CamPbEll, C.a.; hahn, R.a.; ElDER, R.; Et al. the effectiveness of limiting alcohol outlet sizes and the effects of alcohol on memory formation all density as a means of reducing excessive alcohol consumption and alcohol-related complicate the collection of accurate data from traditional harms. American Journal of Preventive Medicine 37(6):556–569, 2009. PmiD: 19944925 self-report surveys. Underreporting of sexual assaults leads to CaREy, K.b.; sCott-shElDon, l.a.; CaREy, m.P.; anD DEmaRtini, K.s. individual-level interven- difficulty in estimating the true extent of the problem. Lack tions to reduce college student drinking: a meta-analytic review. Addictive Behaviors of college identifiers in mortality records and the fact that 32(11):2469–2494, 2007. PmiD: 17590277 alcohol levels are tested too infrequently in non–traffic-related CaREy, K.b.; sCott-shElDon, l.a.; Elliott, J.C.; Et al. Face-to-face versus computer-deliv- deaths leaves uncertainty regarding the actual number of ered alcohol interventions for college drinkers: a meta-analytic review, 1998 to 2010. college students who die each year from alcohol-related Clinical Psychology Review 32(8):690–703, 2012. PmiD: 23022767

causes. Similarly, college identifiers are not present in most CaRPEntER, C., anD DobKin, C. the minimum legal drinking age and public health. Journal crime reports and hospital reports. of Economic Perspectives 25(2):133–156, 2011. PmiD: 21595328 Although it is beyond the scope of this review to examine ChaVEz, P.R.; nElson, D.E.; naimi, t.s.; anD bREWER, R.D. impact of a new gender-specific efforts to prevent excessive drinking on college campuses, it definition for binge drinking on prevalence estimates for women. American Journal of should be noted that important strides have been made in Preventive Medicine 40(4):468–471, 2011. PmiD: 21406282 this area (Carey et al. 2012). In addition, data from MTF ClaPP, J.D.; Johnson, m.; Voas, R.b.; Et al. Reducing Dui among us college students: suggest that levels of binge drinking are decreasing among Results of an environmental prevention trial. Addiction 100(3):327–334, 2005. PmiD: 12th graders, particularly males. Hopefully, as our under- 15733246 standing of the nature of the problem continues to improve ClaPP, J.D.; min, J.W.; tRim, R.s.; Et al. Predictors of error in estimates of blood alcohol with better measurement strategies, improvements in pre- concentration: a replication. Journal of Studies on Alcohol and Drugs 70(5):683–688, vention approaches combined with declines in precollege 2009. PmiD: 19737492 drinking will lead to reductions in both the levels of alcohol CREgo, a.; RoDRiguEz-holguin, s.; PaRaDa, m., Et al. Reduced anterior prefrontal cortex acti- consumption by college students and the negative conse- vation in young binge drinkers during a visual working memory task. Drug and Alcohol quences that result. ■ Dependence 109(1–3):45–56, 2010. PmiD: 20079980

DEJong, W.; sChnEiDER, s.K.; toWVim, l.g.; Et al. a multisite randomized trial of social norms marketing campaigns to reduce college student drinking. Journal of Studies on Financial Disclosure Alcohol 67(6):868–879, 2006. PmiD: 17061004

DEJong, W.; sChnEiDER, s.K.; toWVim, l.g.; Et al. a multisite randomized trial of social The authors declare that they have no competing financial norms marketing campaigns to reduce college student drinking: a replication failure. interests. Substance Abuse 30(2):127–140, 2009. PmiD: 19347752

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 213 DEVos-Comby, l., anD langE, JJE. “my drink is larger than yours”? a literature review of hingson, R.W.; zaKoCs, R.C.; hEEREn, t.; Et al. Effects on alcohol related fatal crashes of a self-defined drink sizes and standard drinks. Current Drug Abuse Reviews community based initiative to increase substance abuse treatment and reduce alcohol 1(2):162–176, 2008. PmiD: 19630715 availability. Injury Prevention 11(2):84–90, 2005b. PmiD: 15805436

ElDER, R.W.; laWREnCE, b.; FERguson, a.; Et al. the effectiveness of tax policy interventions hingson, R.; zha, W.; anD WEitzman, E.R. magnitude of and trends in alcohol-related mor- for reducing excessive alcohol consumption and related harms. American Journal of tality and morbidity among u.s. college students ages 18-24, 1998-2005. Journal of Preventive Medicine 38(2):217–229, 2010. PmiD: 20117579 Studies on Alcohol and Drugs (suppl. 16):12–20, 2009. PmiD: 19538908

Elliott, J.C.; CaREy, K.b.; anD bollEs, J.R. Computer-based interventions for college drink- holDER, h.D.; gRuEnEWalD, P.J.; PoniCKi, W.R.; Et al. Effect of community-based interven- ing: a qualitative review. Addictive Behaviors 33(8):994–1005, 2008. PmiD: 18538484 tions on high-risk drinking and alcohol-related injuries. JAMA: Journal of the American Medical Association 284(18):2341–2347, 2000. PmiD: 11066184 Engs, R.C.; DiEbolD, b.a.; anD hanson, D.J. the drinking patterns and problems of a national sample of college students, 1994. Journal of Alcohol and Drug Education hoWlanD, J.; RohsEnoW, D.J.; gREECE, J.a.; Et al. the effects of binge drinking on college 41:13–33, 1996. students’ next-day academic test-taking performance and mood state. Addiction 105(4):655–665, 2010. PmiD: 20403018 FEll, J.C.; FishER, D.a.; Voas, R.b.; Et al. the impact of underage drinking laws on alcohol- related fatal crashes of young drivers. Alcoholism: Clinical and Experimental Research iChiyama, m.a.; FaiRliE, a.m.; WooD, m.D.; Et al. a randomized trial of a parent-based inter- 33(7):1208–1219, 2009. PmiD: 19389192 vention on drinking behavior among incoming college freshmen. Journal of Studies on Alcohol and Drugs (suppl. 16):67–76, 2009. PmiD: 19538914 FlEming, m.F.; balousEK, s.l.; gRossbERg, P.m.; Et al. brief physician advice for heavy drinking college students: a randomized controlled trial in college health clinics. JEnnison, K.m. the short-term effects and unintended long-term consequences of binge Journal of Studies on Alcohol and Drugs 71(1):23–31, 2010. PmiD: 20105410 drinking in college: a 10-year follow-up study. American Journal of Drug and Alcohol Abuse 30(3):659–684, 2004. PmiD: 15540499 gilEs, s.m.; ChamPion, h.; sutFin, E.l.; Et al. Calorie restriction on drinking days: an exam- ination of drinking consequences among college students. Journal of American College Johnston, l.D.; o’mallEy, P.m.; anD baChman, J.g. Monitoring the Future National Survey Health 57(6):603–609, 2009. PmiD: 19433398 Results on Drug Use, 1975–2000. Volume I: Secondary School Students. nih Publication no. 01–4924. bethesda, mD: national institute on Drug abuse, 2001a. gill, J.s. Reported levels of alcohol consumption and binge drinking within the uK undergraduate student population over the last 25 years. Alcohol and Alcoholism Johnston l.D.; o’mallEy, P.m.; anD baChman, J.g. Monitoring the Future National Survey 37(2):109–120, 2002. PmiD: 11912065 Results on Drug Use,1975–2000. Volume II: College Students and Adults Ages 19–40. nih Publication no. 01–4925. bethesda, mD: national institute on Drug abuse, 2001b. gooDWin, D.W. alcohol amnesia. Addiction 90(3):315–317, 1995. PmiD: 7735016 Johnston, l.D.; o’mallEy, P.m.; baChman, J.g.; anD sChulEnbERg, J.E. Monitoring the Future hanson, K.l.; mEDina, K.l.; PaDula, C.b.; Et al. impact of adolescent alcohol and drug use National Survey Results on Drug Use, 1975–2006. Volume II: College Students and on neuropsychological functioning in young adulthood: 10-year outcomes. Journal of Adults Ages 19–45. nih Publication no. 07–6206. bethesda, mD: national institute on Child & Adolescent Substance Abuse 20(2):135–154, 2011. PmiD: 21532924 Drug abuse, 2007.

haRtzlER, b., anD FRommE, K. Fragmentary and en bloc blackouts: similarity and distinc- Johnston, l.D.; o’mallEy, P.m.; baChman, J.g.; anD sChulEnbERg, J.E. Monitoring the Future tion among episodes of alcohol-induced memory loss. Journal of Studies on Alcohol National Survey Results on Drug Use, 1975–2011. Volume II: College Students and 64(4):547–550, 2003a. PmiD: 12921196 Adults Ages 19–50. ann arbor: institute for social Research, the university of michigan, 2012. haRtzlER, b., anD FRommE, K. Fragmentary blackouts: their etiology and effect on alcohol

expectancies. Alcoholism: Clinical and Experimental Research 27(4):628–637, 2003b. JonEs, R., anD laCEy, J. alcohol and highway safety 2001: A Review of the State of PmiD: 12711925 Knowledge. Report no. Dot hs–809–383. Washington, D.C.: national highway traffic safety administration, 2001. hingson, R.W.; hEEREn, t.; EDWaRDs, E.m.; anD saitz, R. young adults at risk for excess

alcohol consumption are often not asked or counseled about drinking alcohol. Journal JonEs, R., anD Pitt, n. health surveys in the workplace: Comparison of postal, email and of General Internal Medicine 27(2):179–184, 2012. PmiD: 21935753 World Wide Web methods. Occupational Medicine (London) 49(8):556–558, 1999. PmiD: 10658310 hingson, R.; hEEREn, t.; anD WintER, m. lower legal blood-alcohol limits for young drivers. Public Health Reports 109(6):738–744, 1994. PmiD: 7800781 KaysEn, D.; nEighboRs, C.; maRtEll, J.; Et al. incapacitated rape and alcohol use: a prospective analysis. Addictive Behaviors 31(10):1820–1832, 2006. PmiD: 16446044 hingson, R.; hEEREn, t.; WintER, m.R.; anD WEChslER, h. Early age of first drunkenness as a factor in college students’ unplanned and unprotected sex attributable to drinking. KERR, W.C.; anD stoCKWEll, t. understanding standard drinks and drinking guidelines. Pediatrics 111(1):34–41, 2003a. PmiD: 12509551 Drug and Alcohol Review 31(2):200–205, 2012. PmiD: 22050262

hingson, R.; hEEREn, t.; WintER, m.; anD WEChslER, h. magnitude of alcohol-related mortal- Knight, J.R.; WEChslER, h.; Kuo, m.; Et al. alcohol abuse and dependence among u.s. ity and morbidity among u.s. college students ages 18-24: Changes from 1998 to college students. Journal of Studies on Alcohol 63(3):263–270, 2002. PmiD: 12086126 2001. Annual Review of Public Health 26: 259–279, 2005a. PmiD: 15760289 KosKi, a.; oJanPERä, i.; anD VuoRi, E. interaction of alcohol and drugs in fatal poisonings. hingson, R.; hEEREn, t.; zaKoCs, R., Et al. age of first intoxication, heavy drinking, driving Human & Experimental Toxicology 22(5):281–287, 2003. PmiD: 12774892 after drinking and risk of unintentional injury among u.s. college students. Journal of KosKi, a,; VuRoi, E.; anD oJanPERä, i. Relation of postmortem blood alcohol and drug con- Studies on Alcohol 64(1):23–31, 2003b. PmiD: 12608480 centrations in fatal poisonings involving amitriptyline, propoxyphene and promazine. hingson, R.; hEEREn, t.; zaKoCs, R.; Et al. magnitude of alcohol-related mortality and mor- Human & Experimental Toxicology 24(8):389–396, 2005. PmiD: 16138729 bidity among u.s. college students ages 18-24. Journal of Studies on Alcohol 63(2): KRaus, C.l.; salazaR, n.C.; mitChEll, J.R.; Et al. inconsistencies between actual and esti- 136–144, 2002. PmiD: 12033690 mated blood alcohol concentrations in a field study of college students: Do students hingson, R.; mCgoVERn, t.; hoWlanD, J.; Et al. Reducing alcohol-impaired driving in really know how much they drink? Alcoholism: Clinical and Experimental Research massachusetts: the saving lives Program. American Journal of Public health 29(9):1672–1676, 2005. PmiD: 16205367 86(6):791–797, 1996. PmiD: 8659651 KREbs, C.P.; linDQuist, C.h.; WaRnER, t.D.; Et al. College women’s experiences with physi- hingson, R., anD WhitE, a. magnitude and prevention of college alcohol and drug mis- cally forced, alcohol- or drug-enabled, and drug-facilitated sexual assault before and use: u.s. college students ages 18-24. in: Kay, J., and schwartz, V., Eds. Mental Health since entering college. Journal of American College Health 57(6):639–647, 2009. Care in the College Community. london: John Wiley & sons, 2010, pp. 289–324. PmiD: 19433402

214 Alcohol Research: Current Reviews labRiE, J.; EaRlEyWinE, m.; lamb, t.; anD shElEsKy, K. Comparing electronic-keypad respons- moREiRa, m.t.; smith, l.a.; anD FoXCRoFt, D. social norms interventions to reduce alcohol es to paper-and-pencil questionnaires in group assessments of alcohol consumption misuse in university or college students. Cochrane Database of Systematic Reviews and related attitudes. Addictive Behaviors 31(12):2334-2338, 2006. PmiD: 16626878 (Online) (3):CD006748, 2009. PmiD: 19588402 labRiE, J.W.; huChting, K.; taWalbEh, s.; Et al. a randomized motivational enhancement munDt, m.P.; anD zaKlEtsKaia, l.i. Prevention for college students who suffer alcohol- prevention group reduces drinking and alcohol consequences in first-year college induced blackouts could deter high-cost emergency department visits. Health Affairs women. Psychology of Addictive Behaviors 22(1):149–155, 2008. PmiD: 18298242 (Millwood) 31(4):863–870, 2012. labRiE, J.W.; hummER, J.; KEnnEy, s.; Et al. identifying factors that increase the likelihood munDt, m.P.; zaKlEtsKaia, l.i.; bRoWn, D.D.; anD FlEming, .mF. alcohol-induced memory for alcohol-induced blackouts in the prepartying context. Substance Use & Misuse blackouts as an indicator of injury risk among college drinkers. Injury Prevention 46(8):992–1002, 2011. PmiD: 21222521 18(1):44–49, 2012. PmiD: 21708813 laRimER, m.E., anD CRonCE, J.m. identification, prevention and treatment: a review of naimi, t.s.; nElson, D.E.; anD bREWER, R.D. the intensity of binge alcohol consumption individual-focused strategies to reduce problematic alcohol consumption by college among u.s. adults. American Journal of Preventive Medicine 38(2):201–207, 2010. students. Journal of Studies on Alcohol (suppl. 14):148–163, 2002. PmiD: 12022721 PmiD: 20117577

laRimER, m.E., anD CRonCE, J.m. identification, prevention, and treatment revisited: nash, R.a., anD taKaRangi, m.K. Reconstructing alcohol-induced memory blackouts. individual-focused college drinking prevention strategies 1999-2006. Addictive Memory 19(6):566–573, 2011. PmiD: 21919584 Behaviors 32(11):2439–2468, 2007. PmiD: 17604915 national institute on alcohol abuse and alcoholism (niaaa). Alcohol Policy Information lEE, C.m.; lEWis, m.a.; anD nEighboRs, C. Preliminary examination of spring break alcohol System, 2010. available at http://www.alcoholpolicy.niaaa.nih.gov/. accessed February use and related consequences. Psychology of Addictive Behaviors 23(4):689–694, 16, 2010. 2009. PmiD: 20025375 niaaa. NIAAA Newsletter, Winter 2004, Number 3. bethesda, mD: national institutes of liang, l., anD huang, J. go out or stay in? the effects of zero tolerance laws on alcohol health, u.s. Department of health and human services (Dhhs), 2004. available at: use and drinking and driving patterns among college students. Health Economics http://pubs.niaaa.nih.gov/publications/newsletter/winter2004/newsletter_number3.pdf 17(11):1261–1275, 2008. PmiD: 18219708 niaaa. Harmful Interactions: Mixing Alcohol with Medicines. bethesda, mD: national litt, D.m.; lEWis, m.a.; PatRiCK, m.E.; Et al. spring break versus spring broken: Predictive institutes of health, u.s. Department of health and human services (Dhhs), revised utility of spring break alcohol intentions and willingness at varying levels of extremity. 2007a (nih Publication no. 07-5329). available at: http://pubs.niaaa.nih.gov/publica- Prevention Science, 2013 Feb 13. [Epub ahead of print]. PmiD: 23404667 tions/medicine/medicine.htm luCzaK, s.E.; shEa, s.h.; hsuEh, a.C.; Et al. alDh2*2 is associated with a decreased likeli- niaaa. Parents – Spring Break Is Another Important Time to Discuss College Drinking. hood of alcohol-induced blackouts in asian american college students. Journal of bethesda, mD: national institutes of health, Dhhs, 2007b (nih Publication no. 05-5642). Studies on Alcohol 67(3):349–353, 2006. PmiD: 16608143 niaaa. task Force of the national advisory Council on alcohol abuse and alcoholism. lygiDaKis, C.; Rigon, s.; Cambiaso, s.; Et al. a web-based versus paper questionnaire on A Call to Action: Changing the Culture of Drinking at U.S. Colleges. bethesda, mD: alcohol and tobacco in adolescents. Telemedicine Journal and E-Health 16(9):925– national institutes of health, u.s. Department of health and human services (Dhhs), 930, 2010. PmiD: 20958200 2002 (nih Publication no. 02–5010).

mallEtt, K.a.; VaRVil-WElD, l.; boRsaRi, b.; Et al. an update of research examining college niaaa. What Colleges Need to Know Now: An Update on College Drinking Research. student alcohol-related consequences: new perspectives and implications for interventions. bethesda, mD: national institutes of health, u.s. Department of health and human Alcoholism: Clinical and Experimental Research 37(5):709–716, 2013. PmiD: 23241024 services (Dhhs), 2007c (nih Publication no. 07–5010).

mCCabE, s.E.; CRanFoRD, J.a; anD boyD, C.J. the relationship between past-year drinking national Research Council institute of medicine of the national academies. Reducing behaviors and nonmedical use of prescription drugs: Prevalence of co-occurrence in a Underage Drinking: A Collective Responsibility. Washington, DC: the national national sample. Drug and Alcohol Dependence 84(3):281–288, 2006. PmiD: 16621337 academies Press, 2004.

mCCaRtt, a.t.; hEllinga, l.a.; anD WElls, J.K. Effects of a college community campaign on nEighboRs, C.; lEE, C.m.; lEWis, m.a.; Et al. internet-based personalized feedback to drinking and driving with a strong enforcement component. Traffic Injury Prevention reduce 21st-birthday drinking: a randomized controlled trial of an event-specific prevention 10(2):141–147, 2009. PmiD: 19333826 intervention. Journal of Consulting and Clinical Psychology 77(1):51–63, 2009. PmiD: 19170453 mCCaulEy, J.l.; Calhoun, K.s.; anD giDyCz, C.a. binge drinking and rape: a prospective examination of college women with a history of previous sexual victimization. Journal nEighboRs, C.; lEE, C.m.; atKins, D.C.; Et al. a randomized controlled trial of event-specific of Interpersonal Violence 25(9):1655–1668, 2010. PmiD: 20068115 prevention strategies for reducing problematic drinking associated with 21st birthday celebrations. Journal of Consulting and Clinical Psychology 80(5):850–862, 2012. mEDina, K.l.; mCQuEEny, t.; nagEl, b.J.; Et al. Prefrontal cortex volumes in adolescents PmiD: 22823855 with alcohol use disorders: unique gender effects. Alcoholism: Clinical and Experimental Research 32(3):386–394, 2008. PmiD: 18302722 nElson, E.C.; hEath, a.C.; buCholz, K.K.; Et al. genetic epidemiology of alcohol-induced blackouts. Archives of General Psychiatry 61(3):257–263, 2004. PmiD: 14993113 mEilman, P.W.; lEiChlitER, J.s.; anD PREslEy, C.a. greeks and athletes: Who drinks more? Journal of American College Health 47(4):187–190, 1999. PmiD: 9919850 noRbERg, K.E.; biERut, l.J.; anD gRuCza, R.a. long-term effects of minimum drinking age laws on past-year alcohol and drug use disorders. Alcoholism: Clinical and Experimental mEilman, P.W.; PREslEy, C.a.; anD Cashin, J.R. the sober life at the historically black colleges. Research 33(12):2180–2190, 2009. PmiD: 19775322 Journal of Blacks in Higher Education 9:98–100, 1995. noRthCotE, J., anD liVingston, m. accuracy of self-reported drinking: observational verifi- mEilman, P.W.; PREslEy, C.a.; anD lyERla, R. black college students and binge drinking. cation of ‘last occasion’ drink estimates of young adults. Alcohol and Alcoholism Journal of Blacks in Higher Education 8:70–71, 1994. 46(6):709–713, 2011. PmiD: 21949190 millER, n.s.; anD golD, m.s. Alcohol. new york: Plenum Publishing Company, 1991. o’bRiEn, m.C.; mCCoy, t.P.; ChamPion, h.; Et al. single question about drunkenness to miRon, J.a., anD tEtElbaum, E. Does the minimum legal drinking age save lives? detect college students at risk for injury. Academic Emergency Medicine 13(6):629– Economic inquiry 47(2):317–336, 2009. 636, 2006. PmiD: 16614453

mohlER-Kuo, m.; DoWDall, g.b.; Koss, m.P.; anD WEChslER, h. Correlates of rape while o’haRE, t.m. Drinking in college: Consumption patterns, problems, sex differences and intoxicated in a national sample of college women. Journal of Studies on Alcohol legal drinking age. Journal of Studies on Alcohol 51(6):536–541, 1990. PmiD: 65(1):37–45, 2004. PmiD: 15000502 2270062

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 215 ostER-aalanD, l.; lEWis, m.a.; nEighboRs, C.; Et al. alcohol poisoning among college stu- sChulEnbERg, J.; maggs, J.l.; long, s.W.; Et al. the problem of college drinking: insights dents turning 21: Do they recognize the symptoms and how do they help? Journal of from a developmental perspective. Alcoholism: Clinical and Experimental Research Studies on Alcohol and Drugs (suppl 16):122–130, 2009. PmiD: 19538920 25(3):473–477, 2001. PmiD: 11290861

PaRaDa, m.; CoRRal, m.; Caamaño-isoRna, F.; Et al. binge drinking and declarative memory sChWEinsbuRg, a.D.; mCQuEEny, t.; nagEl, b. J.; Et al. a preliminary study of functional in university students. alcoholism: Clinical and Experimental Research 35(8):1475– magnetic resonance imaging response during verbal encoding among adolescent 1484, 2011. PmiD: 21575014 binge drinkers. Alcohol 44(1):111–117, 2010. PmiD: 20113879

PaRaDa, m.; CoRRal, m.; mota, n. Et al. Executive functioning and alcohol binge drinking sCRibnER, R.a.; mason, K.E.; simonsEn, n.R.; Et al. an ecological analysis of alcohol outlet in university students. Addictive Behaviors 37(2):167–172, 2012. PmiD: 21996093 density and campus-reported violence at 32 u.s. colleges. Journal of Studies on PasChall, m.J.; antin, t.; RingWalt, C.l.; anD saltz, R.F. Effects of alcoholEdu for College Alcohol and Drugs 71(2):184–191, 2010. PmiD: 20230715 on alcohol-related problems among freshmen: a randomized multicampus trial. Journal sCRibnER, R.; mason, K.; thEall, K.; Et al. the contextual role of alcohol outlet density in of Studies on Alcohol and Drugs 72(4):642–650, 2011a. PmiD: 21683046 college drinking. Journal of Studies on Alcohol and Drugs 69(1):112–120, 2008. PmiD: PasChall, m.J.; antin, t.; RingWalt, C.l.; anD saltz, R.F. Evaluation of an internet-based 18080071 alcohol misuse prevention course for college freshmen: Findings of a randomized multi-campus trial. American Journal of Preventive Medicine 41(3):300–308, 2011b. sCRibnER, R.a.; thEall, K.P.; mason, K.; Et al. alcohol prevention on college campuses: PmiD: 21855745 the moderating effect of the alcohol environment on the effectiveness of social norms marketing campaigns. Journal of Studies on Alcohol and Drugs 72(2):232-239, 2011. PasChall, m.J., anD FREisthlER, b. Does heavy drinking affect academic performance in PmiD: 21388596 college? Findings from a prospective study of high achievers. Journal of Studies on Alcohol 64(4):515–519, 2003. PmiD: 12921193 shults, R.a.; ElDER, R.W.; slEEt, D.a.; Et al. Reviews of evidence regarding interventions to reduce alcohol-impaired driving. American Journal of Preventive Medicine 21(suppl. PatRiCK, m.E.; lEWis, m.a.; lEE, C.m.; anD maggs, J.l. semester and event-specific motives 4):66-88, 2001. PmiD: 11691562 for alcohol use during Spring Break: Associated protective strategies and negative con- sequences. Addictive Behaviors 38(4):1980–1987, 2013. PmiD: 23384451 singlEton, R.a. Collegiate alcohol consumption and academic performance. Journal of Studies on Alcohol and Drugs 68(4):548–555, 2007. PmiD: 17568960 PERRy, P.J.; aRgo, t.R.; baRnEtt, m.J.; Et al. the association of alcohol-induced blackouts and grayouts to blood alcohol concentrations. Journal of Forensic Sciences 51(4):896– singlEton, R.a. JR., anD WolFson, a.R. alcohol consumption, sleep, and academic perfor- 899, 2006. PmiD: 16882236 mance among college students. Journal of Studies on Alcohol and Drugs 70(3):355–

PoWEll, l.a.; Williams, J.; anD WEChslER, h. study habits and the level of alcohol use 363, 2009. PmiD: 19371486 among college students. Education Economics 12(2):135–149, 2004. smith, g.s.; bRanas, C.C.; anD millER, t.R. Fatal nontraffic injuries involving alcohol: a PREslEy, C.a.; lEiChlitER, J.s.; anD mEilman, P.W. Alcohol and Drugs on American College metaanalysis. Annals of Emergency Medicine 33(6):659–668, 1999. PmiD: 10339681 Campuses: A Report to College Presidents. Third in a Series: 1995, 1996, and 1997, sQuEglia, l.m.; sPaDoni, a.D.; inFantE, m.a.; Et al. initiating moderate to heavy alcohol use Carbondale, il: Core institute, southern illinois university, 1998. predicts changes in neuropsychological functioning for adolescent girls and boys. PREslEy, C.a.; mEilman, P.W.; anD Cashin, J.R. Alcohol and Drugs on American College Psychology of Addictive Behaviors 23(4):715–722, 2009. PmiD: 20025379 Campuses: Use, Consequences, and Perceptions of the Campus Environment. Volume sQuEglia, l.m.; soRg, s.F.; sChWEinsbuRg, a.D.; Et al. binge drinking differentially affects IV: 1992-94. Carbondale, il: the Core institute, 1996a. adolescent male and female brain morphometry. Psychopharmacology (Berlin) PREslEy, C.a.; mEilman, P.W.; Cashin, J.R.; anD lyERa, R. Alcohol and Drugs on American 220(3):529–539, 2012. PmiD: 21952669 College Campuses: Use, Consequences, and Perceptions of the Campus Environment. Volume III: 1991-93. Carbondale, il: the Core institute, 1996b. stRaus, R., anD baCon, s. D. (1953). Drinking in college. new haven, Ct: yale university Press. PREslEy, C.a.; mEilman, P.W., anD lEiChlitER, J.s. College factors that influence drinking. Journal of Studies on Alcohol (suppl 14):82–90, 2002. PmiD: 12022732 substance abuse and mental health services administration (samhsa). The DAWN Report: Trends in Emergency Department Visits Involving Underage Alcohol Use: 2005 PREslEy, C.a., anD PimEntEl, E.R. the introduction of the heavy and frequent drinker: a to 2009. Rockville, mD, substance abuse and mental health services administration, proposed classification to increase accuracy of alcohol assessments in postsecondary 2011 (accessed september 13, 2011). http://www.samhsa.gov/data/2k11/WEb_ educational settings. Journal of Studies on Alcohol 67(2):324–331, 2006. PmiD: DaWn_020/WEb_DaWn_020_html.pdf 16562416 substance abuse and mental health services administration (samhsa). Results from PREussER, D.; ulmER, R.; anD PREissER, C. Obstacles to Enforcement of Youthful (Under 21) Impaired Driving. Washington, DC: national highway traffic safety administration, 1992 the 2001 National Household Survey on Drug Abuse: Volume 1: Summary of National (Dot hs 807–878). Findings. Rockville, mD: substance abuse and mental health services administration, 2002 (Dhhs Publication no. sma 02–3758, 2002). REaD, J.P.; bEattiE, m.; ChambERlain, R.; anD mERRill, J.E. beyond the “binge” threshold: heavy drinking patterns and their association with alcohol involvement indices in col- substance abuse and mental health services administration (samhsa). Results from lege students. Addictive Behaviors 33(2):225–234, 2008. PmiD: 17997047 the 2005 National Survey on Drug Use and Health: National Findings. Rockville, mD: substance abuse and mental health services administration, 2006 (Dhhs Publication REboussin, b.a.; song, E.y.; anD WolFson, m. the impact of alcohol outlet density on the no. sma 06–4194). geographic clustering of underage drinking behaviors within census tracts. Alcoholism: Clinical and Experimental Research 35(8):1541–1549, 2011. PmiD: 21463343 substance abuse and mental health services administration (samhsa). Results from the 2007 national survey on Drug use and health: national Findings. Rockville, mD: RutlEDgE, P.C.; PaRK, a.; anD shER, K.J. 21st birthday drinking: Extremely extreme. Journal of Consulting and Clinical Psychology 76(3):511–516, 2008. PmiD: 18540744 substance abuse and mental health services administration, 2008 (nsDuh series h- 34, Dhhs Publication no. sma 08–4343). saltz, R.F.; WElKER, l.R.; PasChall, m.J.; Et al. Evaluating a comprehensive campus-com- munity prevention intervention to reduce alcohol-related problems in a college popula- substance abuse and mental health services administration (samhsa). Summary of tion. Journal of Studies on Alcohol and Drugs (suppl. 16):21–27, 2009. PmiD: Findings of the 1999 National Household Survey on Drug Abuse. Rockville, mD: 19538909 substance abuse and mental health services administration, 2000 (Dhhs Publication no. sma 00–3466). sChaus, J.F.; solE, m.l.; mCCoy, t.P.; Et al. alcohol screening and brief intervention in a college student health center: a randomized controlled trial. Journal of Studies on tanaKa, E. toxicological interactions between alcohol and benzodiazepines. Journal of Alcohol and Drugs (suppl. 16):131–141, 2009. PmiD: 19538921 Toxicology. Clinical Toxicology 40(1):69–75, 2002. PmiD: 11990206

216 Alcohol Research: Current Reviews tEsta, m., anD liVingston, J.a. alcohol consumption and women’s vulnerability to sexual WEChslER, h.; DoWDall, g.W.; DaVEnPoRt, a.; anD Rimm, E.b. a gender-specific measure of victimization: Can reducing women’s drinking prevent rape? Substance Use & Misuse binge drinking among college students. American Journal of Public Health 85(7):982– 44(9-10):1349–1376, 2009. PmiD: 19938922 985, 1995. PmiD: 7604925

thombs, D.l.; olDs, R.s.; bonDy, s.J.; Et al. undergraduate drinking and academic perfor- WEChslER, h.; DoWDall, g.W.; maEnnER, g.; Et al. Changes in binge drinking and related mance: a prospective investigation with objective measures. Journal of Studies on problems among american college students between 1993 and 1997. Results of the Alcohol and Drugs 70(5):776–785, 2009. PmiD: 19737503 harvard school of Public health College alcohol study. Journal of American College Health 47(2):57–68, 1998. PmiD: 9782661 timbERlaKE, D.s.; hoPFER,C.J.; RhEE, s.h.; Et al. College attendance and its effect on drink- ing behaviors in a longitudinal study of adolescents. Alcoholism: Clinical and WEChslER, h., anD Kuo, m. Watering down the drinks: the moderating effect of college demographics on alcohol use of high-risk groups. American Journal of Public Health, Experimental Research 31(6):1020–1030, 2007. PmiD: 17403064 93(11):1929–1933, 2003. PmiD: 14600068 tREno, a.J.; gRuEnEWalD P.J.; lEE, J.P.; anD REmER, l.g. the sacramento neighborhood WEChslER, h.; Kuh, g.; anD DaVEnPoRt, a.E. Fraternities, sororities and binge drinking: alcohol Prevention Project: outcomes from a community prevention trial. Journal of Results from a national study of american colleges. NASPA Journal 46(3):395–416, 2009. Studies on Alcohol and Drugs 68(2):197–207, 2007. PmiD: 17286338 WEChslER, h.; lEE, J.E.; Kuo, m., Et al. trends in college binge drinking during a period of tuRnER, C.F.; Ku, l.; RogERs, s.m.; Et al. adolescent sexual behavior, drug use, and vio- increased prevention efforts. Findings from 4 harvard school of Public health College lence: increased reporting with computer survey technology. Science 280(5365):867– alcohol study surveys: 1993-2001. Journal of American College Health 50(5):203– 873, 1998. PmiD: 9572724 217, 2002. PmiD: 11990979 tuRRisi, R.; laRimER, m.E.; mallEtt, K.a.; Et al. a randomized clinical trial evaluating a com- WEChslER, h.; lEE, J.E.; Kuo, m.; anD lEE, h. College binge drinking in the 1990s: a con- bined alcohol intervention for high-risk college students. Journal of Studies on Alcohol tinuing problem. Results of the harvard school of Public health 1999 College alcohol and Drugs 70(4):555–567, 2009. PmiD: 19515296 survey. Journal of American College Health 48(5):199–210, 2000. PmiD: 10778020

uPCRaFt, m.l. Today’s First-Year Students and Alcohol. Paper prepared for the task Force WEChslER, h., anD nElson, t.F. What we have learned from the harvard school of Public on College Drinking, national advisory Council on alcohol abuse and alcoholism, health College alcohol study: Focusing attention on college student alcohol consump- bethesda, mD, 2000. tion and the environmental conditions that promote it. Journal of Studies on Alcohol and Drugs 69(4): 481–490, 2008. PmiD: 18612562 VaRVil-WElD, l.; mallEtt, K.a.; tuRRisi, R.; Et al. are certain college students prone to expe- riencing excessive alcohol-related consequences? Predicting membership in a high-risk WEChslER, h., anD nElson, t. F. binge drinking and the american college students: What’s subgroup using pre-college profiles. Journal of Studies on Alcohol and Drugs five drinks? Psychology of Addictive Behaviors, 15(4), 287–291, 2001. PmiD: 11767258

74(4):542–551, 2013. PmiD: 23739017 WEitzman, E.R.; FolKman, a; FolKman, m.P.; anD WEChslER, h. the relationship of alcohol outlet density to heavy and frequent drinking and drinking-related problems among col- Voas, R.b., anD Williams, a.F. age differences of arrested and crash-involved drinking lege students at eight universities. Health & Place 9(1):1–6, 2003. PmiD: 12609468 drivers. Journal of Studies on Alcohol 47(3):244–248, 1986. PmiD: 3724162 WEthERill, R.R., anD FRommE, K. acute alcohol effects on narrative recall and contextual Voas, R.b.; tiPPEtts, a.s.; anD FEll, J. the relationship of alcohol safety laws to drinking memory: an examination of fragmentary blackouts. Addictive Behaviors 36(8):886– drivers in fatal crashes. Accident Analysis and Prevention 32(4):483–492, 2000. PmiD: 889, 2011. PmiD: 21497445 10868751 WEthERill, R.R.; sChnyER, D.m.; anD FRommE, K. acute alcohol effects on contextual memo- WagEnaaR, a.C., anD toomEy, t.l. Effects of minimum drinking age laws: Review and ry bolD response: Differences based on fragmentary blackout history. Alcoholism: analyses of the literature from 1960 to 2000. Journal of Studies on Alcohol (suppl. Clinical and Experimental Research 36(6):1108–1115, 2012. PmiD: 22420742 14):206–225, 2002. PmiD: 12022726 WhitE, a.m. What happened? alcohol, memory blackouts, and the brain. Alcohol WagEnaaR, a.C.; muRRay, D.m.; anD toomEy t.l. Communities mobilizing for Change on Research & Health 27(2):186-196, 2003. PmiD: 15303630 alcohol (CmCa): Effects of a randomized trial on arrests and traffic crashes. Addiction 95(2):209–217, 2000. PmiD: 10723849 WhitE, a.m.; hingson, R.W.; Pan, i.J.; anD yi, h.y. hospitalizations for alcohol and drug overdoses in young adults ages 18-24 in the united states, 1999-2008: Results from WagEnaaR, a.C.; o’mallEy, P.m.; anD laFonD, C. lowered legal blood alcohol limits for the nationwide inpatient sample. Journal of Studies on Alcohol and Drugs 72(5):774– young drivers: Effects on drinking, driving, and driving-after-drinking behaviors in 30 786, 2011. PmiD: 21906505 states. American Journal of Public Health 91(5):801–804, 2001. PmiD: 11344892 WhitE, a.m.; JamiEson–DRaKE, D.W.; anD sWaRtzWElDER, h.s. Prevalence and correlates of WagEnaaR, a.C.; salois, m.J.; anD KomRo, K.a. Effects of beverage alcohol price and tax alcohol–induced blackouts among college students: Results of an e–mail survey. levels on drinking: a meta-analysis of 1003 estimates from 112 studies. Addiction Journal of American College Health 51(3):117–119; 122-131, 2002. PmiD: 12638993 104(2):179–190, 2009. PmiD: 19149811 WhitE, a.m.; KRaus, C.l.; Flom, J.D.; Et al. College students lack knowledge of standard

WagEnaaR, a.C.; ERiCKson, D.J.; haRWooD, E.m.; anD o’mallEy, P.m. Effects of state coali- drink volumes: implications for definitions of risky drinking based on survey data. tions to reduce underage drinking: a national evaluation. American Journal of Alcoholism: Clinical and Experimental Research 29(4):631–638, 2005. PmiD: 15834229 Preventive Medicine 31(4):307–315, 2006. PmiD: 16979455 WhitE, a.m.; KRaus, C.l.; mCCRaCKEn, l.a.; anD sWaRtzWElDER, h.s. Do college students drink more than they think? use of a free-pour paradigm to determine how college Wang, y.C.; lEE, C.m; lEW-ting, C.y.; Et al. survey of substance use among high school students define standard drinks. Alcoholism: Clinical and Experimental Research students in taipei: Web-based questionnaire versus paper-and-pencil questionnaire. 27(11):1750–1756, 2003. PmiD: 14634490 Journal of Adolescent Health 37(4):289–295, 2005. PmiD: 16182139 WhitE, a.m.; KRaus, C.l.; anD sWaRtzWElDER, h. many college freshmen drink at levels far WEChslER, h.; DaVEnPoRt, a.; DoWDall, g.; Et al. health and behavioral consequences of beyond the binge threshold. alcoholism: Clinical and Experimental Research binge drinking in college: a national survey of students at 140 campuses. JAMA: Journal 30(6):1006-1010, 2006. PmiD: 16737459 of the American Medical Association 272(21):1672–1677, 1994. PmiD: 7966895 WhitE, a.m.; signER, m.l.; KRaus, C.l.; anD sWaRtzWElDER, h.s. Experiential aspects of WEChslER, h.; DaVEnPoRt, a.E.; DoWDall, g.W.; Et al. binge drinking, tobacco, and illicit alcohol–induced blackouts among college students. American Journal of Drug and drug use and involvement in college athletics. a survey of students at 140 american Alcohol Abuse 30(1):205–224, 2004. PmiD: 15083562 colleges. Journal of American College Health 45(5):195–200, 1997. PmiD: 9069676 WhitE, a.m., anD sWaRtzWElDER, h.s. inbound college students drink heavily during the WEChslER, h.; DoWDall, g.W.; DaVEnPoRt, a.; anD Castillo, s. Correlates of college student summer before their freshman year: implications for education and prevention efforts. binge drinking. American Journal of Public Health 85(7):921–926, 1995. PmiD: 7604914 American Journal of Health Education 40:909–96, 2009.

the Burden of Alcohol Use: Excessive Alcohol Consumption and Related Consequences Among College Students 217 WILLIAMS, J.; POWELL, L.M.; AND WESCHLER, H. Does alcohol consumption reduce human WOOD, M.D.; SHER, K.J.; AND MCGOWAN, A.K. Collegiate alcohol involvement and role capital accumulation? Evidence from the College Alcohol Study. Applied Economics attainment in early adulthood: Findings from a prospective high-risk study. Journal of 35(10):1227–1245, 2003. Studies on Alcohol 61(2):278–289, 2000. PMID: 10757139

WOLFSON, M.; CHAMPION, H.; MCCOY, T.P.; ET AL. Impact of a randomized campus/commu- World Health Organization (WHO). Alcohol and Injuries: Emergency Department Studies nity trial to prevent high-risk drinking among college students. Alcoholism: Clinical and in an International Perspective. Geneva: World Health Organization, 2009. Experimental Research 36(10):1767–1778, 2012. PMID: 22823091

WOOD, P.K.; SHER, K.J.; ERICKSON, D.J.; AND DEBORD, K.A. Predicting academic problems in ZAGROSEK, A.; MESSROGHLI, D.; SCHULZ, O.; ET AL. Effect of binge drinking on the heart as college from freshman alcohol involvement. Journal of Studies on Alcohol 58(2):200– assessed by cardiac magnetic resonance imaging. JAMA: Journal of the American 210, 1997. PMID: 9065898 Medical Association 304(12):1328–1330, 2010. PMID: 20858877

218 Alcohol Research: Current Reviews marizes recent evidence on associations between women’s Focus on: Women and the Costs alcohol consumption and their physical and mental health. of Alcohol use Drinking Practices and Patterns Among Women in Midlife Rates of drinking decline with age for both men and women Sharon C. Wilsnack, Ph.D.; Richard W. Wilsnack, Ph.D.; in the United States, and drinking remains less prevalent and Lori Wolfgang kantor, M.A. among women compared with men. In 2010, the proportion of people reporting at least one drink in the previous 30 days (i.e., current drinkers) decreased from 70 percent among although light-to-moderate drinking among women is 21- to 25-year-olds to 61.1 percent among 40- to 44-year- associated with reduced risks of some cardiovascular problems, olds and 51.6 percent among 60- to 64-year-olds (SAMHSA strokes, and weakening of bones, such levels of drinking also 2011). The same survey also found that approximately 57.4 are associated with increased risks of breast cancer and liver percent of males aged 12 or older were current drinkers, problems, and heavy drinking increases risks of hypertension compared with 46.5 percent of females of the same age and bone fractures and injuries. Women’s heavy-drinking range (SAMHSA 2011). patterns and alcohol use disorders are associated with Rates of binge drinking also are higher among men than increased likelihood of many psychiatric problems, including women (Centers for Disease Control and Prevention [CDC] depression, posttraumatic stress disorder, eating disorders, and 2012). One survey (National Institute on Alcohol Abuse and suicidality, as well as increased risks of intimate partner Alcoholism [NIAAA] 2012) reported that 28.8 percent of violence and sexual assault, although causality in the women and 43.1 percent of men reported binge drinking associations of drinking with psychiatric disorders and with (i.e., consuming within 2 hours four or more drinks for violence remains unclear. it is important for women to be aware women and five or more drinks for men) in the previous of the risks associated with alcohol use, especially because year. In a multinational study of 35 countries, Wilsnack and gaps between u.s. men’s and women’s drinking may have colleagues (2009) reported that, as expected, men consistently narrowed. however, analyses of health risks and benefits need drank more than women and were more likely to engage in improvement to avoid giving women oversimplified advice high-volume drinking and high-frequency drinking. Women about drinking. kEY WoRDS: Alcohol consumption; alcohol use, were more likely to be lifetime nondrinkers and to be former abuse, and dependence; alcohol use disorder; alcohol drinkers.2 The authors suggest that women may find it easier burden; drinking patterns; prevalence; alcohol burden; than men to quit drinking because (1) women generally are alcohol-related problems; alcohol-related injuries; women; lighter drinkers than men; (2) drinking is not as important pregnancy; cardiovascular disease; stroke; bone mass to women’s social roles as it is to men’s; and/or (3) women density; breast cancer; liver disease; psychiatric disorders; who stop drinking during pregnancy and early childrearing posttraumatic stress disorder (PtSD); depression; eating may not resume drinking later on. disorders; suicidal behavior; intimate partner violence; sexual Despite these findings, Grucza and colleagues (2008) reported assault significant increases between 1990–1991 and 2000–2001 in the lifetime prevalence of drinking for women aged 38–47 in the United States. There also was an increase in lifetime prevalence of alcohol dependence among women drinkers ven though the prevalence of alcohol use in the United aged 38-47. Similar increases were not found for male drinkers, States generally is lower among women compared with suggesting that the gender gap in alcohol use and dependence Emen (Substance Abuse and Mental Health Services is narrowing, at least in these age groups. Administration [SAMHSA] 2011), this gap has narrowed (Grucza et al. 2008). Furthermore, although women con- 2 Former drinkers reported drinking in the past but not in the last 12 months. sume alcohol at lower levels than men, their body composi- tion puts them at higher risk than men of developing some alcohol-related problems, both acutely (because of higher Sharon C. Wilsnack, Ph.D., and Richard W. Wilsnack, blood alcohol levels from a given amount of alcohol1) and chronically (from alcohol-related organ damage). This article Ph.D., are both professors in the Department of Clinical examines alcohol-use patterns (with particular attention to Neuroscience, University of North Dakota School of midlife) and how they differ for men and women and sum- Medicine & Health Sciences, Grand Forks, North Dakota.

Lori Wolfgang kantor, M.A., is a science editor at alcohol 1 because women’s bodies generally have less water than men’s bodies, alcohol becomes less Research: Current Reviews. diluted, and women therefore reach higher blood alcohol levels than men even if both are drinking the same amount.

Women and the Costs of Alcohol Use 219 Drinking During Pregnancy: Patterns and physical impairment, or mental problems). That is, women Predictors who averaged between one-third and one drink per day had about 20 percent higher odds than nondrinkers of good Women who become pregnant in their thirties and forties health at age 70 and older. Also, the women who drank may be more likely to drink during pregnancy than younger frequently during the week (5 to 7 days) had better odds of women. From 2001 to 2005, 17.7 percent of 35- to 44-year- good health at age 70 and older than the women who drank old women reported drinking during pregnancy, compared only once or twice a week (Sun et al. 2011). However, these with 8.6 percent of pregnant women aged 18–24 (Denny findings should be interpreted with caution because the et al. 2009). Among women in eight States who gave birth measures of alcohol consumption were quite limited. between 1997 and 2002, 30.3 percent reported drinking during pregnancy, and 8.3 percent reported binge drinking Effects of Women’s drinking on Cardiovascular Health (four or more drinks on one occasion). Whereas 22.5 percent of the women reported drinking during the first month of Many studies have found that light to moderate alcohol use pregnancy, drinking declined during pregnancy; only 7.9 is associated with lower risks of cardiovascular disease and percent of women reported drinking during the third trimester, mortality, but these studies often have not reported specifically and only 2.7 percent reported drinking during all trimesters. on women’s drinking. However, studies of coronary heart Drinking during pregnancy was more prevalent among disease risk in Denmark (Tolstrup et al. 2006) and England women over 30 (more than 30 percent drank) than among (Ward et al. 2011) found that the risks were lower in women younger women (Ethen at al. 2009). who consumed more alcohol. In the United States, pooled Understanding the predictors of drinking during pregnancy data from nine National Health Interview Surveys (1987– may help target prevention efforts. The eight-State study by 2000) showed that women drinking up to seven drinks per Ethen and colleagues (2009) found that both drinking and week had lower risks of cardiovascular mortality than life- binge drinking during pregnancy were predicted by prepreg- time abstainers (Mukamal et al. 2010). nancy binge drinking. Drinking and binge drinking during Light-to-moderate drinking also may be associated with pregnancy also were more prevalent among women who lower risks of sudden cardiac death (SCD). The study of were non-Hispanic whites, who smoked during pregnancy, nurses in the United States, which examined heart problems and whose pregnancy was unintended. A recent review of in 4-year periods after reported drinking or abstaining, 14 studies of drinking during pregnancy in nine countries found the lowest risk of SCD among women who averaged (Skagerstróm et al. 2011) found that drinking during pregnancy approximately one-half to one drink per day. Women who was associated with heavier drinking prior to pregnancy in drank more heavily (more than 30 g or two drinks per day) all seven studies that measured this; smaller numbers of studies had SCD risks similar to risks of abstainers, but the number consistently found that drinking during pregnancy was associated of SCD cases among women who consumed more than 30 g with higher income/social class and with histories of abuse or per day was limited (Chiuve et al. 2010). As noted earlier, how- exposure to violence and histories of drinking problems. ever, these findings are based on limited measures of drinking. In contrast to studies finding beneficial effects, a meta- analysis of six studies (Samokhvalov et al. 2010) found that Physical Health Effects of Women’s Drinking women’s risks of atrial fibrillation (AF) increased steadily with increasing alcohol consumption. Whereas women who Light to moderate alcohol use has been found to generally averaged up to two drinks a day did not have significantly be beneficial for many health outcomes and is associated higher risks than abstainers, women who consumed more with decreased mortality. Heavier use, however, is associated than two to three drinks daily had a 17 percent increased with poorer health and increased mortality. One meta-analysis risk of AF, and women who consumed more than four of 34 studies in 13 countries found that, compared with drinks daily had twice the risk of AF. abstaining, drinking less than two drinks per day among Women’s risk of hypertension also may increase steadily women and drinking less than four drinks per day among as alcohol consumption increases. A meta-analysis of eight men was associated with significantly reduced total mortality, studies indicated that the risk was reduced somewhat among but higher levels of alcohol use were associated with increased women drinking lightly (averaging less than a drink a day), mortality . These findings should (Di Castelnuovo et al. 2006) but the risk then rose steadily with higher levels of consump- not encourage people to start drinking alcohol for its health tion. Compared with abstainers, women who averaged benefits, because of the significant health problems associated roughly four drinks a day had nearly twice the risk of hyper- with heavier use, as described below. Another study used data from a large survey of middle- tension, and women averaging roughly eight drinks a day aged (median age 58) female nurses in the United States and had nearly three times the risk (Taylor et al. 2009). assessed the health of participants who lived to age 70 and Effects of Women’s drinking on Stroke Risk older. The study found that light to moderate alcohol con- sumption at midlife was associated with modestly increased The risk of stroke is lower among women who are light-to- odds of good health at age 70 or older (no chronic illnesses, moderate drinkers. The U.S. nurses’ study found lower risk of

220 Alcohol Research: Current Reviews strokes among women who were recent light drinkers, aver- on data from 184,418 postmenopausal women aged 50 aging approximately one drink a day (Jimenez et al. 2012). to 71, reported similar findings (Lew et al. 2009). After 7 Among 45,449 Swedish women aged 30 to 50 who were years of follow-up, the researchers found that risks of breast followed up approximately 11 years later, risks of ischemic cancer increased steadily the more women drank. Risks stroke were significantly lower among women averaging less were highest for estrogen- and progesterone-receptor–positive than one drink a day (compared with abstainers). The num- tumors, with risks of these tumors 46 percent higher for bers of women with hemorrhagic strokes and/or strokes after women drinking more than 35 grams of alcohol (more than drinking more heavily were too small for reliable evaluation two drinks) a day. However, when Suzuki and colleagues (Lu et al. 2008). Meta-analyses of five to nine other studies (2010) followed up 50,757 Japanese women (aged 40 to 69) found that women’s light-to-moderate drinking was protec- over 13 years, they found that breast cancer risk increased 6 tive against both ischemic and hemorrhagic strokes (with percent with every additional 10 grams per day of alcohol lowest risks in women averaging about one drink a day), but risks of morbidity and mortality from both types of strokes consumption, but the observed association was not modified increased rapidly as women’s consumption rose above three by menopausal status or use of exogenous estrogens. These to four drinks a day (Patra et al. 2010). findings suggest that breast cancer risks associated with alcohol consumption involve more than just estrogen levels. Effects of Women’s drinking on Liver disease Effects of Women’s drinking on Bone Health Women apparently are more vulnerable than men to liver cirrhosis and other liver injury from alcohol use, possibly Higher bone-mineral density (BMD) is associated with resis- because of estrogens, although the mechanisms are as yet tance to fracture. A recent review of research relevant to 40- unclear (Eagon 2010). A meta-analysis of 12 studies found to 60-year-old women concluded that there was fair evidence that women’s risks of morbidity and mortality from liver that moderate drinking did no harm to BMD (Waugh et al. cirrhosis increased steadily with higher levels of alcohol 2009). In fact, a number of studies have found that light to consumption, with no protective effect of light to moderate moderate drinking is associated with increased BMD, at drinking, and the risks increased more rapidly for women least among postmenopausal women (Maurel et al. 2012). than for men (Rehm et al. 2010). These risks may be increased For example, Tucker and colleagues (2009) found that, in by other personal characteristics and by drinking patterns. women from the Framingham Offspring cohort, hip and In a very large sample of women in the United Kingdom, spine BMD were 5.0 to 8.3 percent greater in postmenopausal followed up for an average of 6.2 years, risks of cirrhosis women who consumed more than two drinks per day than among women averaging two or more drinks a day increased 2 in nondrinkers. A study of 2,043 postmenopausal women greatly if their body mass indexes were greater than 28 kg/m in the United States found that BMD was 3.8 percent higher (Liu et al. 2010). In a large study of women in New York in women who had more than 29 drinking occasions per State, levels of -glutamyl-transferase (GGT), a liver enzyme γ month than those who abstained, although this finding only that increases in all forms of liver disease (Niemelä and Alatalo 2010), were highest not only in women who averaged more was marginally significant (because of small numbers of daily than a drink a day but also in women who did their drink- drinkers) (Wosje and Kalkwarf 2007). Finally, a study in ing only on weekends and without food (Stranges et al. 2004). Scotland of 3,218 women aged 50 to 62 found significant increases in BMD in the femoral neck and lumbar spine in Effects of Women’s drinking on Breast Cancer Risk women who averaged more than one drink a day, compared with lifetime abstainers (McLernon et al. 2012). However, in Even moderate alcohol consumption increases breast cancer general these studies were unable to evaluate effects of heavy risk, and the risk rises as drinking increases. A multinational drinking, and the processes by which alcohol affects BMD meta-analysis of 98 studies found that the risk of breast can- remain uncertain but may involve effects of increased levels cer increased an average of 10 percent for every increase of of estrogen and calcitonin (Maurel et al. 2012). 10 grams per day in alcohol consumption (Key et al. 2006). In contrast, the prevailing wisdom is that heavy drinking A 10-year follow-up study of more than 38,000 U.S. women aged 45 and older found a significant trend of increased risk (averaging multiple drinks per day) increases women’s risks of invasive breast cancer associated with increased alcohol of fractures, such as from falls (Epstein et al. 2007). In a consumption at baseline, with the greatest risk among women combined study of 11,032 women in Canada, Australia, and averaging at least 30 grams of alcohol per day (Zhang et al. the Netherlands, the risks of hip fractures and osteoporotic 2007). The risks from alcohol consumption were clearest for fractures were higher in women averaging two or more drinks estrogen- and progesterone-receptor–positive tumors and for a day than in women averaging up to one drink a day (Kanis women currently taking postmenopausal hormones, consis- et al. 2005). In Sweden, a study of 10,902 middle-aged tent with the hypothesis that part of alcohol’s effect on breast women showed that low-energy fractures were more likely in cancer is to increase estrogen exposure (Garcia-Closas et al. women who had higher levels of GGT, which is associated 2002; Onland-Moret et al. 2005). Another U.S. study, based with chronic heavy drinking (Holmberg et al. 2006).

Women and the Costs of Alcohol Use 221 Women’s Drinking and Psychiatric Disorders depression. Research clearly has established that depressive disorders and symptoms are more likely among people with AUDs (e.g., Grant et al. 2004), but studies have not Alcohol use disorders always examined this connection specifically among women. In addition to physical health risks associated with alcohol However, a large U.S. twin study found that diagnoses of use, women’s risks of mental health problems also are related major depression and alcohol dependence were correlated to their drinking. It is clear that women’s heavy and binge among women (Prescott et al. 2000), and data from the drinking is associated with alcohol use disorders (AUDs). large National Epidemiologic Study on Alcohol and Related For example, U.S. data show that among women aged 50 Conditions (NESARC) showed that women with major or older, those who engage in binge drinking (four or more depressive disorder were more likely to report multiple drinks on a drinking occasion) have more than three times criteria for alcohol abuse and dependence (Lynskey and greater risks of alcohol abuse, and more than five times Agrawal 2008). Research also has repeatedly found associations greater risks of alcohol dependence, than women who drink of women’s depression with binge drinking. For example, but do not engage in binge drinking (Chou et al. 2011). in a major Canadian survey, women’s binge drinking (five However, there has otherwise been limited attention to or more, or eight or more, drinks per day) was associated gender-specific ways in which women’s drinking may be with measures of recent and longer-term depression (Graham related to AUDs. One exception is that women, like men, et al. 2007), and data from the large U.S. Behavioral Risk are at greater risk of AUDs if they begin drinking at early Factor Surveillance System surveys showed that lifetime ages. A large study in Missouri has found elevated risks of depression was significantly more likely in women who AUDs in women who began drinking before age 18 (Jenkins engaged in binge drinking (four or more drinks in a day) et al. 2011), confirming findings from U.S. national surveys (Strine et al. 2008). (Dawson et al. 2008). A second exception is that it has long been thought that development of AUDs is “telescoped” in PtSd. AUDs often have been associated with symptoms women compared with men, occurring in a shorter period of or diagnoses of PTSD. For example, in young adults time after women begin to drink (Greenfield 2002). However, followed up from the U.S. National Survey of Adolescents, this pattern was identified in women in treatment for AUDs, women with PTSD in the past 6 months were more than and U.S. survey data now indicate that telescoping does not twice as likely as other women to meet criteria for a Diagnostic occur in women drinkers in the general population (Keyes et and Statistical Manual of Mental Disorders, 4th Edition al. 2010) but may be related to the experiences that bring diagnosis of alcohol abuse (Danielson et al. 2009). Among women to treatment. women from the large Missouri Adolescent Female Twin Study, PTSD was associated with a greater likelihood of Psychiatric disorders other than Auds AUDs (Sartor et al. 2010). In surveys of three Mexican General-population studies often have found links between cities, lifetime PTSD was more prevalent in women who women’s drinking and psychiatric disorders, but the time misused alcohol (with at least one indicator of alcohol abuse order and causes of these linkages are often unclear. For or dependence) (Slone et al. 2006). In addition, in the example, a German survey found that women with alcohol large California Women’s Health Survey, having symptoms abuse or dependence, or women who drank an average of of PTSD doubled the odds that women engaged in binge at least 20 to 30 grams of alcohol per day, were more likely drinking (Timko et al. 2008). However, most of these than other women to have a variety of psychiatric disorders studies have not found any effects of PTSD beyond the (affective, anxiety, or somatoform), and the connections effects of the traumatic experiences that led to PTSD, a between drinking and disorders were stronger for women pattern also reported in other recent studies of women than for men (Bott et al. 2005). A Danish survey found that who have experienced sexual assaults (Najdowski and Ullman any psychiatric disorders were more likely in women averag- 2009; Testa et al. 2007). Therefore, PTSD may be an ing more than three drinks a day, and anxiety disorders were indicator of experiences distressful enough to lead women specifically more likely among women averaging more than to drink to excess, but PTSD itself may not necessarily be two drinks a day, compared with nondrinkers (Flensborg- a cause of such drinking. Madsen et al. 2011). In addition, U.S. data on women aged 50 and older showed higher risks of both panic disorder and Alcohol and Eating disorders. Research often has found posttraumatic stress disorder (PTSD) in women who engaged that eating disorders in women are associated with problem in any binge drinking, compared with non–binge drinkers drinking. The strongest recent evidence is in a meta-analysis (Chou et al. 2011). Unlike the preceding studies, which of 41 studies, mainly in the U.S. and Canada, in which linked drinking patterns to increased risks of general psychi- women’s eating disorders consistently were associated with atric comorbidity, most studies of women’s alcohol use and AUDs (Gadalla and Piran 2007). The meta-analysis included psychiatric disorders have focused on comorbidity of specific a very large Canadian general-population survey in which disorders with AUDs and risky drinking patterns. These risks of eating disorders also were associated with heavier more specific linkages are discussed in the sections that follow. weekly drinking among women ages 15 to 44 (Piran and

222 Alcohol Research: Current Reviews Gadalla 2007). Hypotheses to explain observed links between related to population drinking patterns and less attention to women’s eating disorders and drinking typically have focused how drinking is related to the risks of injury in individual on possible common antecedents (distress, personality women. However, studies have reported two consistent findings characteristics, and genetic factors) rather than on ways about how individual drinking patterns are linked to injuries. that eating disorders might cause or be caused by drinking First, risks of injury increase among women who have (Conason and Sher 2006). consumed alcohol in the 6 hours before being injured; The meta-analysis by Gadalla and Piran (2007) showed women’s injury risks associated with drinking occur relatively that problem drinking was associated more specifically with rapidly. This conclusion has been confirmed by a combined bulimic behavior than with anorexia nervosa. The associations analysis of 28 hospital emergency-department studies in 16 also were stronger among women in community or student countries (Borges et al. 2006). Additional confirmation has samples but were weaker or absent when women in treat- come from a large emergency-department survey in Sydney, ment for eating disorders were compared with women in the Australia, where the risk was greatest in women who had general population. A multisite European study comparing consumed more than 90 grams of alcohol in the 6 hours individuals (mostly women) in treatment versus healthy before being injured (Williams et al. 2011). individuals in the general population also failed to find that The other consistent finding is that risks of injury are those in eating disorders treatment drank more heavily greatest among women whose drinking patterns are particularly (Krug et al. 2008). It is possible that such negative findings heavy or hazardous. A study of women outpatients at a could result because many women receiving treatment or Veterans Administration hospital found that the likelihood seeking treatment for eating disorders curtail their drinking. of multiple recent injuries was nearly doubled in the heaviest versus the lightest drinkers (Chavez et al. 2012). A study of Alcohol and Suicidal Behavior. Although research often women with high-risk drinking patterns at five U.S. colleges has reported on factors affecting rates of suicide among found that their risks of recent injury were directly related women, only rarely have studies been able to show how to their number of days of drinking five or more drinks individual women’s drinking patterns are related to suicidal (Mundt et al. 2009). In addition, large surveys of women behavior. An exception was a 20-year follow-up of a large aged 45 to 69 in three Eastern European countries found sample of Swedish women hospitalized because of suicidal that the percentage of women with injuries was higher in behavior; those women diagnosed also with alcohol abuse women with high scores on the CAGE3 screening instru- or dependence had a higher risk of later committing suicide ment for problem drinking (Vikhireva et al. 2010). (Tidemalm et al. 2008). Most general-population surveys of individual women have shown that suicidal ideation intimate Partner violence (thinking about committing suicide) was associated with heavier, more frequent, or more hazardous drinking. In Associations between alcohol use and intimate partner vio- the United States, for example, women’s suicidal ideation lence (IPV) have been well documented in research in North was associated with hazardous drinking patterns in a America. Male-to-female IPV perpetration consistently has longitudinal study of women aged 26 to 54 (Wilsnack et been linked to heavy and problem drinking by men (Caetano al. 2004) and was associated with alcohol dependence in the et al. 2000; Thompson and Kingree 2006). The large-scale large National Longitudinal Alcohol Epidemiologic Survey NESARC survey found that past-year IPV victimization was (Grant and Hasin 1999). A large study of active-duty U.S. more likely in women who have symptoms of alcohol abuse Air Force personnel also found that women’s suicidal or dependence (La Flair et al. 2012), and meta-analysis of six ideation was associated with higher levels of alcohol surveys of adolescents and young adults showed that women’s problems, but only among women who were not mothers frequency and/or quantity of drinking was positively related (Langhinrichsen-Rohling et al. 2011). In Seoul, Korea, to their perpetration of IPV (Rothman et al. 2012). Further- women aged 18 to 64 showed a strong association of more, a comparative study of alcohol consumption and IPV suicidal ideation with drinking nearly daily (Park et al. in Canada, the United States, and eight countries in Latin 2010). Finally, a French survey of women aged 18 to 30 America found that in all 10 countries, rates of physical part- found that suicidal ideation was more common in heavier ner aggression were higher among drinkers than nondrinkers drinkers, although the relationship no longer was statistically (men and women); and among drinkers, rates were higher significant after controlling for effects of depression and among persons who reported drinking larger amounts per other adverse experiences (Legleye et al. 2010). occasion. Women reported being victims of more severe aggression than men, and men were more likely than women to be drinking at the time of an incident of physical aggres- Alcohol-Related injuries 3 the CagE is a screening instrument (Ewing 1984) consisting of the following four questions: have Similar to research on women’s suicidality, research on women’s you ever felt you should cut down on your drinking? have people annoyed you by criticizing your drinking? have you ever felt bad or guilty about your drinking? Eye opener: have you ever had a drink alcohol-related injuries has given more attention to gender first thing in the morning to steady your nerves or to get rid of a hangover? two positive responses differences in injury rates and how women’s injury rates are are considered a positive test and indicate further assessment is warranted.

Women and the Costs of Alcohol Use 223 sion (Graham et al. 2008). Other multinational studies have women who have established patterns of binge drinking or shown that odds of IPV were greater where one or both problem drinking. For example, in a large national survey of partners had alcohol problems (Abramsky et al. 2011) and college women in 1999, women with alcohol problems were that aggression severity was significantly higher if one or more likely to report experiencing unwanted sexual advances both partners had been drinking when the aggression (Pino and Johnson-Johns 2009). At a large New York State occurred (Graham et al. 2011). However, in all this research, university, women who increased their drinking during their it is unclear to what extent drinking is a cause or an effect of first year in college (and who averaged more than four drinks IPV, or both. per drinking occasion, with frequent such occasions) had higher odds of sexual victimization (Parks et al. 2008). Second, women are more likely to experience rape or Alcohol and Sexual Assault other severe sexual assault if they become intoxicated, at the time of the assault or as a typical drinking pattern. A large It has been known for some time that women’s drinking is U.S. survey of college women found that the percentage positively associated with their risks of sexual assault, but who had been raped was high in women with any recent how and why this association occurs remains unsettled experience of binge drinking (four or more drinks per occa- (Abbey et al. 2004). Part of the association results because sion) and that more than two-thirds of the women who had women often drink with men who drink, and the men’s been raped reported being intoxicated at the time (Mohler- intoxication makes them more likely to be sexually aggressive Kuo et al. 2004). A study of more than 300 young women toward women (Abbey 2011). Other links between women’s who had been sexually assaulted since age 14 found that the drinking and sexual assaults are harder to interpret because odds of sexual penetration were greater only among women investigators often lack time-ordered data, they differ in the reporting high levels of intoxication (Testa et al. 2004). An types of sexual activity they evaluate (ranging from rape to earlier national survey of college women who had experi- much broader categories of unwanted sexual advances), and enced sexual victimization found that the severity of the most of their studies are limited to college women (as a high- assault was predicted in part by the women’s frequency of risk group). intoxication (Ullman et al. 1999). Nevertheless, certain patterns have become clear in recent Findings like these have led some investigators to conclude years. First, risks of sexual assault are most clearly higher in that one reason why drinking may increase women’s risks of

30

25 No HED HED 20

15 Percent 10 *

* * 5 *

0

Costa Peru Peru Argentina Belize Brazil Canada Rica Mexico Nicaragua (Lima) (Ayacucho) USA Uruguay 10 Countries of the Americas

Figure Female intimate-partner violence victimization by women’s past-12-month heavy episodic drinking (hED) (10 countries of the americas).

notE: * p < .05 for logistic regression, controlling for age. souRCE: graham, K.; bernards, s.; munné, m.; and Wilsnack, s.C.; Eds. Unhappy Hours: Alcohol and Partner Aggression in the Americas. Washington, DC: Pan american health organization, 2008.

224 Alcohol Research: Current Reviews sexual assault is that highly intoxicated women may be inca- outcomes (such as psychiatric disorders and intimate partner pacitated, unable to resist unwanted sexual advances. A violence). Finally, research findings often are presented as if national survey of college women found that a past-year history they applied similarly to all women drinkers, without dis- of binge drinking (five or more drinks at a sitting) was cussing how other conditions and contexts (such as a specifically associated with experiencing incapacitated rape drinker’s other health conditions) might modify how alcohol (McCauley et al. 2009). A study of first-year college students affects health. (One exception, for example, is the research found that reported maximum consumption per occasion by Liu et al. [2010] showing that risks of cirrhosis from rela- during the fall semester was strongly associated with experi- tively heavy drinking are greater in women with high body encing incapacitated rape (Testa and Hoffman 2012). A mass indices.) Therefore, what we should strive for is infor- number of related studies reviewed by Testa and Livingston mation about health effects of women’s drinking that shows (2009) led to the conclusions that in many rapes, especially not only the effect sizes, but also when and where and of college students, women are incapacitated by some form among which women the effects are greatest. of substance use, and that many rapes associated with alcohol Keeping those limitations in mind, the findings summarized use involve incapacitation. here may offer some guidelines for women making personal decisions about drinking in midlife. Light-to-moderate Conclusions drinking is associated to some extent with reduced risks of some cardiovascular problems, strokes, and weakening of Because alcohol consumption has become a more normal bones. On the other hand, even low levels of alcohol con- activity for women, it is important for women to have science- sumption may cause women some increase in risks of breast based information to help them decide whether and when cancer and liver problems, and heavy drinking also increases to drink, and in what amounts, based on potential risks or risks of hypertension and bone fractures and injuries. Women’s benefits of drinking. Such past and current information has heavy drinking patterns and AUDs are associated with had some important limitations. Some of these limitations increased likelihood of many psychiatric problems, including have been addressed in recent decades. In most recent studies depression, PTSD, eating disorders, and suicidality. Women’s (e.g., Mukamal et al. 2010; Patra et al. 2010), apparent heavy drinking and AUDs also are associated with increased health benefits of moderate drinking now are based on com- risks of intimate partner violence and sexual assault, although parisons with lifetime abstainers, excluding potentially sicker causality in the associations of drinking with psychiatric dis- ex-drinkers who were part of some earlier comparisons. Also, orders and with violence remains unclear. On balance, the long-term studies of alcohol consumption in women now evidence summarized here suggests that, for those women are likely to include more detailed measures of baseline who choose to drink, moderation in consumption is the drinking (Moore et al. 2005; Wilsnack et al. 2006) than earlier studies used (Stampfer et al. 1988). However, some safest or least costly strategy to adopt toward alcohol. ■ research findings are still presented in terms of rates of health outcomes in whole groups of women (such as for injuries and suicidality; Landberg 2010; Ramstedt 2005), which can be Financial Disclosure misleading if these results are used to draw conclusions The authors declare that they have no competing financial about the effects of drinking on individuals. Finally, research interests. on long-term health effects of women’s drinking can measure only some of the lifestyle characteristics (such as eating pat- terns and exercise) that may be associated with how women References drink and that may account for some of the apparent effects of drinking (Mukamal et al. 2010; Rimm & Moats 2007). abbEy, a. alcohol’s role in sexual violence perpetration: theoretical explanations, exist- A major current limitation of information about alcohol ing evidence and future directions. Drug and Alcohol Review 30 (5):481–489, 2011. effects is that such effects often are reported, in scientific PmiD: 21896070 papers but particularly in the news media, as simple associa- abbEy, a.; zaWaCKi, t.; buCK, P.o.; Et al. sexual assault and alcohol consumption: What tions (this drinking pattern is associated with that health do we know about their relationship and what types of research are still needed? outcome). Less is said about how large the effects are (not Aggression and Violent Behavior 9(3):271–303, 2004. very large for some cardiovascular benefits of moderate abRamsKy, t.; Watts, C.h.; gaRCia-moREno, C.; Et al. What factors are associated with drinking), and adverse effects often are implied to increase in recent intimate partner violence? Findings from the Who multi-country study on women’s a linear way with each unit increase in drinking. There is too health and domestic violence. BMC Public Health 11:109, 2011. PmiD: 21324186 little attention paid to how the effects of drinking may not boRgEs, g.; ChERPitEl, C.J.; oRozCo, R.;. Et al. acute alcohol use and the risk of non-fatal be linear (with the exception of research on cardiovascular injury in sixteen countries. Addiction 101(7):993–1002, 2006. PmiD: 16771891 benefits versus hazards at different levels of drinking). There bott, K.; mEyER, C.; RumPF, h.-J.; Et al.. Psychiatric disorders among at-risk consumers of also is too little attention paid to how drinking may be both alcohol in the general population. Journal of Studies on Alcohol 66(2):246–253, 2005. a cause and an effect of some adverse health and behavioral PmiD: 15957676

Women and the Costs of Alcohol Use 225 CaEtano, R.; CunRaDi, C.b.; ClaRK, C.l.; anD sChaFER, J. intimate partner violence and gRant, b.F., anD hasin, D.s. suicidal ideation among the united states drinking popula- drinking patterns among White, black, and hispanic couples in the u.s. Journal of tion: Results from the national longitudinal alcohol Epidemiologic survey. Journal of Substance Abuse 11(2):123–138, 2000. PmiD: 10989773 Studies on Alcohol 60(3):422–429, 1999. PmiD: 10371272

Centers for Disease Control and Prevention (CDC). Alcohol and Public Health Fact Sheet. gRant, b.F.; stinson, F.s.; DaWson, D.a.; Et al. Prevalence and co-occurrence of sub- atlanta, ga: CDC. available at: http://www.cdc.gov/alcohol/ fact-sheets/binge-drinking.htm. stance use disorders and independent mood and anxiety disorders: Results from the accessed august 31, 2012. national Epidemiologic survey on alcohol and Related Conditions. Archives of General Psychiatry 61(8):807–816, 2004. PmiD: 15289279 ChaVEz, l.J.; Williams, E.C.; laPham, g.; anD bRaDlEy, K.a. association between alcohol

screening scores and alcohol-related risks among female Veterans affairs patients. gREEnFiElD, s.F. Women and alcohol disorders. Harvard Review of Psychiatry Journal of Studies on Alcohol and Drugs 73(3):391–400, 2012. PmiD: 22456244 10(2):76–85, 2002. PmiD: 11897748

ChiuVE, s.E.; Rimm, E.b.; muKamal, K.J. Et al. light-to-moderate alcohol consumption and gRuCza, R.a.; buCholz, K.K.; RiCE, J.P.; anD biERut, l.J. secular trends in the lifetime preva- risk of sudden cardiac death in women. Heart Rhythm 7(10):1374–1380, 2010. PmiD: lence of alcohol dependence in the united states: a re-evaluation. Alcoholism: Clinical 20638933 and Experimental Research 32(5):763–770, 2008. PmiD: 18336633 Chou, K.-l.; liang, K.; anD maCKEnziE, C.s. binge drinking and axis i psychiatric disorders holmbERg, a.h.; JohnEll, o.; nilsson, P.m.; Et al. Risk factors for fragility fracture in middle in community-dwelling middle-aged and older adults: Results from the national age: a prospective population-based study of 33,000 men and women. Osteoporosis Epidemiologic survey on alcohol and Related Conditions (nEsaRC). Journal of Clinical International 17(7):1065–1077, 2006. PmiD: 16758143 Psychiatry 72(5):640–647, 2011. PmiD: 21294995 JEnKins, m.b.; agRaWal, a.; lynsKEy, m.t.; Et al. Correlates of alcohol abuse/dependence Conason, a.h., anD shER, l. alcohol use in adolescents with eating disorders. International in early-onset alcohol-using women. American Journal on Addictions 20(5):429–434, Journal of Adolescent Medicine and Health 18(1):31–36, 2006. PmiD: 16639856 2011. PmiD: 21838841 DaniElson, C.K.; amstaDtER, a.b.; DangElmaiER, R.E.; Et al. trauma-related risk factors for substance abuse among male versus female young adults. Addictive Behaviors JimEnEz, m.; ChiuVE, s.E.; glynn, R.J.; Et al. alcohol consumption and risk of stroke in 34(4):395–399, 2009. PmiD: 19110381 women. Stroke 43(4):939–945, 2012. PmiD: 22403048

DaWson, D.a.; golDstEin, R.b.; Chou, s.P.; Et al. age at first drink and the first incidence Kanis, J.a.; Johansson, h.; JohnEll, o.; Et al. alcohol intake as a risk factor for fracture. of adult-onset Dsm-iV alcohol use disorders. Alcoholism: Clinical and Experimental Osteoporosis International 16(7):737–742, 2005. PmiD: 15455194 Research 32(12):2149–2160, 2008. PmiD: 18828796 KERR, W.C.; subbaRaman, m.; anD yE, y. Per capita alcohol consumption and suicide mor- DEnny, C.h; tsai, J.; FloyD, R.l.; anD gREEn, P.P. alcohol use among pregnant and non- tality in a panel of us states from 1950 to 2002. Drug and Alcohol Review pregnant women of childbearing age: united states, 1991–2005. Morbidity and 30(5):473–480, 2011. PmiD: 21896069 Mortality Weekly Report 58(19):529–532, 2009. PmiD: 19478721 KEy, J.; hoDgson, s.; omaR, R.z.; Et al. meta-analysis of studies of alcohol and breast Di CastElnuoVo, a.; Costanzo, s.; bagnaRDi, V.; Et al. alcohol dosing and total mortality in cancer with consideration of the methodological issues. Cancer Causes & Control men and women: an updated meta-analysis of 34 prospective studies. Archives of 17(6):759–770, 2006. PmiD: 16783604 Internal Medicine 166(22):2437–2445, 2006. PmiD: 17159008 KEyEs, K.m.; maRtins, s.s.; blanCo, C.; anD hasin, D.s. telescoping and gender differences Eagon, P.K. alcoholic liver injury: influence of gender and hormones. World Journal of in alcohol dependence: new evidence from two national surveys. American Journal of Gastroenterology 16(11):1377–1384, 2010. PmiD: 20238405 Psychiatry 167(8):969–976, 2010. PmiD: 20439391

EPstEin, E.E.; FisChER-ElbER, K.; anD al-otaiba, z. Women, aging, and alcohol use disor- KRug, i.; tREasuRE, J.; anDERluh, m.; Et al. Present and lifetime comorbidity of tobacco, ders. Journal of Women & Aging 19(1–2):31–48, 2007. PmiD: 17588878 alcohol and drug use in eating disorders: a European multicenter study. Drug and Alcohol Dependence 97(1–2):169–179, 2008. PmiD: 18571341 EthEn, m.K.; RamaDhani, t.a.; sChEuRlE, a.E.; Et al. alcohol consumption by women before and during pregnancy. Maternal and Child Health Journal 13(2):274–285, 2009. la FlaiR, l.n.; bRaDshaW, C.P., stoRR, C.l.; Et al. intimate partner violence and patterns of PmiD: 18317893 alcohol abuse and dependence criteria among women: a latent class analysis. Journal

EWing, J. a. Detecting alcoholism: the CagE questionnaire. JAMA, 252(14):1905– of Studies on Alcohol and Drugs 73(3):351–360, 2012. PmiD: 22456240 1907, 1984. PmiD: 6471323 lanDbERg, J. Population drinking and fatal injuries in Eastern Europe: a time-series anal- FlEnsboRg-maDsEn, t.; bECKER, u.; gRønbaEK, m.; Et al. alcohol consumption and later risk ysis of six countries. European Addiction Research 16(1):43–52, 2010. PmiD: 20009445 of hospitalization with psychiatric disorders: Prospective cohort study. Psychiatry langhinRiChsEn-Rohling, J.; snaRR, J.D.; slEP, a.m.; Et al. Risk for suicidal ideation in the Research 187(1–2):214–219, 2011. PmiD: 21146876 u.s. air Force: an ecological perspective. Journal of Consulting and Clinical gaDalla, t., anD PiRan, n. Co-occurrence of eating disorders and alcohol use disorders in Psychology 79(5):600–612, 2011. PmiD: 21787046 women: a meta analysis. Archives of Women’s Mental Health 10(4):133–140, 2007. lEglEyE, s.; bECK, F.; PEREtti-WatEl, P.; Et al. suicidal ideation among young French PmiD: 17533558 adults: association with occupation, family, sexual activity, personal background and gaRCia-Closas, m.; hERbstman, J.; sChiFFman, m.; Et al. Relationship between serum hor- drug use. Journal of Affective Disorders 123(1–3):108–115, 2010. PmiD: 19892406 mone concentrations, reproductive history, alcohol consumption and genetic polymor- lEW, J.Q.; FREEDman, n.D.; lEitzmann, m.F.; Et al. alcohol and risk of breast cancer by his- phisms in pre-menopausal women. International Journal of Cancer 102(2):172–178, 2002. PmiD: 12385014 tologic type and hormone receptor status in postmenopausal women: the nih-aaRP Diet and health study. American Journal of Epidemiology 170(3):308–317, 2009. gRaham, K.; bERnaRDs, s.; WilsnaCK, s.C.; anD gmEl, g. alcohol may not cause partner vio- PmiD: 19541857 lence but it seems to make it worse: a cross national comparison of the relationship between alcohol and severity of partner violence. Journal of Interpersonal Violence liu, b.; balKWill, a.; REEVEs, g.; Et al. body mass index and risk of liver cirrhosis in middle 26(8):1503–1523, 2011. PmiD: 20522883 aged uK women: Prospective study. BMJ 340:c912, 2010. PmiD: 20223875

gRaham, K.; bERnaRDs, s.; munnÉ, m.; anD WilsnaCK, s.C. Unhappy Hours: Alcohol and lu, m.; yE, W.; aDami, h.o.; anD WEiDERPass, E. stroke incidence in women under 60 years Partner Aggression in the Americas. Washington, DC: Pan american health of age related to alcohol intake and smoking habit. Cerebrovascular Diseases organization, 2008. 25(6):517–525, 2008. PmiD: 18480604

gRaham, K.; massaK, a.; DEmERs, a.; anD REhm, J. Does the association between alcohol lynsKEy, m.t, anD agRaWal, a. Differential criterion functioning of alcohol use symptoma- consumption and depression depend on how they are measured? Alcoholism: Clinical tology in major depressive disorder? Psychological Medicine 38(3):441–449, 2008. and Experimental Research 31(1):78–88, 2007. PmiD: 17207105 PmiD: 17935644

226 Alcohol Research: Current Reviews mauREl, D.b.; boissEau, n.; bEnhamou, C.l.; anD JaFFRE, C. alcohol and bone: Review of Rothman, E.F.; mCnaughton REyEs, l.; Johnson, R.m.; anD laVallEy, m. Does the alcohol dose effects and mechanisms. Osteoporosis International, 23(1):1–16, 2012. PmiD: make them do it? Dating violence perpetration and drinking among youth. 21927919 Epidemiologic Reviews 34(1):103–119, 2012. PmiD: 22128086

mCCaulEy, J.; RuggiERo, K.J.; REsniCK, h.s.; Et al. Forcible, drug-facilitated, and incapaci- samoKhValoV, a.V.; iRVing, h.m.; anD REhm, J. alcohol consumption as a risk factor for atri- tated rape in relation to substance use problems: Results from a national sample of al fibrillation: a systematic review and meta-analysis. European Journal of college women. Addictive Behaviors 34(5):458–462, 2009. PmiD: 19162407 Cardiovascular Prevention and Rehabilitation 17(6):706–712, 2010. PmiD: 21461366 mClERnon, D.J.; PoWEll, J.J.; JugDaohsingh, R.; anD maCDonalD, h.m. Do lifestyle choices saRtoR, C.E.; mCCutChEon, V.V.; PommER, n.E.; Et al. Posttraumatic stress disorder and explain the effect of alcohol on bone mineral density in women around menopause? alcohol dependence in young women. Journal of Studies on Alcohol and Drugs American Journal of Clinical Nutrition 95(5):1261–1269, 2012. PmiD: 22456655 71(6):810–818, 2010. PmiD: 20946737 mohlER-Kuo, m.; DoWDall, g.W.; Koss, m.P.; anD WEChslER, h. Correlates of rape while sKagERstRóm, J.; Chang, g.; anD nilsEn, P. Predictors of drinking during pregnancy: a sys- intoxicated in a national sample of college women. Journal of Studies on Alcohol tematic review. Journal of Women’s Health 20(6):901–913, 2011. PmiD: 21671775 65(1):37–45, 2004. PmiD: 15000502 slonE, l.b.; noRRis, F.h.; gutiÉRREz RoDRiguEz, F.; Et al. alcohol use and misuse in urban mooRE, a.a.; goulD, R.; REubEn, D.b.; Et al. longitudinal patterns and predictors of mexican men and women: an epidemiologic perspective. Drug and Alcohol alcohol consumption in the united states. American Journal of Public Health 95(3): Dependence 85(2):163–170, 2006. PmiD: 16735099 458–464, 2005. PmiD: 15727977 stamPFER, m.J.; ColDitz, g.a.; WillEtt, W.C.; Et al. a prospective study of moderate alco- muKamal, K.J.; ChEn, C.m.; Rao, s.R.; anD bREsloW, R.a. alcohol consumption and cardio- hol consumption and the risk of coronary disease and stroke in women. New England vascular mortality among u.s. adults, 1987 to 2002. Journal of the American College Journal of Medicine 319(5):267–273, 1988. PmiD: 3393181 of Cardiology 55(13):1328–1335, 2010. PmiD: 20338493 stRangEs, s.; FREuDEnhEim, J.l.; muti, P.; Et al. Differential effects of alcohol drinking pat- munDt, m.P.; zaKlEtsKaia, l.i.; anD FlEming, m.F. Extreme college drinking and alcohol- related tern on liver enzymes in men and women. Alcoholism: Clinical and Experimental injury risk. Alcoholism: Clinical and Experimental Research 33(9):1532–1538, 2009. Research 28(6):949–956, 2004. PmiD: 15201638 PmiD: 19485974 stRinE, t.W.; moKDaD, a.h.; DubE, s.R.; Et al. the association of depression and anxiety naJDoWsKi, C.J., anD ullman, s.E. Prospective effects of sexual victimization on PtsD with obesity and unhealthy behaviors among community-dwelling us adults. General and problem drinking. Addictive Behaviors 34(11):965–968, 2009. PmiD: 19501469 Hospital Psychiatry 30(2):127–137, 2008. PmiD: 18291294 national institute on alcohol abuse and alcoholism (niaaa). Drinking Statistics. substance abuse and mental health services administration (samhsa). Results From available at: http://www.niaaa.nih.gov/alcohol-health/overview-alcohol-consumption/ the 2010 national survey on Drug use and health: summary of national Findings. drinking-statistics. accessed august 31, 2012. nsDuh series h–41, hhs Publication no. (sma) 11–4658. Rockville, mD: samsha, 2011. niEmElä, o., anD alatalo, P. biomarkers of alcohol consumption and related liver disease. sun, Q.; toWnsEnD, m.K.; oKEREKE, o.i. Et al. alcohol consumption at midlife and success- Scandinavian Journal of Clinical and Laboratory Investigation 70(5):305–312, 2010. ful ageing in women: a prospective cohort analysis in the nurses’ health study. PLoS PmiD: 20470213 Medicine 8(9):e1001090, 2011. PmiD: 21909248

onlanD-moREt, n.C.; PEEtERs, P.h.; Van DER sChouW, y.t.; Et al. alcohol and endogenous suzuKi, R.; iWasaKi, m.; inouE, m.; Et al. alcohol consumption-associated breast cancer sex steroid levels in postmenopausal women: a cross-sectional study. Journal of incidence and potential effect modifiers: the Japan Public health Center-based Clinical Endocrinology and Metabolism 90(3):1414–1419, 2005. PmiD: 15572431 Prospective study. International Journal of Cancer 127(3):685–695, 2010. PmiD: 19960437 PaRK, s.m.; Cho, s.i.; anD moon, s.s. Factors associated with suicidal ideation: Role of emotional and instrumental support. Journal of Psychosomatic Research 69(4):389– tayloR, b.; iRVing, h.m.; baliunas, D.; EEt al. alcohol and hypertension: gender differ- 397, 2010. PmiD: 20846540 ences in dose-response relationships determined through systematic review and meta- analysis. Addiction 104(12):1981–1990, 2009. PmiD: 19804464 PaRKs, K.a.; Romosz, a.m.; bRaDizza, C.m.; anD hsiEh, y.P. a dangerous transition: Women’s drinking and related victimization from high school to the first year at college. tEsta, m.; anD hoFFman, J.h. naturally occurring changes in women’s drinking from high Journal of Studies on Alcohol and Drugs 69(1):65–74, 2008. PmiD: 18080066 school to college and implications for sexual victimization. Journal of Studies on Alcohol and Drugs 73(1):26–33, 2012. PmiD: 22152659 PatRa, J.; tayloR, b.; iRVing, h.; Et al. alcohol consumption and the risk of morbidity and mortality for different stroke types: a systematic review and meta-analysis. BMC Public tEsta, m., anD liVingston, J.a. alcohol consumption and women’s vulnerability to sexual Health 10:258, 2010. PmiD: 20482788 victimization: Can reducing women’s drinking prevent rape? Substance Use & Misuse 44(9–10):1349–1376, 2009. PmiD: 19938922 Pino, n.W., anD Johnson-Johns, a.m. College women and the occurrence of unwanted sexual advances in public drinking settings. The Social Science Journal 46(2):252– tEsta, m.; liVingston, J.a.; anD hoFFman, J.h. Does sexual victimization predict subse- quent alcohol consumption? a prospective study among a community sample of 267, 2009. women. Addictive Behaviors 32(12):2926–2939, 2007. PmiD: 17597304 PiRan, n., anD gaDalla, t. Eating disorders and substance abuse in Canadian women: a tEsta, m.; VanzilE-tamsEn, C.; anD liVingston, J.a. the role of victim and perpetrator intoxi- national study. Addiction 102(1):105–113, 2007. PmiD: 17207128 cation on sexual assault outcomes. Journal of Studies on Alcohol 65(3):320–329, PREsCott, C.a.; aggEn, s.h.; anD KEnDlER, K.s. sex-specific genetic influences on the 2004. PmiD: 15222588 comorbidity of alcoholism and major depression in a population-based sample of us thomPson, m.P., anD KingREE, J.b. the roles of victim and perpetrator alcohol use in inti- twins. Archives of General Psychiatry 57(8):803–811, 2000. PmiD: 10920470 mate partner violence outcomes. Journal of Interpersonal Violence 21(2):163–177, RamstEDt, m. alcohol and suicide at the population level: the Canadian experience. 2006. PmiD: 16368759 Drug and Alcohol Review 24(3):203–208, 2005. PmiD: 16096123 tiDEmalm, D.; långstRöm, n.; liChtEnstEin, P.; anD RunEson, b. Risk of suicide after suicide REhm, J.; tayloR, b.; mohoPatRa, s.; Et al. alcohol as a risk factor for liver cirrhosis: a sys- attempt according to coexisting psychiatric disorder: swedish cohort study with long tematic review and meta-analysis. Drug and Alcohol Review 29(4):437–445, 2010. term follow-up. BMJ 337:a2205, 2008. PmiD: 19018040 PmiD: 20636661 timKo, C.; sutKoWi, a.; PaVao, J.; anD KimERling, R. Women’s childhood and adult adverse Rimm, E.b., anD moats, C. alcohol and coronary heart disease: Drinking patterns and experiences, mental health, and binge drinking: the California Women’s health survey. mediators of effect. Annals of Epidemiology 17(5):s3–s7, 2007. Substance Abuse Treatment, Prevention, and Policy 3:15, 2008. PmiD: 18538028

Women and the Costs of Alcohol Use 227 tolstRuP, J.; JEnsEn, m.K.; tJønnElanD, a.; Et al. Prospective study of alcohol drinking pat- Williams, m.; mohsin, m.; WEbER, D.; Et al. alcohol consumption and injury risk: a case- terns and coronary heart disease in women and men. BMJ 332(7552): 1244–1248, crossover study in sydney, australia. Drug and Alcohol Review 30(4):344–354, 2011. 2006. PmiD: 16672312 PmiD: 21355919

tuCKER, K.l.; JugDaohsingh, R.; PoWEll, J.J.; Et al. Effects of beer, wine, and liquor intakes WilsnaCK, R.W.; KRistJanson, a.F.; WilsnaCK, s.C.; anD CRosby, R.D. are u.s. women drink- on bone mineral density in older men and women. American Journal of Clinical ing less (or more)? aging and historical trends, 1981–2001. Journal of Studies on Nutrition 89(4):1188–1196, 2009. PmiD: 19244365 Alcohol 67(3):341–348, 2006. PmiD: 16608142

WilsnaCK, R.W.; WilsnaCK, s.C.; KRistJanson, a.F.; Et al. gender and alcohol consump- ullman, s.E.; KaRabatsos, g.; anD Koss, m.P. alcohol and sexual assault in a national tion: Patterns from the multinational gEnaCis project. Addiction 104(9):1487–1500, sample of college women. Journal of Interpersonal Violence 14(6):603–625, 1999. 2009. PmiD: 19686518 ViKhiREVa, o.; PiKhaRt, h.; PaJaK, a.; Et al. non-fatal injuries in three Central and Eastern WilsnaCK, s.C.; WilsnaCK, R.W.; KRistJanson, a.F.; Et al. alcohol use and suicidal behavior European urban population samples: the haPiEE study. European Journal of Public in women: longitudinal patterns in a u. s. national sample. Alcoholism: Clinical and Health 20(6):695–701, 2010. PmiD: 19959615 Experimental Research 28(5 suppl):38s–47s, 2004. PmiD: 15166635

WaRD, h.; lubEn, R.n.; WaREham, n.J.; anD KhaW, K.-t. ChD risk in relation to alcohol WosJE, K.s., anD KalKWaRF, h.J. bone density in relation to alcohol intake among men intake from categorical and open-ended dietary instruments. Public Health Nutrition and women in the united states. Osteoporosis International 18(3):391–400, 2007. 14(3):402–409, 2011. PmiD: 20707945 PmiD: 17091218

Waugh, E.J,; lam, m.-a.; haWKER, g.a.; Et al. Risk factors for low bone mass in healthy zhang, s.m.; lEE, i.m.; manson, J.E.; Et al. alcohol consumption and breast cancer risk in 40-60 year old women: a systematic review of the literature. Osteoporosis International the Women’s health study. American Journal of Epidemiology 165(6):667–676, 2007. 20(1):1–21, 2009. PmiD: 18523710 PmiD: 17204515 now Available – nIAAA Spectrum volume 5, issue 3 the latest issue of the nIAAA Spectrum, an online magazine featuring information from NiAAA and the alcohol research field. Each issue includes feature-length stories, news updates, engaging graphics, and an interview with an NiAAA staff member or prominent researcher in the field.

For more information and to subscribe, go to http://www.spectrum.niaaa.nih.gov/

228 Alcohol Research: Current Reviews disproportionate to their consumption. Differences in social Focus on: Ethnicity and the Social and socioeconomic factors and biological differences related to alcohol metabolism also could contribute to alcohol’s and Health Harms From drinking varying effects across populations. This article reviews current research examining the harms of drinking for U.S. ethnic groups. It examines such social harms as driving under the influence and alcohol-attributed violence but primarily focuses on health harms like fetal alcohol syndrome (FAS), liver diseases, and cancers. karen G. Chartier, Ph.D.; Patrice A.C. vaeth, Dr.P.H.; and The research reviewed focuses on Whites, Blacks, Raul Caetano, M.D., Ph.D. Hispanics, Asians, and Native Americans (i.e., American Indians and Alaska Natives) in the United States as general ethnic groups, although significant subgroup differences alcohol consumption is differentially associated with social and within populations also are evident. There are limitations to health harms across u.s. ethnic groups. native americans, using these general categories because ethnicity encompasses hispanics, and blacks are disadvantaged by alcohol-attributed a combination of characteristics such as tribe, ancestry, harms compared with Whites and asians. Ethnicities with higher national group, birthplace, and language, which could have rates of risky drinking experience higher rates of drinking harms. distinct relationships to patterns of drinking and alcohol- other factors that could contribute to the different effects of related harms (Caetano 1986; Cheung 1993; Heath 1990– alcohol by ethnicity are social disadvantage, acculturation, drink 1991). People with multiethnic backgrounds also are not preferences, and alcohol metabolism. this article examines the well represented by these general groups. Nevertheless, studies relationship of ethnicity and drinking to (1) unintentional that examine ethnicity and alcohol-attributed harms provide injuries, (2) intentional injuries, (3) fetal alcohol syndrome important information about public health and serve to (Fas), (4) gastrointestinal diseases, (5) cardiovascular identify high-risk groups in the population. This article diseases, (6) cancers, (7) diabetes, and (8) infectious diseases. shows that Native Americans, Hispanics, and Blacks are Reviewed evidence shows that native americans have a disproportionately affected by the adverse social and health disproportionate risk for alcohol-related motor vehicle fatalities, harms from alcohol consumption. suicides and violence, Fas, and liver disease mortality. hispanics are at increased risk for alcohol-related motor vehicle fatalities, suicide, liver disease, and cirrhosis mortality; and Drinking Patterns and other Determinants of Risk blacks have increased risk for alcohol-related relationship for Alcohol-Related Harms violence, Fas, heart disease, and some cancers. however, the scientific evidence is incomplete for each of these harms. more Heavy drinking and binge drinking contribute to a variety research is needed on the relationship of alcohol consumption of alcohol-attributed social and health harms (Naimi et al. to cancers, diabetes, and hiV/aiDs across ethnic groups. 2003; Rehm et al. 2010). Heavy alcohol use, as defined by studies also are needed to delineate the mechanisms that give the National Institute on Alcohol Abuse and Alcoholism’s rise to and sustain these disparities in order to inform prevention (NIAAA’s) Helping Patients Who Drink Too Much: A Clinician’s strategies. kEY WoRDS: Alcohol consumption; alcohol-attributable Guide (NIAAA 2005), is defined as consuming more than 4 fractions; alcohol burden; harmful drinking; alcohol and other standard drinks per day (or more than 14 per week) for men drug–induced risk; risk factors; ethnicity; ethnic groups; racial and more than 3 per day (or more than 7 per week) for groups; cultural patterns of drinking; Native Americans; women. One standard drink is equivalent to 12 ounces of Hispanics; Blacks; African Americans; Asian Americans; Whites; Caucasians; injury; intentional injury; unintentional injury; fetal karen G. Chartier, Ph.D., is an instructor and assistant alcohol syndrome; gastrointestinal diseases; cardiovascular diseases; cancers; diabetes; infectious diseases professor at the Virginia Commonwealth University School of Social Work and Department of Psychiatry with the Virginia Institute for Psychiatric and Behavioral Genetics, Richmond, Virginia.

esearch has shown differential social and health effects Patrice A.C. vaeth, Dr.P.H., is a scientist at the Prevention from alcohol use across U.S. ethnic groups, including Research Center, Pacific Institute for Research and RWhites, Blacks, Hispanics, Asians, and Native Americans. The relationship of ethnicity to alcohol-related social and Evaluation, Berkeley, California. health harms partially is attributed to the different rates and patterns of drinking across ethnicities. Some ethnic groups Raul Caetano, M.D., Ph.D., is regional dean and professor have higher rates of alcohol consumption, putting them at at the University of Texas School of Public Health, Dallas greater risk of drinking harms. However, other ethnic Regional Campus, Dallas Texas. minorities experience health harms from drinking that are

Ethnicity and the Social and Health Harms From Drinking 229 beer, 5 ounces of wine, or 1.5 ounces of 80-proof spirits. reported higher volume for Hispanics. Other ethnic minority Binge drinking is defined as consuming five or more drinks groups with higher abstinence rates include Blacks (24.7 per- in approximately 2 hours for men and four or more drinks cent) and Asians (39.1 percent). Native Americans (17.14 per- for women (NIAAA 2004). cent) have lower rates of abstinence than other minority groups. Other than these patterns of consumption, the volume Alternatively, the negative effects from drinking could of alcohol intake, defined as the total alcohol consumed over be explained by factors other than alcohol consumption. a time period, is linked to social and health harms. Most dis- Mulia and colleagues (2009) showed that Black and Hispanic eases (e.g., injury, some cancers, and liver cirrhosis) have a adult drinkers were more likely than White drinkers to detrimental dose-response relationship with alcohol as risk report alcohol dependence symptoms and social problems increases with higher-volume alcohol consumption, whereas from drinking at the no/low level of heavy drinking. Blacks coronary heart disease and diabetes display a J- or U-shaped also experience negative health effects from alcohol use relationship (Howard et al. 2004; Rehm et al. 2010; Roerecke despite showing a later onset of use and levels of use often and Rehm 2012). The J and U shapes are characterized by comparable with, if not lower than, Whites (Chartier et al. both detrimental and beneficial (e.g., increased high-density 2011; Chen et al. 2006; Russo et al. 2004). Other factors lipoprotein “good cholesterol”) (Goldberg and Soleas 2001) associated with ethnic disparities in alcohol-related harms effects of alcohol use, with higher risks for abstainers and include social disadvantage, characterized by lower socioeco- heavy drinkers compared with light or moderate drinkers. nomic status, neighborhood poverty, greater neighborhood However, this relationship is complex and varies by age, alcohol availability, reduced alcohol treatment utilization, gender, and ethnicity (Roerecke and Rehm 2012). Drinking and unfair treatment or discrimination (Chae et al. 2008; levels that may be protective of cardiovascular health among Chartier and Caetano 2011; Cunradi et al. 2000; Mulia et men also may increase the risk for other harms such as al. 2008; Nielsen et al. 2005; Zemore et al. 2011). Some injury, violence, gastrointestinal disease, and some cancers. ethnic subgroups are more likely to consume high-alcohol- Epidemiological studies show that these high-risk pat- content beverages (e.g., malt liquor), which could result in terns of drinking and drinking volume vary by U.S. ethnic greater social and health harms (Vilamovska et al. 2009). group. Ethnicities with greater drinking volume and higher Preference for such beverages seems to be more common rates of daily and weekly heavy drinking could be at greater in lower-income ethnic minority communities (Bluthenthal risk for experiencing alcohol-attributed harms. Among adult et al. 2005). Some ethnic minority groups also face stressors drinkers in the United States, based on the 2001–2002 related to the acculturation process. Higher acculturation, National Epidemiologic Survey on Alcohol and Related U.S.-born nativity, and longer residence in the United States Conditions (NESARC) (Chen et al. 2006), Native Americans are risk factors associated with alcohol use disorders and and Hispanics have greater alcohol consumption than other alcohol-related social problems among Hispanics, particu- ethnic minority groups. Rates of daily heavy drinking were larly women (Alegria et al. 2007, 2008; Caetano et al. 2009, higher among Hispanics (33.9 percent), Native Americans 2012; Zemore 2007). Another potential contributor is eth- (28.4 percent), and Whites (27.3 percent) compared with nic differences in the alcohol content of poured drinks. Kerr Blacks (22.5 percent) and Asians (19.2 percent). Weekly and colleagues (2009) showed that Black men had drink heavy drinking was highest among Native Americans (21.9 sizes with larger average alcohol content compared with percent), followed by Blacks (16.4 percent), Whites (16.3 other groups, which partially could explain the higher risks percent), Hispanics (11.8 percent), and Asians (9.8 percent). for alcohol-related harms. Genes responsible for alcohol Based on the 2001–2002 NESARC data, Caetano and col- metabolism also vary across ethnic groups and could be asso- leagues (2010) reported that White men consumed a higher ciated with susceptibility for alcohol-related diseases. Among volume of alcohol (22.3 drinks per month) than Black men Whites, Blacks, and Asians, alcohol dehydrogenase (ADH) (18.9 drinks per month) and Hispanic men (17.8 drinks per and aldehyde dehydrogenase (ALDH) genotypes have been month) and that White women consumed more (6.2 drinks linked in combination with drinking to alcohol-related can- per month) compared with Black women (4.9 drinks per cers, birth defects, and pancreatitis (Yin and Agarwal 2001). month) and Hispanic women (3.9 drinks per month). The sample for these estimates of drinking volume was the U.S. population of Whites, Blacks, and Hispanics and included Ethnicity and Alcohol-Attributed Harms abstainers. However, a study by Mulia and colleagues (2009) of current drinkers in the United States showed that Whites Alcohol-attributed harms can be both acute and chronic consumed less alcohol than Hispanics and more than Blacks. conditions that are wholly caused (e.g., alcoholic liver cirrhosis) The differences between these two studies could reflect a or associated with alcohol use via intoxication, alcohol depen- higher rate of abstinence from alcohol among Hispanics dence, and the toxic effects of alcohol (Rehm et al. 2010). (25.7 percent) compared with Whites (13.4 percent) in the The major injury and disease categories linked to alcohol U.S. population (Chen et al. 2006). The study that included consumption include (1) unintentional injuries, (2) intentional abstainers (Caetano et al. 2010), who by definition consume injuries, (3) FAS, (4) gastrointestinal diseases, (5) cardiovas- zero drinks, showed higher drinking volume for Whites, cular diseases, (6) cancers, (7) diabetes, and (8) infectious whereas the study excluding abstainers (Mulia et al. 2009) diseases (World Health Organization [WHO] 2011). Evidence

230 Alcohol Research: Current Reviews is incomplete on the relationship between ethnicity, drinking, on 2005–2006 data from the National Violent Death and each of these categories. Below, those alcohol-related harms Reporting System presented findings on alcohol and suicide are described that have available findings by ethnic group in across ethnic groups. Recent alcohol use was reported among addition to important gaps in this scientific literature. Alcohol suicides in 46 percent of Native Americans, 30 percent of use disorders are causally linked to drinking and vary by eth- Hispanics, 26 percent of Whites, 16 percent of Blacks, and nicity (i.e., more likely in Native Americans and Whites) (Hasin 15 percent of Asians. Among those tested for alcohol, the et al. 2007), but this disease category is not described here. rates of intoxication (BAC higher than or equal to 0.08) were highest for Native Americans (37 percent), followed by Hispanics (29 percent), Whites (24 percent), Blacks (14 Unintentional injuries percent), and Asians (12 percent). Age-groups identified as being at high risk for alcohol-involved suicide included Unintentional injuries associated with alcohol use include Native Americans ages 30 to 39 (54 percent of suicide falls, drowning, and poisoning (WHO 2011). However, victims had BACs higher than or equal to 0.08), Native most available research on ethnicity, alcohol use, and injuries Americans and Hispanics ages 20 to 29 (50 percent and 37 is focused on motor vehicle crashes. Alcohol-impaired driving percent, respectively), and Asians ages 10 to 19 (29 percent). and crash fatalities vary by ethnicity, with Native Americans Males were at higher risk than female drinkers in all ethnic and Hispanics being at higher risk than other ethnic minority groups except Native Americans; the percentages of alcohol groups. Past-year driving under the influence (DUI) estimates intoxication among Native American suicides were equal for based on the 2007 National Survey on Drug Use and Health males and females (37 percent). were highest for Whites (15.6 percent) and Native Americans (13.3 percent) relative to Blacks (10.0 percent), Hispanics Violence (9.3 percent), and Asians (7.0 percent) (Substance Abuse and Mental Health Services Administration [SAMHSA] 2008). Ethnic groups are differentially affected by alcohol-attributed National surveys generally show lower DUI rates for Hispanics violence, including intimate-partner violence (IPV). Alcohol than Whites, but studies based on arrest data identify Hispanics plays an important role in IPV and other types of relation- as another high-risk group for DUI involvement (Caetano ship conflicts (Field and Caetano 2004; Leonard and Eiden and McGrath 2005; SAMHSA 2005). The DUI arrest rate 2007). Based on data from the National Study of Couples, for Native Americans in 2001, according to the U.S. Department general rates of male-to-female partner violence (MFPV) of Justice (Perry 2004), was 479 arrestees per 100,000 resi- and female-to-male partner violence (FMPV), are highest dents compared with 332 for all other U.S. ethnic groups. among Black couples (23 percent and 30 percent, respec- tively), followed by Hispanic (17 percent and 21 percent) Based on a 1999–2004 report from the National Highway and White (12 percent and 16 percent) couples (Caetano et Traffic Safety Administration (Hilton 2006), rates of intoxi- al. 2000). The National Study of Couples provides general cation (i.e., blood alcohol concentration [BAC] more than population data on IPV, which includes mostly moderate or equal to 0.08 percent) for drivers who were fatally injured violence and may differ from other studies of severe violence. in a motor vehicle crash were highest for Native Americans In this study, regardless of ethnicity, men were more likely (57 percent) and Hispanics (47 percent) and lowest for Asians than women to report drinking during partner violence. (approximately 20 percent), with Whites and Blacks falling Drinking during a violent episode by the male or the female in between. Across ethnic groups, most drinking drivers killed partner, respectively, was more frequent among Blacks were male, although the proportion of female drivers who (MFPV: 41.4 percent and 23.6 percent; FMPV: 33.7 percent were intoxicated among fatally injured drivers was highest and 22.4 percent) than among Whites (MFPV: 29.4 percent (i.e., more than 40 percent) for Native Americans. Centers for and 11.4 percent; FMPV: 27.1 percent and 14.7 percent) Disease Control and Prevention (CDC) (2009b) statistics on and Hispanics (MFPV: 29.1 percent and 5.4 percent; alcohol-related motor vehicle crash deaths also point to an FMPV: 28.4 percent and 3.8 percent). Longitudinal findings, important subgroup difference for Asians. In 2006, the overall using 5-year National Study of Couples data, identified death rate among Asians (1.8 per 100,000 people) obscured female-partner alcohol problems (i.e., alcohol dependence the death rate among Native Hawaiians and other Pacific symptoms and social problems) in Black couples and male- Islanders (5.9), which was less than the rate for Native Americans and female-partner alcohol consumption in White couples but similar to that for Hispanics (14.5 and 5.2, respectively). as risk factors for IPV (Field and Caetano 2003). Some evidence also suggests that interethnic couples, involving White, Black, and Hispanic partners of different ethnic backgrounds, are a intentional injuries high-risk group for relationship violence. Relative to intraethnic couples, these interethnic couples had higher prevalence rates Suicide of IPV, which was associated with binge drinking and alcohol problems among male partners (Chartier and Caetano 2012). Native Americans are overrepresented in national estimates Alcohol also contributes to violence victimization of alcohol-involved suicides. A CDC report (2009a) based among Native Americans (Yuan et al. 2006). Several studies

Ethnicity and the Social and Health Harms From Drinking 231 indicate that Native Americans are at greater risk for alcohol- respectively) compared with other ethnic groups (0 percent related trauma (e.g., IPV, rape, and assault) compared with to 10.7 percent and 21.0 percent to 37.3 percent). White other U.S. ethnic groups (Oetzel and Duran 2004; Wahab women in this study were at greater risk for an alcohol- and Olson 2004). Based on 1992–2001 National Crime exposed pregnancy. However, other studies found that Black, Victimization Survey data, the U.S. Department of Justice Hispanic, and Asian women were less likely to reduce or quit (Perry 2004) reported that 42 percent of all violent crimes heavy drinking after becoming pregnant (Morris et al. 2008; (i.e., rape, sexual assault, robbery, aggravated assault, and Tenkku et al. 2009). Blacks and Native Americans are at simple assault) were committed by an offender who was greater risk than Whites for FAS and fetal alcohol spectrum under the influence of alcohol. In particular, Native American disorders (Russo et al. 2004). From 1995 to 1997, FAS rates violent crime victims were more likely (62 percent) than averaged 0.4 per 1,000 live births across data-collection sites other violent crime victims to report alcohol use by their for the Fetal Alcohol Syndrome Surveillance Network and offender, including Whites (43 percent), Blacks (35 percent), were highest for Black (1.1 percent) and Native American and Asians (33 percent). (3.2 percent) populations (CDC 2002).

Fetal Alcohol Syndrome Gastrointestinal Diseases Using data from the 2001–2002 NESARC, Caetano and Liver disease is an often-cited example of the disproportionate colleagues (2006) examined alcohol consumption, binge effect of alcohol on health across ethnic groups. Native Americans drinking, and alcohol abuse and dependence among women have higher mortality rates for alcoholic liver disease than who were pregnant during the past year. Most women (88 other U.S. ethnic groups (see figure). According to the percent) who reported being pregnant and also a drinker at National Vital Statistical Reports (Miniño et al. 2011) on any point in the past 12 months indicated that they did not 2008 U.S. deaths, age-adjusted death rates attributed to drink during pregnancy. Rates of past-year alcohol abuse alcoholic liver disease for Native American men and women (0.8 percent to 2.3 percent) and dependence (1.2 percent to were 20.4 and 15.3 per 100,000 people, respectively, com- 2.8 percent) were similar and low in White, Black, Hispanic, pared with 6.9 and 2.4 per 100,000 for men and women in and Asian pregnant women. Binge drinking and alcohol the general population. consumption without binge drinking among pregnant women Blacks and Hispanics have greater risk for developing were highest in Whites (21.1 percent and 45.0 percent, liver disease compared with Whites (Flores et al. 2008), and

25

Male

20 Female

15

10 Age-adjusted death rates Age-adjusted

(per 100,000 population) 5

0 U.S. Total White Black Native American Asian Hispanic Non-Hispanic Ethnicity

Figure in 2008, age-adjusted death rates attributed to alcoholic liver disease for native american men and women were 20.4 and 15.3 per 100,000 people, respectively, compared with 6.9 and 2.4 for men and women in the general population.

souRCE: miniño, a.m. et al., Deaths: Final data for 2008. National Vital Statistics Reports 59(10):1–52, 2011.

232 Alcohol Research: Current Reviews death rates attributed to alcohol-related cirrhosis across pop- higher among men than women (National Cancer Institute ulations of Whites, Blacks, and Hispanics are highest for 2011c, d, e). The incidence and mortality rates for these can- White Hispanic men (Yoon and Yi 2008). Blacks show a cers also vary across ethnic groups. Regarding cancers of the greater susceptibility than Whites to alcohol-related liver oral cavity and pharynx, incidence rates among White and damage, with risk differences amplified at higher levels of Black men are comparable (16.1 and 15.6 per 100,000, consumption (Stranges et al. 2004). Based on data from the respectively); however, mortality rates are higher among National Center for Health Statistics, 1991–1997, mortality Black men (6.0 versus 3.7 per 100,000 for White men) rates for cirrhosis with mention of alcohol were higher in (National Cancer Institute 2011e). For cancer of the larynx, White Hispanics and Black non-Hispanics compared with both incidence and mortality rates are higher among Black White non-Hispanics (Stinson et al. 2001). Male mortality men than among White men (incidence, 9.8 and 6.0; rates for alcohol-related cirrhosis in White Hispanics and mortality, 4.4 and 2.0) (National Cancer Institute 2011c). non-Hispanic Blacks were 114 percent and 24 percent higher, Although these differences may be explained by differential respectively, than the overall male rate (5.9 deaths per 100,000 use of alcohol and tobacco in relation to gender and ethnic- people); female rates in White Hispanics and non-Hispanic ity, there is some evidence that even after controlling for Blacks were 16 percent and 47 percent higher than the overall alcohol and tobacco use, Blacks continue to be at increased female rate (1.9 deaths per 100,000 people). In contrast, risk for squamous cell esophageal cancer and cancers of the death rates for White non-Hispanic and Black Hispanic males oral cavity and pharynx (Brown et al. 1994; Day et al. 1993). and females were lower than overall rates for each gender. In The majority (approximately 90 percent) of all primary addition, there is considerable variation in deaths from liver liver cancers are hepatocellular carcinomas (HCC) (Altekruse cirrhosis across Hispanic subgroups, with mortality rates et al. 2009). Alcohol-related and non–alcohol-related liver highest in Puerto Ricans and Mexicans and lowest in Cubans cirrhosis usually precede HCC and are the two most common (Yoon and Yi 2008). risk factors (Altekruse et al. 2009; El-Serag 2011; Pelucchi et al. 2006). The relative risk for developing this cancer increases with increased levels of alcohol consumption (Pelucchi Cardiovascular Diseases et al. 2006). By ethnic group, 2003–2005 age-adjusted inci- dence rates for HCC per 100,000 persons were highest Although moderate alcohol consumption has been associated among Asians (11.7), followed by Hispanics (8.0), Blacks with a reduced risk for coronary heart disease (CHD) (Goldberg (7.0), Native Americans (6.6), and Whites (3.9) (Altekruse and Soleas 2001), there is some evidence that ethnic groups et al. 2009). Death rates for HCC per 100,000 people also differ in terms of this protective effect, particularly for Blacks are higher among minority groups (i.e., 8.9, 6.7, 5.8, 4.9, compared with Whites. Sempos and colleagues (2003) and 3.5 for Asians, Hispanics, Blacks, Native Americans, and found no protective health effect for moderate drinking in Blacks for all-cause mortality, as previously reported in Whites, respectively). Whites. Kerr and colleagues (2011) reported the absence In 2007, the IARC reconvened and added breast and of this protective effect for all-cause mortality in Blacks and colorectal cancers to the list of cancers related to alcohol use Hispanics. Similar findings have been described for hyper- (Baan et al. 2007). Research has demonstrated consistent, tension and CHD risks in Black men compared with White albeit weak, dose-response relationships between alcohol men and women (Fuchs et al. 2001, 2004) and for mortality consumption and these cancers (Cho et al. 2004; Collaborative among Black women without hypertension (Freiberg et al. Group on Hormonal Factors in Breast Cancer 2002; Moskal 2009). Mukamal and colleagues (2010) also showed that the et al. 2007; Singletary and Gapstur 2001). Alcohol con- protective effects of light and moderate drinking in cardio- sumption also contributes to the stage at which breast cancer vascular mortality were stronger among Whites than non- is diagnosed (Hebert et al. 1998; Trentham-Dietz et al. Whites. Pletcher and colleagues (2005) found evidence that 2000; Vaeth and Satariano 1998; Weiss et al. 1996). This the dose-response relationship between alcohol consumption could be because of the timing of disease detection, since and increased coronary calcification, a marker for CHD, was heavy drinking has been associated with a lack of mammog- strongest among Black men. raphy utilization (Cryer et al. 1999). Alcohol consumption also may contribute to more rapid tumor proliferation (Singletary and Gapstur 2001; Weiss et al. 1996). Data from Cancers the Surveillance, Epidemiology, and End Results (SEER) Program indicate that White women, relative to women In 1988, the WHOInternational Agency for Research on from ethnic minority groups, have higher incidence rates Cancer (IARC) reviewed the epidemiologic evidence on the of breast cancer (i.e., Whites, 127.3; Blacks, 119.9; Asians, association between alcohol consumption and cancer and 93.7; Native Americans, 92.1; and Hispanics, 77.9 per found a consistent association between alcohol consumption 100,000 people) (National Cancer Institute 2011a). Black and increased risk for cancers of the oral cavity, pharynx, women, however, are more likely to be diagnosed with larynx, esophagus, and liver (IARC 1988). Regardless of advanced disease (Chlebowski et al. 2005) and have signifi- ethnicity, the risk of developing these cancers is significantly cantly higher mortality rates than White women (i.e., 32.0

Ethnicity and the Social and Health Harms From Drinking 233 per 100,000 versus 22.8 per 100,000 people) (Chlebowski et Cook et al. 2001; Samet et al. 2004) and HIV disease pro- al. 2005; National Cancer Institute 2011a). Regarding colorectal gression (Conigliaro et al. 2003; Samet et al. 2003). Despite cancer, Blacks have higher incidence (67.7) and mortality these strong individual associations between ethnicity and (51.2) rates than all ethnic groups combined (55.0 and 41.0, HIV/AIDS and alcohol and HIV/AIDS, there is limited respectively) (National Cancer Institute 2011b). Unfortunately, research across ethnicities on alcohol use and HIV infection little is known about how drinking differentially affects ethnic or disease progression. differences in breast and colorectal cancers. Conclusions Diabetes This article identifies U.S. ethnic-group differences in alcohol- In 2010, the prevalence of diabetes was 7.1 percent, 12.6 attributed social and health-related harms. Three minority percent, 11.8 percent, and 8.4 percent among Whites, Blacks, ethnicities are particularly disadvantaged by alcohol-related Hispanics, and Asians, respectively (National Institute of harms. Native Americans, relative to other ethnic groups, Diabetes and Digestive and Kidney Diseases 2011). Age- have higher rates of alcohol-related motor vehicle fatalities, adjusted mortality rates in 2007 were 20.5, 42.8, 28.9, and suicide, violence, FAS, and liver disease mortality. Unlike 16.2 per 100,000 people among Whites, Blacks, Hispanics, other ethnic groups, in which men are primarily at risk for and Asians (National Center for Health Statistics 2011). alcohol-related harms, both Native American men and Data on mortality rates for diabetes among Hispanics may women are high-risk groups. Hispanics have higher rates of be underreported as a result of inconsistencies in the report- alcohol-related motor vehicle fatalities, suicide, and cirrhosis ing of Hispanic origin on death certificates (Heron et al. mortality. Blacks have higher rates of FAS, intimate partner 2009). Despite higher risks for the development of and violence, and some head and neck cancers, and there is lim- death from diabetes in Hispanics and Blacks compared with ited empirical support in Blacks for a protective health effect Whites, little evidence is available to delineate the relation- from moderate drinking. These patterns of findings provide ship of alcohol to diabetes across ethnic groups. Studies recognition of the health disparities in alcohol-attributed among both diabetics and nondiabetics demonstrate a J- or harms across U.S. ethnicities. However, further research U-shaped curve between alcohol consumption and insulin is needed to identify the mechanisms that give rise to and sensitivity (Bell et al. 2000; Davies et al. 2002; Greenfield et al. 2003; Kroenke et al. 2003). Likewise, two large epi- sustain these disparities in order to develop prevention strate- demiologic studies among diabetic subjects show that mod- gies. The contributing factors include the higher rates of erate alcohol consumption is associated with better glycemic consumption found in Native Americans and Hispanics, control (Ahmed et al. 2008; Mackenzie et al. 2006). An but more broadly range from biological factors to the social important limitation of these studies, however, is that few environment. More research on the relationship of alcohol to included ethnic minority groups or failed to emphasize some cancers, diabetes, and HIV/AIDs across ethnic groups possible differences in relation to ethnicity in their analyses. is also needed. There is limited evidence for how drinking differentially affects ethnic differences in breast and colorec- tal cancers and in diabetes and HIV/AIDS onset and care, infectious Diseases and few findings for how alcohol-attributed harms vary across ethnic subgroups. ■ Among the infectious diseases attributable to alcohol (e.g., pneumonia, tuberculosis) (WHO 2011), human immunod- eficiency virus (HIV) and acquired immunodeficiency syndrome Acknowledgements (AIDS) are most relevant to U.S. ethnic health disparities. In 2009, Blacks represented 44 percent of new HIV infections The authors thank Kierste Miller for her help in preparing and Hispanics represented 20 percent. Infection rates by this paper. This work was supported by NIAAA grant gender for Blacks were 15 times (for men) and 6.5 times (for R01–AA–016319. women) those of Whites, and rates for Hispanics were 4.5 times for men and 2.5 times for women, compared with rates for Whites (CDC 2011). In addition, alcohol consumption Financial Disclosure has been associated with increased HIV infection risk (Bryant et al. 2010). Caetano and Hines (1995) showed that heavy The authors declare that they have no competing financial drinking predicted high-risk sexual behaviors in White, interests. Black, and Hispanic men and women, with more Blacks than Whites and Hispanics reporting risky sexual behaviors. Among HIV-infected patients, there also is evidence that References increased alcohol consumption negatively affects adherence ahmED, a.t.; KaRtER, a.J.; WaRton, E.m.; Et al. the relationship between alcohol consump- to antiretroviral medication regimens (Chander et al. 2006; tion and glycemic control among patients with diabetes: the Kaiser Permanente

234 Alcohol Research: Current Reviews northern California Diabetes Registry. Journal of General Internal Medicine 23(3):275– ChaE, D.h.; taKEuChi, D.t.; baRbEau, E.m.; Et al. alcohol disorders among asian americans: 282, 2008. PmiD: 18183468 associations with unfair treatment, racial/ethnic discrimination, and ethnic identification (the national latino and asian americans study, 2002–2003). Journal of Epidemiology alEgRia, m.; Canino, g.J.; shRout, P.E.; Et al. Prevalence of mental illness in immigrant and Community Health 62(11):973–979, 2008. PmiD: 18854501 and non-immigrant u.s. latino groups. American Journal of Psychiatry 165(3):359– 369, 2008. PmiD: 18245178 ChanDER, g.; lau, b.; anD mooRE, R.D. hazardous alcohol use: a risk factor for non-adher- ence and lack of suppression in hiV infection. Journal of Acquired Immune Deficiency alEgRía, m.; sRibnEy, W.; Woo, m.; Et al. looking beyond nativity: the relation of age of immi- gration, length of residence, and birth cohorts to the risk of onset of psychiatric disorders Syndromes: JAIDS 43(4):411–417, 2006. PmiD: 17099312 for latinos. Research in Human Development 4(1–2):19–47, 2007. PmiD: 19412354 ChaRtiER, K.g., anD CaEtano, R. trends in alcohol services utilization from 1991–1992 to altEKRusE, s.F.; mCglynn, K.a.; anD REiChman, m.E. hepatocellular carcinoma incidence, 2001–2002: Ethnic group differences in the u.s. population. Alcoholism: Clinical and mortality, and survival trends in the united states from 1975 to 2005. Journal of Experimental Research 35(8):1485–1497, 2011. PmiD: 21575015 Clinical Oncology 27(9):1485–1491, 2009. PmiD: 19224838 ChaRtiER, K.g., anD CaEtano, R. intimate partner violence and alcohol problems in baan, R.; stRaiF, K.; gRossE, y.; Et al. Carcinogenicity of alcoholic beverages. Lancet interethnic and intraethnic couples. Journal of Interpersonal Violence, 27(9):1780– Oncology 8(4):292–293, 2007. PmiD: 17431955 1801, 2012. PmiD: 22203625 bEll, R.a.; mayER-DaVis, E.J.; maRtin, m.a.; Et al. associations between alcohol consump- ChaRtiER, K.g.; hEssElbRoCK, m.n.; anD hEssElbRoCK, V.m. alcohol problems in young adults tion and insulin sensitivity and cardiovascular disease risk factors: the insulin Resistance transitioning from adolescence to adulthood: the association with race and gender. and atherosclerosis study. Diabetes Care 23(11):1630–1636, 2000. PmiD: 11092284 Addictive Behaviors 36(3):167–174, 2011. PmiD: 21115225 bluthEnthal, R.n.; bRoWn tayloR, D.; guzman-bECERRa, n.; anD Robinson, P.l. Characteristics ChEn, C.m.; yi, h.-y.; FalK, D.E.; Et al. Alcohol Use and Alcohol Use Disorders in the United of malt liquor beer drinkers in a low-income, racial minority community sample. States: Main Findings from the 2001–2002 National Epidemiologic Survey on Alcohol Alcoholism: Clinical and Experimental Research 29(3):402–409, 2005. PmiD: 15770116 and Related Conditions (NESARC). bethesda, mD: national institutes of health, 2006. bRoWn, l.m.; hooVER, R.n.; gREEnbERg, R.s.; Et al. are racial differences in squamous cell ChEung, y.W. approaches to ethnicity: Clearing roadblocks in the study of ethnicity and esophageal cancer explained by alcohol and tobacco use? Journal of the National substance use. International Journal of the Addictions 28(12):1209–1226, 1993. Cancer Institute 86(17):1340–1345, 1994. PmiD: 8064893 PmiD: 8282450 bRyant, K.J.; nElson, s.; bRaithWaitE, R.s.; anD RoaCh, D. integrating hiV/aiDs and alcohol ChlEboWsKi, R.t.; ChEn, z.; anDERson, g.l.; Et al. Ethnicity and breast cancer: Factors influ- research. Alcohol Research and Health 33(3):167–178, 2010. encing differences in incidence and outcome. Journal of the National Cancer Institute CaEtano, R. alternative definitions of hispanics: Consequences in an alcohol survey. 97(6):439–448, 2005. PmiD: 15770008 Hispanic Journal of Behavioral Science 8(4):331–344, 1986. Cho, E.; smith-WaRnER, s.a.; Ritz, J.; Et al. alcohol intake and colorectal cancer: a pooled CaEtano, R.; baRuah, J.; RamisEtty-miKlER, s.; anD Ebama, m.s. sociodemographic predictors analysis of 8 cohort studies. Annals of Internal Medicine 140(8):603–613, 2004. PmiD: of pattern and volume of alcohol consumption across hispanics, blacks, and Whites: 15096331 10-year trend (1992–2002). Alcoholism: Clinical and Experimental Research 34(10): 1782–1792, 2010. PmiD: 20645935 Collaborative group on hormonal Factors in breast Cancer. alcohol, tobacco and breast cancer: Collaborative reanalysis of individual data from 53 epidemiological studies, CaEtano, R.; CunRaDi, C.b.; ClaRK, C.l.; anD sChaFER J. intimate partner violence and drink- including 58,515 women with breast cancer and 95,067 women without the disease. ing patterns among White, black, and hispanic couples in the u.s. Journal of British Journal of Cancer 87(11):1234–1245, 2002. PmiD: 12439712 Substance Abuse 11(2):123–138, 2000. PmiD: 10989773 ConigliaRo, J.; goRDon, a.J.; mCginnis, K.a.; Et al. how harmful is hazardous alcohol use CaEtano, R., anD hinEs, a.m. alcohol, sexual practices, and risk of aiDs among blacks, and abuse in hiV infection: Do health care providers know who is at risk? Journal of hispanics, and Whites. Journal of Acquired Immune Deficiency Syndromes and Human Acquired Immune Deficiency Syndromes: JAIDS 33(4):521–525, 2003. PmiD: 12869842 Retrovirology 10(5):554–561, 1995. PmiD: 8548335 CooK, R.l.; sEREiKa, s.m.; hunt, s.C.; Et al. Problem drinking and medication adherence CaEtano, R., anD mCgRath, C. Driving under the influence (Dui) among u.s. ethnic among persons with hiV infection. Journal of General Internal Medicine 16(2):83–88, groups. Accident Analysis and Prevention 37(2):217–224, 2005. PmiD: 15667807 2001. PmiD: 11251758 CaEtano, R.; RamisEtty-miKlER, s.; FloyD, l.R.; anD mCgRath, C. the epidemiology of drink- ing among women of child-bearing age. Alcoholism: Clinical and Experimental CRyER, P.C.; JEnKins, l.m.; CooK, a.C.; Et al. the use of acute and preventative medical Research 30(6):1023–1030, 2006. PmiD: 16737461 services by a general population: Relationship to alcohol consumption. Addiction 94(10):1523–1532, 1999. PmiD: 10790904 CaEtano, R.; RamisEtty-miKlER, s.; anD RoDRiguEz, l.a. the hispanic americans baseline alcohol survey (hablas): the association between birthplace, acculturation and alco- CunRaDi, C.b.; CaEtano, R.; ClaRK, C.; anD sChaFER J. neighborhood poverty as a predictor hol abuse and dependence across hispanic national groups. Drug and Alcohol of intimate partner violence among White, black, and hispanic couples in the united Dependence 99(1–3):215–221, 2009. PmiD: 18945554 states: a multilevel analysis. Annals of Epidemiology 10(5):297–308, 2000. PmiD: 10942878 CaEtano, R.; VaEth, P.a.; anD RoDRiguEz, l.a. the hispanic americans baseline alcohol survey (hablas): acculturation, birthplace and alcohol-related social problems across DaViEs, m.J.; baER, D.J.; JuDD, J.t.; Et al. Effects of moderate alcohol intake on fasting hispanic national groups. Hispanic Journal of Behavioral Sciences 34(1):95–117, insulin and glucose concentrations and insulin sensitivity in postmenopausal women. 2012. PmiD: 22438607 JAMA: Journal of the American Medical Association 287(19):2559–2562, 2002. PmiD: Centers for Disease Control and Prevention (CDC). Fetal alcohol syndrome: alaska, 12020337 arizona, Colorado, and new york, 1995–1997. Morbidity and Mortality Weekly Report Day, g.l.; blot, W.J.; austin, D.F.; Et al. Racial differences in risk of oral and pharyngeal 51(20):433–435, 2002. PmiD: 12056499 cancer: alcohol, tobacco, and other determinants. Journal of the National Cancer CDC. alcohol and suicide among racial/ethnic populations: 17 states, 2005–2006. Institute 85(6):465–473, 1993. PmiD: 8445674 Morbidity and Mortality Weekly Report 58(23):637–641, 2009a. PmiD: 19543198 El-sERag, h.b. hepatocellular carcinoma. New England Journal of Medicine CDC. DATA2010, The Healthy People 2010 Database, substance abuse, 2009b. 365(12):1118–1127, 2011. PmiD: 21992124 available at: http://wonder.cdc.gov/data2010/focus.htm. accessed January 27, 2010. FiElD, C.a., anD CaEtano, R. longitudinal model predicting partner violence among White, CDC. HIV in the United States. atlanta, ga: national Center for hiV/aiDs, Viral hepatitis, black and hispanic couples in the united states. Alcoholism: Clinical and Experimental stD, and tb Prevention, Divisions of hiV/aiDs Prevention, 2011. Research 27(9):1451–1458, 2003. PmiD: 14506406

Ethnicity and the Social and Health Harms From Drinking 235 FiElD, C.a., anD CaEtano, R. Ethnic differences in intimate partner violence in the u.s. moRRis, D.s.; tEnKKu, l.E.; salas, J.; Et al. Exploring pregnancy-related changes in alcohol general population: the role of alcohol use and socioeconomic status. Trauma, consumption between black and White women. Alcoholism: Clinical and Experimental Violence, & Abuse 5(4):303–317, 2004. PmiD: 15361585 Research 32(3):505–512, 2008. PmiD: 18302726

FloREs, y.n.; yEE, h.F., JR.; lEng, m.; Et al. Risk factors for chronic liver disease in blacks, mosKal, a.; noRat, t.; FERRaRi, P.; anD Riboli, E. alcohol intake and colorectal cancer risk: mexican americans, and Whites in the united states: Results from nhanEs iV, 1999– a dose-response meta-analysis of published cohort studies. International Journal of 2004. American Journal of Gastroenterology 103(9):2231–2238, 2008. PmiD: 18671818 Cancer 120(3):664–671, 2007. PmiD: 17096321

FREibERg, m.s.; Chang, y.-F.; KRaEmER, K.l.; Et al. alcohol consumption, hypertension, and muKamal, K.J.; ChEn, C.m.; Rao, s.R.; anD bREsloW, R.a. alcohol consumption and cardio- total mortality among women. American Journal of Hypertension 22(11):1212–1218, vascular mortality among u.s. adults, 1987 to 2002. Journal of the American College 2009. PmiD: 19730413 of Cardiology 55(13):1328–1335, 2010. PmiD: 20338493

FuChs, F.D.; ChamblEss, l.E.; Folsom, a.R.; Et al. association between alcoholic beverage mulia, n.; yE, y.; gREEnFiElD, t.K.; anD zEmoRE, s.E. Disparities in alcohol-related problems consumption and incidence of coronary heart disease in Whites and blacks: the among White, black, and hispanic americans. Alcoholism: Clinical and Experimental atherosclerosis Risk in Communities study. American Journal of Epidemiology 160(5): Research 33(4):654–662, 2009. PmiD: 19183131 466–474, 2004. PmiD: 15321844 mulia, n.; yE, y.; zEmoRE, s.E.; anD gREEnFiElD, t.K. social disadvantage, stress, and alco- FuChs, F.D.; ChamblEss, l.E.; WhElton, P.K., Et al. alcohol consumption and the incidence hol use among black, hispanic, and White americans: Findings from the 2005 u.s. of hypertension : the atherosclerosis Risk in Communities study. Hypertension 37(5): national alcohol survey. Journal of Studies on Alcohol and Drugs 69(6):824–833, 1242–1250, 2001. PmiD: 113588935 2008. PmiD: 18925340

golDbERg, D.m., anD solEas, g.J. beverage alcohol consumption as a negative risk factor naimi, t.s.; bREWER, R.D.; moKDaD, a.; Et al. binge drinking among us adults. JAMA: for coronary heart disease. in agrawal, D.P., and seitz, h.K., Eds. Alcohol in Health and Journal of the American Medical Association 289(1):70–75, 2003. PmiD: 12503979 Disease. new york: marcel Dekker, 2001. national Cancer institute (nCi). SEER Stat Fact Sheets: Breast: Surveillance gREEnFiElD, J.R.; samaRas, K.; JEnKins, a.b.; Et al. moderate alcohol consumption, estrogen Epidemiology and End Results, 2011a. available at: http://seer.cancer.gov/statfacts/ replacement therapy, and physical activity are associated with increased insulin sensi- html/breast.html. accessed January 3, 2012. tivity: is abdominal adiposity the mediator? Diabetes Care 26(10):2734–2740, 2003. PmiD: 14514572 nCi. SEER Stat Fact Sheets: Colon and Rectum: Surveillance Epidemiology and End Results, 2011b. available at: http://seer.cancer.gov/statfacts/html/colorect.html. hasin, D.s.; stinson, F.s.; ogbuRn, E.; anD gRant, b.F. Prevalence, correlates, disability, accessed January 3, 2012. and comorbidity of Dsm-iV alcohol abuse and dependence in the united states: Results from the national Epidemiologic survey on alcohol and Related Conditions. Archives of nCi. SEER Stat Fact Sheets: Larynx: Surveillance Epidemiology and End Results, 2011c. General Psychiatry 64(7):830–842, 2007. PmiD: 17606817 available at: http://seer.cancer.gov/statfacts/html/laryn.html. accessed January 3, 2012.

hEath, D.b. uses and misuses of the concept of ethnicity in alcohol studies: an essay in nCi. SEER Stat Fact Sheets: Liver and Intrahepatic Bile Duct: surveillance Epidemiology deconstruction. International Journal of the Addictions 25(5a–6a):607–628, and End Results, 2011d. available at: http://seer.cancer.gov/statfacts/html/livibd.html. 1990–1991. PmiD: 2101395 accessed January 3, 2012.

hEbERt, J.R.; huRlEy, t.g.; anD ma, y. the effect of dietary exposures on recurrence and nCi. SEER Stat Fact Sheets: Oral Cavity and Pharynx: Surveillance Epidemiology and mortality in early stage breast cancer. Breast Cancer Research and Treatment End Results, 2011e. available at: http://seer.cancer.gov/statfacts/html/oralcav.html. 51(1):17–28, 1998. PmiD: 9877026 accessed January 3, 2012.

hERon, m.; hoyERt, D.l.; muRPhy, s.l.; Et al. Deaths: Final data for 2006. National Vital national Center for health statistics. Health, United States, 2010: With Special Feature Statistics Reports 57(14):1–134, 2009. PmiD: 19788058 on Death and Dying. hyattsville, mD: Centers for Disease Control and Prevention, 2011.

hilton, J. Race and Ethnicity in Fatal Motor Vehicle Traffic Crashes 1999–2004. national institute of Diabetes and Digestive and Kidney Diseases (niDDK). National Washington, DC: us national highway traffic safety administration, 2006. Diabetes Statistics, 2011, 2011. available at: http://diabetes.niddk.nih.gov/Dm/Pubs/ statistics/#Racial. accessed January 9, 2012. hoWaRD, a.a.; aRnstEn, J.h.; anD gouREVitCh, m.n. Effect of alcohol consumption on dia- betes mellitus: a systematic review. Annals of Internal Medicine 140(3):211–219, national institute on alcohol abuse and alcoholism (niaaa). niaaa Council approves 2004. PmiD: 14757619 definition of binge drinking. NIAAA Newsletter, Winter, 2004, no. 3, page 3. bethesda, international agency for Research on Cancer. Alcohol Drinking. Vol. 44. geneva, mD: niaaa. switzerland: World health organization, 1988. niaaa. Helping Patients Who Drink Too Much: A Clinician’s Guide, 2005. available at:

KERR, W.C.; gREEnFiElD, t.K.; bonD, J.; Et al. Racial and ethnic differences in all-cause http://pubs.niaaa.nih.gov/publications/Practitioner/Cliniciansguide2005/clinicians_guid mortality risk according to alcohol consumption patterns in the national alcohol e.htm. accessed December 19, 2011.

surveys. American Journal of Epidemiology 174(7):769–778, 2011. PmiD: 21856649 niElsEn, a.l.; maRtinEz, R.; anD lEE, m.t. alcohol, ethnicity, and violence: the role of alco-

KERR, W.C.; PattERson, D.; anD gREEnFiElD, t.K. Differences in the measured alcohol con- hol availability for latino and black aggravated assaults and robberies. Sociological tent of drinks between black, White and hispanic men and women in a us national Quarterly 46(3):479–502, 2005. sample. Addiction 104(9):1503–1511, 2009. PmiD: 19438419 oEtzEl, J., anD DuRan, b. intimate partner violence in american indian and/or alaska KRoEnKE, C.h.; Chu, n.F.; RiFai, n.; Et al. a cross-sectional study of alcohol consumption native communities: a social ecological framework of determinants and interventions. patterns and biologic markers of glycemic control among 459 women. Diabetes Care American Indian and Alaska Native Mental Health Research 11(3):49–68, 2004. PmiD: 26(7):1971–1978, 2003. PmiD: 12832298 15536589

lEonaRD, K.E., anD EiDEn, R.D. marital and family processes in the context of alcohol use PEluCChi, C.; gallus, s.; gaRaVEllo, W.; Et al. Cancer risk associated with alcohol and and alcohol disorders. Annual Review of Clinical Psychology 3:285–310, 2007. PmiD: tobacco use: Focus on upper aero-digestive tract and liver. Alcohol Research & Health 17716057 29(3):193–198, 2006. PmiD: 17373408

maCKEnziE, t.; bRooKs, b.; anD o’ConnoR, g. beverage intake, diabetes, and glucose con- PERRy, s.W. American Indians and Crime: A BJS Statistical Profile, 1992–2002. trol of adults in america. Annals of Epidemiology 16(9):688–691, 2006. PmiD: Washington, DC: u.s. Department of Justice, bureau of Justice statistics, 2004. 16458538 available at: http://www.bjs.gov/content/pub/pdf/aic.pdf. accessed December 9, 2011.

miniño, a.m.; muRPhy, s.l.; Xu, J.; anD KoChanEK, K.D. Deaths: Final data for 2008. PlEtChER, m.J.; VaRosy, P.; KiEFE, C.i.; Et al. alcohol consumption, binge drinking, and National Vital Statistics Reports 59(10):1–126, 2011. PmiD: 22808755 early coronary calcification: Findings from the Coronary artery Risk Development in

236 Alcohol Research: Current Reviews young adults (CaRDia) study. American Journal of Epidemiology 161(5):423–433, 2008. available at: http://oas.samhsa.gov/nsDuh/2k7nsDuh/tabs/sect2petabs43to84. 2005. PmiD: 15718478 htm#tab2.46b. accessed october 14, 2010.

REhm, J.; baliunas, D.; boRgEs, g.l.; Et al. the relation between different dimensions of tEnKKu, l.E.; moRRis, D.s.; salas, J.; anD XaVERius P.K. Racial disparities in pregnancy-relat- alcohol consumption and burden of disease: an overview. Addiction 105(5):817–843, ed drinking reduction. Maternal and Child Health Journal 13(5):604–613, 2009. PmiD: 2010. PmiD: 20331573 18780169

RoERECKE, m., anD REhm, J. the cardioprotective association of average alcohol consump- tREntham-DiEtz, a.; nEWComb, P.a.; stoRER, b.E.; anD REmington P.l. Risk factors for carcino- tion and ischaemic heart disease: a systematic review and meta-analysis. Addiction ma in situ of the breast. Cancer Epidemiology, Biomarkers & Prevention 9(7):697–703, 107(7):1246–1260, 2012. PmiD: 22229788 2000. PmiD: 10919740

Russo, D.; PuRohit, V.; FouDin, l.; anD salin, m. Workshop on alcohol use and health VaEth, P.a., anD sataRiano, W.a. alcohol consumption and breast cancer stage at diagnosis. Disparities 2002: a call to arms. Alcohol 32(1):37–43, 2004. PmiD: 15066702 Alcoholism: Clinical and Experimental Research 22(4):928–934, 1998. PmiD: 9660324

samEt, J.h.; hoRton, n.J.; mEli, s.; Et al. alcohol consumption and antiretroviral adher- VilamoVsKa, a.-m.; bRoWn tayloR, D.; anD bluthEnthal, R.n. adverse drinking-related conse- ence among hiV-infected persons with alcohol problems. Alcoholism: Clinical and quences among lower income, racial, and ethnic minority drinkers: Cross-sectional results. Experimental Research 28(4):572–577, 2004. PmiD: 15100608 Alcoholism: Clinical and Experimental Research 33(4):645–653, 2009. PmiD: 19183136

samEt, J.h.; hoRton, n.J.; tRaPhagEn, E.t.; Et al. alcohol consumption and hiV disease Wahab, s., anD olson, l. intimate partner violence and sexual assault in native american progression: are they related? Alcoholism: Clinical and Experimental Research communities. Trauma, Violence, & Abuse 5(4):353–366, 2004. PmiD: 15361588

27(5):862–867, 2003. PmiD: 12766632 WEiss, h.a.; bRinton, l.a.; bRogan, D.; Et al. Epidemiology of in situ and invasive breast sEmPos, C.t.; REhm, J.; Wu, t.; Et al. average volume of alcohol consumption and all- cancer in women aged under 45. British Journal of Cancer 73(10):1298–1305, 1996. cause mortality in african americans: the nhEFs cohort. Alcoholism: Clinical and World health organization (Who). The Global Status Report on Alcohol and Health. Experimental Research 27(1):88–92, 2003. PmiD: 12544011 geneva, switzerland: Who, 2011. singlEtaRy, K.W., anD gaPstuR, s.m. alcohol and breast cancer: Review of epidemiologic yin, s., anD agaRWal, D.P. Functional polymorphism of alcohol and aldehyde dehydroge- and experimental evidence and potential mechanisms. JAMA: Journal of the American nases. in agrawal D.P., and seitz h.K., Eds. Alcohol in Health and Disease. new york: Medical Association 286(17):2143–2151, 2001. PmiD: 11694156 marcel Dekker, 2001. stinson, F.s.; gRant, b.F.; anD DuFouR, m.C. the critical dimension of ethnicity in liver cir- yoon, y., anD yi, h. Surveillance Report #83: Liver Cirrhosis Mortality in the United rhosis mortality statistics. Alcoholism: Clinical and Experimental Research States, 1970–2005. bethesda, mD: niaaa, Division of Epidemiology and Prevention 25(8):1181–1187, 2001. PmiD: 11505049 Research, alcohol Epidemiologic Data system, 2008. available at: http://pubs.niaaa. nih.gov/publications/surveillance83/Cirr05.htm. accessed march 2, 2009. stRangEs, s.; FREuDEnhEim, J.l.; muti, P.; Et al. greater hepatic vulnerability after alcohol intake in african americans compared with Caucasians: a population-based study. yuan, n.P.; Koss, m.P.; PolaCCa, m.; anD golDman, D. Risk factors for physical assault and Journal of the National Medical Association 96(9):1185–1192, 2004. PmiD: 15481746 rape among six native american tribes. Journal of Interpersonal Violence 21(12): 1566–1590, 2006. PmiD: 17065655 substance abuse and mental health services administration (samhsa). Arrests for Driving Under the Influence Among Adult Drivers. Rockville, mD: office of applied zEmoRE, s.E. acculturation and alcohol among latino adults in the united states: a com- studies, samhsa, 2005. available at: http://oas.samhsa.gov/2k5/Duiarrests/Duiarrests.pdf. prehensive review. Alcoholism: Clinical and Experimental Research 31(12):1968– accessed april 3, 2012. 1990, 2007. PmiD: 18034692

samhsa. 2007 National Survey on Drug Use & Health: Detailed tables [table 2.46b: zEmoRE, s.E.; KaRRiKER-JaFFE, K.J.; KEithly, s.; anD mulia, n. Racial prejudice and unfair alcohol use, binge alcohol use, and heavy alcohol use in the past month among per- treatment: interactive effects with poverty and foreign nativity on problem drinking. sons aged 18 or older, by demographic characteristics: percentages, 2006 and 2007], Journal of Studies on Alcohol and Drugs 72(3):361–370, 2011. PmiD: 21513672

Ethnicity and the Social and Health Harms From Drinking 237 selective (i.e., clinical) prevention and treatment approaches Gaps in Clinical for heavy drinkers and people with alcohol use disorders (AUDs) in reducing the burden associated with excessive alcohol use. As used here, selective, or clinical, prevention Prevention and treatment refers to strategies targeted at individuals at higher risk of experiencing adverse alcohol effects, such as screening and brief counseling of heavy drinkers in health care settings or for Alcohol Use Disorders internet-based screening and advice provided to college stu- dents. The term “treatment” refers to services for alcohol dependence provided by a professional, such as a counselor, social worker, nurse, psychologist, or physician. Community Costs, Consequences, and Strategies peer-led support groups such as Alcoholics Anonymous are considered to be distinct from professional treatment ser- vices, much like a diabetes support group would be distin- guished from endocrinology services. The article focuses on the following three questions: (1) Can selective prevention and treatment reduce the disease burden attributable to Mark L. Willenbring, M.D. heavy drinking? (2) Are some treatment approaches more cost-effective than others? (3) Do gaps exist in the current continuum of care? After addressing these issues, the review heavy drinking causes significant morbidity, premature suggests research priorities to help close existing gaps and mortality, and other social and economic burdens on society, reduce the burden of disease. prompting numerous prevention and treatment efforts to avoid or ameliorate the prevalence of heavy drinking and its consequences. however, the impact on public health of current Selective Prevention and treatment: selective (i.e., clinical) prevention and treatment strategies is Effectiveness, Cost-Effectiveness, unclear. screening and brief counseling for at-risk drinkers in ambulatory primary care has the strongest evidence for efficacy, and Disease Burden and some evidence indicates this approach is cost-effective and Screening and brief advice for at-risk (i.e., nondependent) reduces excess morbidity and dysfunction. Widespread drinkers, commonly known as screening and brief interven- implementation of screening and brief counseling of tion (SBI), is effective at reducing drinking for a year or nondependent heavy drinkers outside of the medical context more and in many studies also has been shown to reduce has the potential to have a large public health impact. For alcohol-related harms, such as motor-vehicle crashes and people with functional dependence, no appropriate treatment driving violations. Its efficacy is supported by numerous and prevention approaches currently exist, although such randomized controlled trials and multiple meta-analyses; as a strategies might be able to prevent or reduce the morbidity and result, the U.S. Prevention Task Force has listed it as a Type other harmful consequences associated with the condition B recommendation for medical prevention services (Babor et before its eventual natural resolution. For people with alcohol al. 2007; Whitlock et al. 2004). The evidence is strongest for use disorders, particularly severe and recurrent dependence, nondependent heavy drinkers who present for primary care treatment studies have shown improvement in the short term. services in ambulatory settings. Unfortunately, a recent however, there is no compelling evidence that treatment of meta-analysis of studies of SBI in primary care settings failed alcohol use disorders has resulted in reductions in overall to show significant reductions in subsequent health care disease burden. more research is needed on ways to address utilization (Bray et al. 2011). The efficacy of SBI in other functional alcohol dependence as well as severe and recurrent settings, such as emergency departments (EDs) or hospitals, alcohol dependence. kEY WoRDS: Alcohol use, abuse, and has not been established, although several randomized con- dependence; heavy drinking; alcohol use disorders (AUDs); trolled trials have been conducted (Field et al. 2010). One alcohol-related problems; alcohol burden; burden of disease; explanation for the observed differences may be the patient morbidity; mortality; prevention; treatment; prevention populations analyzed. Thus, in most of the outpatient pri- strategy; treatment strategy; screening and brief intervention; mary care studies, participants with alcohol dependence were primary care; cost-effectiveness of AoD health services excluded from the analysis, whereas that generally was not

Mark L. Willenbring, M.D., former director of Treatment eavy drinking takes a high toll on society. Other articles and Recovery Research at the National Institute on Alcohol in this issue summarize the disease burden and eco- Abuse and Alcoholism, is founder and CEO of Alltyr: Hnomic cost to society attributable to alcohol use, which Transforming Treatment for Addictions, St. Paul, Minnesota. provide a powerful incentive to develop and implement ways to reduce them. The focus of this article is on the role of

238 Alcohol Research: Current Reviews the case for studies conducted in EDs or hospital settings. • Professional treatment only has a small effect in deter- Moreover, patients with alcohol dependence are much more mining outcome compared with other, nontreatment commonly encountered in ED and hospital settings than in factors, such as social control (e.g., driving-while-intoxicated primary ambulatory care. In summary, at this time, SBI in laws, family pressure, or employer mandate), natural his- primary care ambulatory settings for adults can be strongly tory of alcohol dependence, and the tendency to revert to recommended as highly efficacious, whereas SBI in EDs or usual levels of drinking following resolution of a crisis hospitals cannot. where drinking had peaked (i.e., regression to the mean). SBI also seems to be effective among select groups when delivered through internet-based or computerized applica- This last explanation is supported by recent research tions. In particular, there is strong evidence that digital SBI demonstrating that changes in drinking habits begin weeks can effectively reduce drinking and associated consequences before treatment entry (Penberthy et al. 2007). Likewise, among college students (Moreira et al. 2009). It is not clear in another study of treatment of alcohol dependence that whether or to what extent this finding might generalize to examined events leading to treatment seeking (Orford et al. other population subgroups, but it is certainly plausible that 2006), the findings suggested that the change point occurred it could, provided the target population has easy access to prior to treatment entry. Thus, it is unclear how much of the computers and is computer literate. The same holds true for positive change can be attributed to the treatment processes other methods, such as telephone-based SBI or use of the themselves as opposed to other factors leading to and follow- relatively new publication and Web site called Rethinking ing treatment seeking. Drinking, which is published by the National Institute on What is clear, however, is that researchers and clinicians Alcohol Abuse and Alcoholism (NIAAA). do not yet understand how or why some people change in Despite the evidence supporting its effectiveness, SBI is response to treatment and others do not. To address this not yet being implemented widely (Hingson et al. 2012). issue, NIAAA led the way at the National Institutes of Health Widespread dissemination of information about recom- (NIH) in shifting the focus of behavioral treatment research mended drinking limits and easy access to screening and to identifying the mechanisms of behavior change rather brief counseling has the potential to make a significant public than encouraging more comparisons of different psychotherapy health impact. Because at-risk drinkers are much more approaches (Willenbring 2007). The NIH subsequently numerous than alcohol-dependent people, at-risk drinking developed a major initiative on basic behavioral research (Li 2009). This research initiative provides an opportunity to contributes a much greater disease burden than alcohol investigate many obvious questions. For example, what are dependence. Accordingly, widespread implementation of the social forces that either support or impede positive health SBI has the potential to reduce a greater proportion of disease behavior change? What determines their impact, in terms of burden than even very effective treatment, a concept known the response of the individual? Why and how do people as the prevention paradox (Rose 1981). Therefore, more begin to change, and what determines the resilience of that research is needed to expand the implementation of SBI in change? What is the basic science underlying behavior change, the at-risk population and further increase its effectiveness. at all levels from genetic and genomic to cellular, organic, Estimating the effectiveness and cost-effectiveness of individual, and social interactions? Research elucidating the treatment is more complex. Most reviews conclude that treat- basic science of behavior change is an exciting and promising ment is effective at reducing drinking and associated conse- area that has the potential to substantially change the types quences. Multiple behavioral treatment approaches—such as of interventions that are available, making them more pow- cognitive– behavioral therapy, motivational enhancement erful, available, and cost-effective. therapy, 12-step facilitation, behavioral marital therapy, and The lack of clarity about what causes change in drinking community reinforcement—have similar and relatively high behavior also results in uncertainty as to whether treatment levels of short-term success in reducing drinking and associ- of alcohol dependence reduces disease burden. The commu- ated consequences, at least when treatment is provided by nity prevalence of alcohol dependence, which is about 4 percent the highly trained, motivated, and closely supervised clini- in any year, has not changed substantially in recent years cians participating in clinical efficacy trials (Project MATCH (Substance Abuse and Mental Health Services Administration Research Group 1998). Why these technically diverse coun- 2011). Earlier studies found a cost offset of treatment—that seling techniques produce almost identical drinking outcomes is, lower health care costs after treatment than before treatment is unclear. Three alternative explanations have been offered: (Holder 1998). More recent studies, however, have found that heavy drinkers who are not in crisis underutilize health • The specific technique is less important than other, care, at least in an employed population, suggesting that the mostly unidentified, factors associated with psychotherapy. observed cost reduction is more a reflection of the natural history of drinking behavior and of a regression to the mean • Each approach works via different mechanisms but produces (Finney 2008; Zarkin et al. 2004). In other words, people similar results on average, much like different antidepres- suffering from any disease tend to seek treatment when their sants acting through different mechanisms produce similar condition is most severe. In the case of alcohol dependence, outcomes in the treatment of depression. treatment seeking therefore would be preceded by an escala-

Gaps in Clinical Prevention and treatment for Alcohol Use Disorders 239 tion of drinking, complications, and utilization of medical ment in outcomes. In addition, confidential treatment by services and, consequently, high costs before treatment entry. their usual primary care physician involving only routine Because chronic conditions such as alcohol dependence wax clinic visits may attract more people, thus expanding access and wane, most people will tend to improve after a period to effective treatments. of greater severity, even without effective treatment, so that subsequent reduced costs may not necessarily be associated with treatment. Also, every patient’s disease trajectory is dif- Gaps in the Continuum of Care ferent, so that when drinkers are assessed before and after treatment, some of them will be well at followup, whereas There are several gaps in the continuum of care that deserve for others their condition will be more severe. The average attention, affecting drinkers across the spectrum of alcohol severity, however, will be less following treatment, because involvement. Recent epidemiological research has demon- for all patients studied, their disease severity at treatment strated that alcohol involvement varies along a continuum entry will have been high. The most rigorous study of cost- ranging from asymptomatic heavy drinking (i.e., at-risk effectiveness of alcoholism treatment, the COMBINE trial, drinking), through functional alcohol dependence, and to found that treatment was cost-effective, especially pharma- severe and recurrent alcohol dependence (Willenbring et al. cotherapy with medical management (Zarkin et al. 2008, 2009). The continuum of care ideally should correspond to 2010). The interpretation of these findings is limited, however, this epidemiology but does not at this time. Most studies by the study’s highly rigorous trial design, intensive follow and treatment approaches have focused on the more severe up, and exclusion criteria (Anton et al. 2006), and it is end of the spectrum—that is, people with severe, recurrent unknown to what extent these findings generalize to com- dependence. However, the vast majority of heavy drinkers munity treatment programs and participants. either does not have alcohol dependence or has a relatively Another limitation when estimating the effects of treatment milder, self-limiting form (Moss et al. 2007). This spectrum on public health is that relatively few affected people seek of severity is similar to that for other chronic diseases, such treatment. For example, among people who develop alcohol as asthma. Likewise, examining treatment seekers in the dependence at some point in their lives only 12 percent seek current system of care yields similar results to studying treatment in a specialty treatment program (Hasin et al. hospitalized asthmatics: thus, heavy drinkers in treatment 2007). Among people who have AUDs and who perceive a exhibit more severe dependence, more comorbidity, less need for treatment, almost two-thirds (i.e., 65 percent) fail response to treatment, and a less supportive social network to obtain it because they are not ready to stop drinking or compared with people who do not seek intensive treatment feel they can handle it on their own. Other common reasons (Bischof et al. 2003; Dawson et al. 2005; Sobell et al. 2000). for the failure to seek treatment include practical barriers, In contrast, people with functional alcohol dependence2 such as lack of health insurance, the cost of treatment, and predominantly exhibit “internal” symptoms, such as impaired lack of transportation or access to treatment, which are control; a persistent desire to cut down on their drinking but reported by 59 percent of respondents, and stigma, which is finding it hard to do; and alcohol use despite internal symptoms reported by 31 percent (Center for Behavioral Health Statistics such as insomnia, nausea, or hangover. These individuals and Quality 2012).1 Thus, more people might seek treat- generally drink much less than more seriously affected people ment if it was less expensive, stigmatizing, and disruptive (Moss et al. 2007). Functional alcohol dependence typically than most treatment approaches. Efforts to improve access, resolves after a few years, mostly without requiring specialty affordability, and attractiveness of treatment, especially for treatment (Hasin et al. 2007). Large gaps in services exist individuals with less severe AUDs should be encouraged. for people at both ends of the spectrum of dependence Despite these limitations, some tentative conclusions severity—that is, both for people at the milder end of the can be drawn as to which approaches to treating alcohol spectrum (i.e., at-risk drinkers and people with functional dependence are more cost effective. Studies found no significant alcohol dependence) and for those at the most severe end difference in outcomes between residential and outpatient (i.e., with recurrent, treatment-refractory dependence). treatment and no clear relationship between intensity of There currently are few services for at-risk drinkers and treatment and outcome (Fink et al. 1985; Longabaugh et al. people with functional alcohol dependence. In primary 1983; McCrady 1986). For example, medical management medical care, very few patients are screened and positive plus pharmacotherapy with naltrexone generated similar screening results addressed (McGlynn et al. 2003). Furthermore, outcomes to more expensive counseling approaches, even functional alcohol dependence largely is ignored because when counseling was performed once weekly and on an although these individuals meet diagnostic criteria for outpatient basis (Anton et al. 2006; O’Malley et al. 2003). dependence, they rarely seek treatment in the current system These studies suggest that a more individualized, outpatient, (Moss et al. 2007). These gaps are significant from a public and medically based approach may provide a cost-effective health perspective because the prevalence of at-risk drinking alternative to approaches favoring intensive psycho-education, which often are provided in residential settings. Treatment 1 the numbers add up to more than 100 percent because respondents could endorse multiple reasons. provided in residential rather than outpatient settings may 2 People with functional alcohol dependence are those who meet the criteria for a medical diagnosis add considerable expense without a commensurate improve- of alcohol dependence but remain functional in society (i.e., in their jobs, families, and social lives).

240 Alcohol Research: Current Reviews and functional dependence is much higher than that of lapse medications) combined with medical management more severe disorders and these conditions therefore account offers an attractive possible approach for this group, and evi- for the majority of excess morbidity, mortality, and associ- dence suggests that this combination yields comparable ated costs attributable to alcohol consumption (Centers for results to state-of-the-art counseling (Anton et al. 2006; Disease Control and Prevention 2012). Whether wider O’Malley et al. 2003). Such an approach would allow most implementation of SBI would result in a reduction in disease people with functional dependence to be treated in primary burden is not known at this time. However, enhancement care and mental health care settings, similar to people with of these approaches, especially among young people and mild to moderate depression. More research, especially community-dwelling heavy drinkers not seeking medical regarding effectiveness and implementation, is needed on this care, might reduce disease burden, although the two popula- approach. Although most people with functional alcohol tions require somewhat distinct approaches. More studies of dependence eventually recover without any treatment (Hasin secondary prevention efforts outside of medical settings et al. 2007; Moss et al. 2007), their period of illness is associ- therefore are needed. ated with less severe but still significant dysfunction, such SBI in primary care settings to identify people with as absenteeism, attending work or school while sick (i.e., AUDs at the milder end of the severity spectrum is effective presenteeism), and reduced productivity. Early identification and may be cost-effective (Solberg et al. 2008), but many and treatment could reduce or hopefully eliminate these costs questions remain. For example, is it more cost-effective to tar- to the affected individuals and society. get higher-risk groups (e.g., young people) for routine screen- Gaps in treatment also exist for people with severe recur- ing or is universal screening better overall? And when should rent alcohol dependence—the group that most people tend screening occur (e.g., only during annual prevention visits to think of when they think of “alcoholism.” A recent exhaus- or at every new patient visit) and how often should it be tive report examining the current treatment system concluded repeated? However, the biggest problem remains that effective that “Most of those who are providing addiction treatment selective prevention interventions such as SBI are not widely are not medical professionals and are not equipped with the implemented. Although implementation has worked well in knowledge, skills or credentials necessary to provide the full situations where additional grant funds were available, it still range of evidence-based services to address addiction effectively,” is unknown whether physicians will engage in this widely or (p. 3) and that “Addiction treatment facilities and programs how to best facilitate implementation. The Veterans Affairs are not adequately regulated or held accountable for provid- health services system has been the most effective at imple- ing treatment consistent with medical standards and proven menting annual screening, but this system is unique in its treatment practices.” (National Center on Addiction and structure and hierarchical nature. Implementation of such Substance Abuse at Columbia University 2012, pp. 3–4). approaches in private health care organizations is much more The current addiction treatment system first was conceptual- complex and difficult. Therefore, more research is needed on ized in the middle of the last century, as documented by low-cost ways to encourage wider adoption of SBI in primary White (2002), and has changed little since. No other chronic care settings. Additional research should focus on SBI in disease is treated with brief stints in a program with limited other medical settings, especially mental health settings and follow up care. Instead, for other chronic conditions patients medical specialties particularly affected by heavy drinking, are followed closely by physicians and other professionals over such as gastroenterology (with patients with alcohol-related long periods of time, with the goal of minimizing symptoms liver disease, gastritis, and pancreatitis) and otolaryngology and relapses, treating complications, and maximizing func- (with patients with alcohol-related head and neck cancers). tion. In these cases, care is provided indefinitely, often for life. Because so many hospitalized heavy drinkers have Such a longitudinal-care approach also offers considerable dependence, SBI is much less effective in this group (Saitz et promise in treating people with severe recurrent alcohol al. 2007) and its effectiveness with patients in EDs or trauma dependence. Several studies have found a highly significant centers also is unknown. Although some early studies showed positive effect for longitudinal care in people who have one positive results, subsequent research has yielded as many neg- or more medical complications of alcohol dependence ative as positive findings (Field et al. 2010). Current efforts to (Kristenson et al. 1984; Lieber et al. 2003), including two implement SBI in these more acute-care settings therefore are studies that found significant reduction in 2-year mortality premature, and more research is needed to determine if heavy (Willenbring and Olsen 1999; Willenbring et al. 1995). drinkers encountered in such settings require more intensive Some findings also indicate that integrating treatment for services, linkage to ambulatory care services, or both. substance use disorders into that for severe and persistent People with functional alcohol dependence likely require mental illness may be effective at reducing substance use, more than brief counseling, but there is a major gap in although no high-quality randomized controlled trials of research concerning optimal treatment strategies. Currently, this approach have been published (Drake et al. 2006). few, if any, services are available for this group because they Pharmacotherapy for AUDs also may be effective in people fall between at-risk drinkers and those with severe recurrent with severe mental illnesses (Petrakis et al. 2004, 2005, alcohol dependence (who are most likely to enter the current 2006; Salloum et al. 2005). Finally, the ongoing need for specialty treatment system). Pharmacotherapy (e.g., antire- recovery support and maintenance should be addressed.

Gaps in Clinical Prevention and treatment for Alcohol Use Disorders 241 Thus, more research is needed on the best long-term man- DaWson, D.a.; gRant, b.F.; stinson, F.s.; Et al. Recovery from Dsm–iV alcohol depen- agement strategies for recurrent alcohol dependence. dence: united states, 2001–2002. Addiction 100(3):281–292, 2005. PmiD: 15733237 DRaKE, R.E.; mChugo, g.J.; XiE, h.; Et al. ten-year recovery outcomes for clients with co- occurring schizophrenia and substance use disorders. Schizophrenia Bulletin Conclusion 32(3):464–473, 2006. PmiD: 16525088 FiElD, C.a.; baiRD, J.; saitz, R.; Et al. the mixed evidence for brief intervention in emer- At this time no solid conclusions can be drawn as to whether gency departments, trauma care centers, and inpatient hospital settings: What should current approaches to prevention of and treatment for AUDs we do? Alcoholism: Clinical and Experimental Research 34(12):2004–2010, 2010. PmiD: 20860610 reduce the disease burden attributable to heavy drinking, although these strategies have shown positive outcomes in FinK, E.b.; longabaugh, R.; mCCRaDy, b.m.; Et al. Effectiveness of alcoholism treatment in partial versus inpatient settings: twenty-four month outcomes. Addictive Behaviors the short term. SBI for at-risk drinkers in ambulatory primary 10(3):235–248, 1985. PmiD: 3936343 care settings has the strongest evidence for efficacy, and some FinnEy, J.W. Regression to the mean in substance use disorder treatment research. evidence supports its cost-effectiveness and associated reduc- Addiction 103(1):42–52, 2008. PmiD: 17999707 tion in excess morbidity and dysfunction. However, these benefits do not necessarily indicate that health care costs for hasin, D.s.; stinson, F.s.; ogbuRn, E.; anD gRant, b.F. Prevalence, correlates, disability, and comorbidity of Dsm–iV alcohol abuse and dependence in the united states: these patients are reduced. Widespread implementation of Results from the national Epidemiologic survey on alcohol and Related Conditions. SBI for nondependent heavy drinkers outside of the medical Archives of General Psychiatry 64(7):830–842, 2007. PmiD: 17606817

context has the potential to have a large public health impact. hingson, R.W.; hEEREn, t.; EDWaRDs, E.m.; anD saitz, R. young adults at risk for excess For heavy drinkers with more severe conditions (i.e., recur- alcohol consumption are often not asked or counseled about drinking alcohol. Journal rent alcohol dependence), time-limited counseling may of General Internal Medicine 27(2):179–184, 2012. PmiD: 21935733

improve short-term recovery rates, but its long-term impact holDER, h.D. Cost benefits of substance abuse treatment: an overview of results from is less clear. Moreover, recent research findings have not been alcohol and drug abuse. Journal of Mental Health Policy and Economics 1(1):23–29, widely implemented. Scientifically based, medically anchored 1998. PmiD: 11964488 treatment approaches may provide a more attractive and KRistEnson, h.; ohlin, h.; hultEn-nosslin, m.b.; Et al. identification and intervention of cost-effective approach than the current intensive but time- heavy drinking in middle-aged men: Results and follow-up of 24–60 months of long- limited treatment. More research is needed on ways to term study with randomized controls. Alcoholism: Clinical and Experimental Research address functional alcohol dependence as well as severe and 7(2):203–209, 1983. PmiD: 6135365 recurrent alcohol dependence. ■ li, R. NIH Science of Behavior Change. Retrieved 03/29/2012, 2012. liEbER, C.s.; WEiss, D.g.; gRoszmann, R.; Et al. i. Veterans affairs Cooperative study of polyenylphosphatidylcholine in alcoholic liver disease: Effects on drinking behavior by Financial Disclosure nurse/physician teams. Alcoholism: Clinical and Experimental Research 27(11):1757–1764, 2003. PmiD: 14634491

The author declares that he has no competing financial longabaugh, R.; mCCRaDy, b.; FinK, E.; Et al. Cost effectiveness of alcoholism treatment interests. in partial vs inpatient settings. six-month outcomes. Journal of Studies on Alcohol 44(6):1049–1071, 1983. PmiD: 6420619

mCCRaDy, b.; longabaugh, R.; FinK, E.; Et al. Cost effectiveness of alcoholism treatment in References partial hospital versus inpatient settings after brief inpatient treatment: 12-month out- comes. Journal of Consulting and Clinical Psychology 54(5):708–713, 1986. PmiD:

anton, R.F.; o’mallEy, s.s.; CiRaulo, D.a.; Et al. Combined pharmacotherapies and 3095404

behavioral interventions for alcohol dependence: the CombinE study: a randomized mCglynn, E.a.; asCh, s.m.; aDams, J.; Et al. the quality of health care delivered to adults controlled trial. Jama: Journal of the American Medical Association 295(17):2003– in the united states. New England Journal of Medicine 348(26):2635–2645, 2003. 2017, 2006. PmiD: 16670409 PmiD: 12826639

baboR, t.F.; mCREE, b.g.; KassEbaum, P.a.; Et al. screening, brief intervention, and Referral moREiRa, m.; smith, l.a.; anD FoXCRoFt, D. social norms interventions to reduce alcohol to treatment (sbiRt): toward a public health approach to the management of sub- misuse in university or college students. Cochrane Database of Systematic Reviews. stance abuse. Substance Abuse 28(3):7–30, 2007. PmiD: 18077300 8(3):CD006748, 2009. PmiD: 19588402

bisChoF, g.; RumPF, h.J; haPKE, u.; Et al. types of natural recovery from alcohol depen- moss, h.b.; ChEn, C.m.; anD yi, h.y. subtypes of alcohol dependence in a nationally rep- dence: a cluster analytic approach. Addiction 98(12):1737–1746, 2003. PmiD: resentative sample. Drug and Alcohol Dependence 91(2-3):149–158, 2007. PmiD: 14651506 17597309

bRay, J.W.; CoWEll, a.J.; anD hinDE, J.m. a systematic review and meta-analysis of health national Center on addiction and substance abuse at Columbia university. Addiction care utilization outcomes in alcohol screening and brief intervention trials. Medical Medicine: Closing the Gap Between Science and Practice. new york: the national Care 49(3):287–294, 2011. PmiD: 21263359 Center for addiction and substance abuse at Columbia university, 2012.

Center for behavioral health statistics and Quality. Results from the 2011 national o’ mallEy, s.s.; RounsaVillE, b.J.; FaRREn, C.; Et al. initial and maintenance naltrexone survey on Drug use and health: summary of national findings 2012 (hhs Publication treatment for alcohol dependence using primary care vs specialty care: a nested no. sma 12-4713, nsDuh series h-44). Rockville, mD: substance abuse and mental sequence of 3 randomized trials. Archives of Internal Medicine 163(14):1695–1704, health services administration. 2003. PmiD: 12885685

Centers for Disease Control and Prevention(CDC). Vital signs: binge drinking preva- oRFoRD, J.; hoDgson, R.; CoPEllo, a.; Et al. the clients’ perspective on change during lence, frequency, and intensity among adults—united states, 2010. MMWR: Mortality treatment for an alcohol problem: Qualitative analysis of follow-up interviews in the uK and Morbidity Report Weekly 61(1):14–19, 2012. PmiD: 22237031 alcohol treatment trial. Addiction 101(1):60–68, 2006. PmiD: 16393192

242 Alcohol Research: Current Reviews PEnbERthy, J.K.; ait-DaouD, n.; bREton, m.; Et al. Evaluating readiness and treatment seek- substance abuse and mental health services administration (samhsa). Results From ing effects in a pharmacotherapy trial for alcohol dependence. Alcoholism: Clinical and the 2010 national survey on Drug use and health: summary of national Experimental Research 31(9):1538–1544, 2007. PmiD: 17624996 Findings.Rockville, mD: u.s. Department of health and human servicess, samhsa.

PEtRaKis, i.l.; o’ mallEy, s.; RounsaVillE, b; Et al. naltrexone augmentation of neuroleptic WhitloCK, E.P.; PolEn, m.R.; gREEn, C.a. behavioral counseling interventions in primary treatment in alcohol abusing patients with schizophrenia. Psychopharmacology care to reduce risky/harmful alcohol use by adults: a summary of the evidence for the 172(3):291–297, 2004. PmiD: 14634716 u.s. Preventive services task Force. Annals of Internal Medicine 140(7):557– PEtRaKis, i.l.; Poling, J.; lEVinson, C.; Et al. naltrexone and disulfiram in patients with 568+i564, 2004. PmiD: 15068985 alcohol dependence and comorbid post-traumatic stress disorder. Biological Psychiatry 60(7):777–783, 2006. PmiD: 17008146 WillEnbRing, m.l. a broader view of change in drinking behavior. Alcoholism: Clinical and Experimental Research 31(suppl s3):84s–86s, 2007. PmiD: 17880354 PEtRaKis, i.l.; Poling, J.; lEVinson, C.; Et al. naltrexone and disulfiram in patients with alcohol dependence and comorbid psychiatric disorders. Biological Psychiatry WillEnbRing, m.l., anD olson, D.h. a randomized trial of integrated outpatient treatment 57(10):1128–1137, 2005. PmiD: 15866552 for medically ill alcoholic men. Archives of Internal Medicine 159(16):1946–1952, Project matCh Research group. matching alcoholism treatments to client heterogene- 1999. PmiD: 10493326 ity: treatment main effects and matching effects on drinking during treatment. Project matCh Research group. Journal of Studies on Alcohol 59(6):631–639, 1998. PmiD: WillEnbRing, m.l.; massEy, s.h.; anD gaRDnER, m.b. helping patients who drink too much: 9811084 an evidence-based guide for primary care clinicians. American Family Physician 80(1):44–50, 2009. PmiD: 19621845 RosE, g. strategy of prevention: lessions from cardiovascualar disease. British Medical Journal (Clinical Research Edition) 282(6279):1847–1851, 1981. PmiD: 6786649 WillEnbRing, m.l.; olson, D.h.; anD biElinsKi, J. integrated outpatient treatment for medi- cally ill alcoholic men: results from a quasi-experimental study. Journal of Studies on saitz, R.; PalFai, t.P.; ChEng, D.m.; Et al. brief intervention for medical inpatients with unhealthy alcohol use: a randomized, controlled trial. Annals of Internal Medicine Alcohol 56(3):337–343, 1995. PmiD: 7623473 146(3):167–176, 2007. PmiD: 17283347 zaRKin, g.a.; bRay, J.W.; alDRiDgE, a., Et al. the effect of alcohol treatment on social salloum, i.m.; CoRnElius, J.R.; DalEy, D.C.; Et al. Efficacy of valproate maintenance in costs of alcohol dependence: Results from the CombinE study. Medical Care patients with bipolar disorder and alcoholism: a double-blind placebo-controlled study. 48(5):396–401, 2010. PmiD: 20393362 Archives of General Psychiatry 62(1):37–45, 2005. PmiD: 15630071 zaRKin, g.a.; bRay, J.W.; alDRiDgE, a.; Et al. Cost and cost-effectiveness of the CombinE sobEll, l.C.; EllingstaD, t.P.; anD sobEll, m.b. natural recovery from alcohol and drug study in alcohol-dependent patients. Archives of General Psychiatry 65(10): problems: methodological review of the research with suggestions for future directions. 1214–1221, 2008. PmiD: 18838638 Addiction 95(5):749–764, 2000. PmiD: 10885050 solbERg, l.i.; maCiosEK, m.V.; anD EDWaRDs, n.m. Primary care intervention to reduce alco- zaRKin, g.a.; bRay, J.W.; baboR, t.F.; anD higgins-biDDlE, J.C. alcohol drinking patterns hol misuse: Ranking its health impact and cost effectiveness. American Journal of and health care utilization in a managed care organization. Health Services Research Preventive Medicine 34(2):143-152, 2008. PmiD: 18201645 39(3):553–570, 2004. PmiD: 15149478

Gaps in Clinical Prevention and treatment for Alcohol Use Disorders 243 the World Health organization’s Global Monitoring System on Alcohol and Health vladimir Poznyak, M.D., Ph.D.; Alexandra Fleischmann, Ph.D.; Dag Rekve, M.S.C.; Margaret Rylett, M.A.; Jürgen Rehm, Ph.D.; and Gerhard Gmel, Ph.D. vladimir Poznyak, MD., Ph.D., is ith growing awareness of the General Assembly on the Prevention the coordinator of Management of impact of alcohol consumption and Control of Non-Communicable Won global health (Rehm et al. Diseases (NCDs) mandated the Substance Abuse unit at the World 2004; World Health Organization Health Organization, Geneva, development of a global monitoring [WHO] 2002, 2009) the demand framework, including indicators, and Switzerland; Alexandra Fleischmann, for global information on alcohol a set of voluntary global targets for Ph.D., is a technical officer at the consumption and alcohol-attributable the prevention and control of NCDs. Management of Substance Abuse and alcohol-related harm as well as This mandate explicitly mentioned unit at the World Health Organization, related policy responses has increased the harmful use of alcohol as one of Geneva, Switzerland; Dag Rekve, significantly. Public health problems attributable to harmful alcohol con- the four common risk factors for M.S.C., is a technical officer at sumption have become the focus NCDs along with tobacco use, the Management of Substance of several World Health Assembly unhealthy diet, and lack of physical Abuse unit at the World Health resolutions, including one adopted activity (United Nations 2011). This Organization, Geneva, Switzerland; in 2005 that requested the Director- work yielded a set of nine voluntary Margaret Rylett, M.A., is a General of the WHO “to strengthen targets, including at least a 10 per- researcher at the Social and global and regional information cent relative reduction in the harm- systems through further collection ful use of alcohol and a set of 25 Epidemiological Research (SER) and analysis of data on alcohol con- indicators, including the following Department, Centre for Addiction sumption and its health and social possible indicators for monitoring and Mental Health (CAMH), Toronto, consequences, providing technical the harmful use of alcohol as appro- Canada; Jürgen Rehm, Ph.D., support to Member States and priate, within the national context: is director of the Social and promoting research where such data (1) total (recorded and unrecorded) Epidemiological Research are not available” (WHO 2005). alcohol per capita consumption Department at the Centre for Monitoring and surveillance are crucial (among those ages 15 and older) in setting objectives for national alcohol Addiction and Mental Health, chair within a calendar year in liters of plans and in evaluating success (for pure alcohol; (2) age-standardized and professor in the Dalla Lana more details see Rehm and Scafato School of Public Health, University 2011). In recognition of the increas- prevalence of heavy episodic drink- of Toronto, Canada, and section ing demand from WHO Member ing among adolescents and adults; and (3) alcohol-related morbidity head at the Institute for Clinical States for global health information, and mortality among adolescents Psychology and Psychotherapy, the WHO’s 11th General Programme of Work called for monitoring health and adults (WHO 2012). Inclusion Technische Universität Dresden, situations and assessing trends as one of the alcohol target and indicators Dresden, Germany. Gerhard Gmel, of six core functions for the period in the global monitoring framework Ph.D., is a senior scientist at the 2006–2015 (WHO 2006). for NCDs and their risk factors will Swiss Institute for the Prevention In 2010, the World Health increase the demand for high-quality of Alcohol and Drug Problems, Assembly endorsed the Global Strategy global data on alcohol consumption Lausanne, Switzerland. to Reduce the Harmful Use of Alcohol and alcohol-related harm and attention (WHO 2010), which targeted the to the WHO monitoring activities in monitoring and surveillance of this area. harmful alcohol consumption and alcohol-attributable harm as one of 10 areas for action. The Global Strategy History of the WHo Global also identified production and dis- Monitoring System on Alcohol semination of knowledge as one of and Health the key components for global action (WHO 2010). The WHO Program on Substance Most recently, the Political Abuse established the Global Alcohol Declaration of the High-level Database in 1996, creating the world’s Meeting of the United Nations largest single source of information

244 Alcohol Research: Current Reviews on levels and patterns of alcohol con- Database, to continue efforts to WHo Global information sumption, its health consequences, collect, compile, and analyze alcohol System on Alcohol and Health and policy responses in WHO monitoring and surveillance infor- Member States. Prior to this, WHO mation based on comparable data The WHO created the Global monitoring of alcohol and health and agreed definitions (WHO 2007). Information System on Alcohol and activities largely was focused on collecting countries’ alcohol policy and prevention program data (Moser 1974, 1980, 1992). With the estab- Finland lishment of the Global Alcohol SOCIOECONOMIC CONTEXT Database, the WHO Secretariat Total population: 5 261 000 ³Population 15+ years: 83% ³Population in urban areas: 61% ³Income group (World Bank): High income started to implement regular global Data source: United Nations, data range 1990–2006. questionnaire surveys on alcohol RECORDED ADULT (15+) ALCOHOL CONSUMPTION BY TYPE OF ALCOHOLIC ALCOHOL CONSUMPTION and health among the governmental BEVERAGE (IN % OF PURE ALCOHOL), 2005 Population data (refer to the population 15 years and older and are in litres of pure alcohol). officials of WHO Member States 3% RECORDED ADULT (15+) PER CAPITA CONSUMPTION, 1961–2007 OTHER nominated to provide information 20 to WHO in the areas of alcohol con- 15 28% sumption, alcohol-related harm, and SPIRITS 10 46% BEER 5 policy responses. The data collection Litres of pure alcohol tools were developed by WHO staff 23% 0 WINE 1961 1966 1971 1976 1981 1986 1991 1996 2001 2006 in collaboration with external experts. Year „ Beer „ Wine „ Spirits „ Other „ Total The first Global Status Report on Beer includes malt beers. Wine includes wine made from grapes. Spirits include all distilled beverages. Other includes one or several other alcoholic beverages, such as fermented beverages made from sorghum, maize, millet, rice, or cider, fruit wine, fortified wine, etc.  Alcohol was published in 1999 (WHO ENLARGEMENT OF RECORDED ADULT (15+) PER CAPITA CONSUMPTION, 2000–2007 1999), followed by the Global Status Adult (15+) per capita Robust estimate of five-year 15 consumption, average 2003–2005 change in recorded adult Report on Alcohol and Young People (in litres of pure alcohol): (15+) per capita consumption, 10 2001–2005: (WHO 2001). In 2004, the WHO Recorded 9.7 5 ³ INCREASE

Unrecorded 2.8 Litres of pure alcohol produced two global status reports  STABLE 0 Total 12.5 DECREASE 20002001 2002 2003 2004 2005 2006 2007 based on the data collected from WHO European Region 12.2 INCONCLUSIVE Year Member States and other sources during 2002: one on alcohol con- PATTERNS OF DRINKING HEALTH CONSEQUENCES sumption and related harm (WHO ABSTAINERS (15+ years), 2000 MORBIDITY Males Females Total Prevalence estimates (12-month prevalence for 2004): Males Females 2004a) and the second focused on Lifetime abstainers 3.3% 10.5% 7.1% Alcohol use disorders (15+ years) 6.39% 1.17% alcohol policy (WHO 2004b). The Former drinkers 5.8% 3.7% 4.7% Abstainers* 9.1% 14.2% 11.8% ALL CAUSE MORTALITY latest Global Status Report On Alcohol * Persons who did not drink in the past 12 months. Age-standardized deaths rates, 15+ years (per 100,000 population) and Health contained newly devel- 2000 2001 2002 2003 2004 2005 DRINKERS ONLY MFMFMFMFMFMF oped country profiles (see figure) Adult (15+ years) per capita consumption*, total 14.20 Liver cirrhosis 18.1 6.7 19.5 7.0 21.0 7.6 20.4 6.6 26.2 8.5 27.8 10.2 Road traffic Adult (15+ years) per capita consumption*, males 20.55 12.2 5.1 14.7 5.2 14.2 4.5 12.0 4.7 12.1 4.8 12.1 3.6 based on 30 key indicators related accidents (1) Adult (15+ years) per capita consumption*, females 8.70 Data source: WHO Mortality Database, data as reported by countries (1) refer to transport accidents. to alcohol consumption, health Heavy episodic drinkers** (15–85+ years), males, 2000 16.5% consequences, and policy responses Heavy episodic drinkers** (15–85+ years), females, 2000 3.7% ALCOHOL POLICY * (Recorded + unrecorded) in litres of pure alcohol, average 2003–2005. Excise tax on beer / wine / spirits Yes / Yes / Yes (WHO 2011). The reports also pro- ** Had at least 60 grams or more of pure alcohol on at least one occasion weekly. National legal minimum age for off-premise sales of alcoholic beverages (selling) (beer / wine / spirits) 18 / 18 / 20 vided valuable information on levels PATTERNS OF DRINKING SCORE National legal minimum age for on-premise sales of alcoholic and patterns of alcohol consumption Patterns of drinking score* LEAST RISKY 1 2 3 4 5 MOST RISKY beverages (serving) (beer / wine / spirits) 18 / 18 / 18 * Given the same level of consumption, the higher the patterns of drinking score, Restrictions for on-/off-premise sales of alcoholic beverages: at global and regional levels, and the greater the alcohol-attributable burden of disease for the country. Time (hours and days) / location (places and density) Yes / Yes Specific events / intoxicated persons / petrol stations Yes / Yes / No contained estimates of alcohol- National maximum legal blood alcohol concentration (BAC) when attributable disease burden. In 2006, driving a vehicle (general / young / professional), in % 0.05 / 0.05 / 0.05 Legally binding regulations on alcohol advertising / product placement Yes / Yes the WHO Expert Committee on Legally binding regulations on alcohol sponsorship / sales promotion Yes / Yes

Problems Related to Alcohol Consumption recommended the establishment of a global information Figure Example of country profile as presented in the WHO Global Status Report on system on alcohol, based on the Alcohol and Health (Who 2011) (reproduced with permission from the Who). current WHO Global Alcohol

the World Health organization’s Global Monitoring System on Alcohol and Health 245 Health (GISAH) to collect, compile, Since its development, the Project; data from national and analyze, and disseminate global GISAH and its regional components international surveys including ques- information on alcohol and health. have become the central global infor- tions on alcohol consumption and From the very beginning of its devel- mation tool for dynamic presenta- related harm from the WHO STEPS opment by the WHO Department tion of worldwide data on levels and (http://www.who.int/chp/steps/instr of Mental Health and Substance patterns of alcohol consumption, ument/ en/index.html) survey Abuse in collaboration with the alcohol-attributable health and social instrument; and data in the public Centre for Addiction and Mental consequences, and policy responses domain from economic operators in Health (CAMH) in Canada, the at all levels. The WHO Global alcohol production and trade, global information system was con- Strategy to reduce the harmful use of including industry-supported organi- ceived as integrated with the regional alcohol explicitly mentions strength- zations, published scientific articles, information systems on alcohol, ening the GISAH and developing or data from the United Nations (UN) although at that time such a system refining appropriate data-collection Food and Agricultural Organization existed only in the WHO European mechanisms, based on comparable (FAO) and other UN agencies, and region. GISAH now is part of the data and agreed indicators and defi- intergovernmental organizations WHO Global Health Observatory nitions, as the key activity of the such as Organization International and integrates four regional informa- WHO Secretariat in support of de la Vigne et du Vin. The Canadian tion systems from countries in the WHO Member States in producing CAMH conducts passive surveillance Americas, Europe, Southeast Asia, and disseminating knowledge on of the relevant published as well as and Western Pacific regions (http:// alcohol and health (WHO 2010). grey literature. The WHO Secretariat www.who.int/gho/alcohol/en/index. Among the remaining key chal- convenes regular meetings with key html). The GISAH functions as one lenges for improving international data providers on alcohol consumption single data repository, with common comparisons of data on alcohol con- to discuss and triangulate available data collection and data quality– sumption and alcohol-attributable data for achieving better estimates control procedures to prevent discrep- health consequences are the following: when national data are either ancies between the global and regional (1) national monitoring systems on unavailable or incomplete. information systems on alcohol alcohol and health in many countries The WHO Global Survey on and health. are weak, fragmented or lacking; Alcohol and Health, a key data-col- Within GISAH, data are (2) difficulties exist in estimating lection mechanism, is implemented organized under a broad set of seven consumption of informally and in collaboration with WHO regional categories of indicators: levels of illicitly produced alcohol; (3) poor and country offices, the Canadian alcohol consumption; patterns of comparability of indicators used in CAMH and several other academic consumption; harms and conse- different jurisdictions; (4) limited centers and institutions. The WHO quences; economic aspects; alcohol geographical representation of studies Global Survey Instrument on Alcohol control policies; prevention, research, on the association of alcohol con- and Health, developed by WHO and treatment resources; and youth sumption with health outcomes; and in collaboration with all partners and alcohol. (5) a paucity of international multi- involved in the survey, is forwarded GISAH currently encompasses country research projects on alcohol to all WHO Member States through more than 150 alcohol-related indi- epidemiology using common the WHO regional and country cators, with data for more than 225 research protocols. offices for completion by focal points countries and territories and includes and national counterparts explicitly indicators that are comparable across nominated by governments to col- countries. The information on pre- Processes and Procedures laborate with WHO on this activity. vention and treatment resources is Underlying the WHo GiSAH For countries belonging to the presented in another information sys- European Union (EU), the survey tem (i.e., Resources for the Prevention Data sources for the GISAH include is implemented in collaboration and Treatment of Substance Use results of the WHO Global Survey with and support from the European Disorders) (http://www.who.int/ on Alcohol and Health; government Commission. In 2008, the survey gho/substance_abuse/en/index.html) documents and national statistics instrument contained 69 questions , which also is a part of the WHO available in the public domain; data grouped into three sections: (1) alcohol Global Health Observatory. from the Global Burden of Disease policy; (2) alcohol consumption; and

246 Alcohol Research: Current Reviews (3) alcohol-related health indicators. country, regional, and global levels. Secretariat to decide which statistics The questionnaire was developed in Surveys are thought to underestimate to give preference to, depending on English and translated into French, per capita consumption by more than the validity of the data (see Portuguese, Russian, and Spanish. 50 percent (Midanik 1988, 1982; http://who.int/gho/gisah). In 2008, completed questionnaires Rehm et al. 2007) and survey errors Unrecorded consumption obvi- were received from 84 percent of are larger (Shield and Rehm 2012). ously is harder to estimate and moni- WHO Member States, representing The alcohol per capita con- tor at the country level. Only a few 97 percent of the world’s population. sumption indicator is based on the countries have regular monitoring of In 2012, 177 Member States partici- estimates of per capita consumption unrecorded consumption. For all pated in the survey, which repre- of recorded and unrecorded alcohol, others, unrecorded alcohol consump- sented a 90 percent response rate and the latter referring to alcohol that is tion is estimated based on one-time covered 98 percent of the world pop- not taxed and is outside the usual studies and expert opinion. For the ulation. system of governmental control, 2012 Global Survey on Alcohol and In 2012, the survey tool was because it is produced, distributed, Health an additional questionnaire modified to strengthen the alcohol and sold outside formal channels component on unrecorded alcohol policy section in line with the main and, therefore, not registered by rou- consumption has been developed suggested areas for national action tine data collection (Rehm et al. and implemented based on the prin- specified in the WHO Global 2003, 2007). It is critical to include ciples of the Delphi survey method- Strategy to reduce the harmful use of unrecorded consumption in the esti- ology (for a description, see Linstone alcohol. In 2012 the survey was par- mates of overall levels of alcohol and Turoff 1975; Rehm and tially implemented using the Web- exposure in populations, because Gadenne 1990). The questionnaire based data-collection tool. more than one-fourth of global con- in this component covers estimates sumption stems from unrecorded of unrecorded alcohol consumption Alcohol Per Capita Consumption alcohol (WHO 2011). However, in its major categories, such as home One of the most important indica- contrary to some conjectures, production (of spirits, wine, and tors of alcohol consumption in the unrecorded consumption does not beer), alcohol brought over the Global Survey on Alcohol and seem to be linked to more health border (smuggling, duty free, and Health is per capita consumption problems than recorded consump- cross-border shopping), illegal (among those aged 15 and older) tion, if volume and patterns of production (including counterfeit in liters of pure alcohol. Notwith- drinking are controlled for (Rehm alcoholic beverages), and surrogate standing some limitations associated et al. 2010). Recorded consumption alcohol (liquids usually containing with its aggregate-level nature can be measured via sales and taxa- ethanol and industrial spirits not (Bloomfield et al. 2003), alcohol per tion or via production, export, and intended for consumption as bever- capita consumption is a key indica- import. Many national governments ages). The questionnaire also tor for measuring levels of alcohol regularly monitor alcohol per capita addresses perceived importance of exposure in populations (WHO consumption, and reliable data is unrecorded alcohol consumption 2000, 2007). Despite the potential available from a significant number from a public health perspective as measurement bias in unrecorded of countries, though predominantly well as the measures implemented at consumption, per capita consump- high-income. These national statis- the country level to reduce the pub- tion is considered the most reliable tics, if based on validated methodol- lic health impact of illicit and infor- and valid indicator for alcohol con- ogy, are given highest preference in mally produced alcohol in line with sumption in a population (Gmel reporting in GISAH. However, even a set of policy options and interven- and Rehm 2004) and is particularly if data on alcohol consumption are tions listed in the Global Strategy to appropriate for monitoring purposes. unavailable from national statistics, reduce the harmful use of alcohol Population-based survey data are per capita consumption can be esti- (WHO 2010). extremely important for further esti- mated, either via industry data in the Tourist consumption also is mates of alcohol consumption in public domain, or by using data sup- being considered in estimating alcohol different age and gender groups but plied from by the FAO and its statis- per capita consumption in popula- currently cannot be considered as a tical database (FAOSTAT) (http:// tions, where tourist consumption is valid and reliable basis for estimates faostat3.fao.org/home/index.html). significant (because the number of of alcohol per capita consumption at An algorithm is used by the WHO tourists per year is at least the num-

the World Health organization’s Global Monitoring System on Alcohol and Health 247 ber of inhabitants) and is not bal- consequences and related policy evaluation. This system, which anced by drinking by national inhab- responses, while also strengthening includes the global surveys, GISAH, itants abroad during vacations. This national capacity for analyzing and and WHO global status reports on mainly is the case for smaller countries. disseminating the information, also alcohol and health, is the central Alcohol per capita consumption through the WHO’s global and mechanism for monitoring imple- is one example of the approximately regional information systems on mentation of the WHO global strategy 200 indicators monitored via the alcohol and health. To further improve to reduce the harmful use of alcohol GISAH at the country, regional, and comparability of data generated in and report on its implementation global levels. countries, consistent data collection to WHO Member States (WHO mechanisms, agreed indicators and 2013), other constituencies, and the data Validation definitions, and enhanced dissemina- public health community at large. ■ Before releasing the national, tion of data is needed. The new alcohol Note: The views expressed in this regional, and global data on alcohol module in the WHO STEPS instru- article are those of the authors and, consumption, alcohol-related harm, ment (http://www.who.int/chp/ except as specifically noted, do not and policy responses, the WHO steps/instrument/en/index.html), represent the official policies or Secretariat undertakes an intensive which is the main data collection positions of the WHO. process of data validation by compil- tool used in WHO surveillance ing country profiles with all the activities on risk factors for chronic available data for key indicators and diseases, is an attempt to prioritize Financial Disclosure forwarding the country profiles to some well-defined key indicators and improve consistency between the rel- The authors declare that they have each country for validation. At this no competing financial interests. stage, any discrepancies are resolved evant national surveillance activities by considering new data for the peri- and the WHO global monitoring system on alcohol and health. Work ods covered in the survey, further tri- References angulation of available information, continues on WHO tools for alcohol epidemiology and monitoring, and building consensus around dis- bloomFiElD, K.; stoCKWEll, t.; gmEl, g., anD REhn, n. puted qualitative indicators. After such as the International Guide for international comparisons of alcohol consumption. the validation process, the data are Monitoring Alcohol Consumption and Alcohol Research & Health 27:95–109, 2003. PmiDl uploaded in the WHO GISAH and Related Harm (WHO 2000). 15301404 subsequently presented in the WHO One of the challenges for the gmEl, g., anD REhm, J. measuring alcohol consump- global status reports on alcohol and WHO global monitoring system on tion. Contemporary Drug Problems 31:467–540, 2004. alcohol and health continues to be a health. lim, s. s.; Vos, t.; FlaXman, a.D., Et al. a comparative time lag between the alcohol expo- risk assessment of burden of disease and injury sure data collected from countries attributable to 67 risk factors and risk factor clusters Further Developments and their dissemination through in 21 regions, 1990–2010: a systematic analysis for GISAH and WHO global and the global burden of Disease study 2010. Lancet 380:2224–2260, 2012. PmiD: 23245609 Both the depth of the GISAH and regional status reports on alcohol the rigor of data collection and vali- and health. Efforts to reduce this linstonE, h.a., anD tuRoFF, m. The Delphi: Method: Techniques and Applications. Reading, ma: addison- dation make it an indispensable tool time lag will involve data collection Wesley, 1975. for policy development and evalua- through Web-based data collection tools, optimizing data validation and miDaniK, l. Validity of self-reported alcohol use: a liter- tion, as well as for global research ature review and assessment. British Journal of (e.g., the Global Burden of Disease dissemination procedures, as well as Addiction 83:1019–1029, 1988. PmiD: 3066418 2010 estimates were based on WHO strengthening partnerships and miDaniK, l.t. the validity of self-reported alcohol con- global monitoring [Lim et al. 2012; resource mobilization for effective sumption and alcohol problems: a literature review. Shield et al. 2013]). One of the key functioning of the global monitoring British Journal of Addiction 77:357–382, 1982. tasks for the WHO in the area of system. PmiD: 6762224

global monitoring and surveillance The ultimate objective for the mosER, J. Problems and programmes related to alco- on alcohol and health is to support WHO global monitoring system on hol and drug dependence in 33 countries. geneva, the development of effective national alcohol and health is strengthening switzwerland: World health organization, 1974.

systems for monitoring alcohol the link between monitoring activi- mosER, J. Prevention of Alcohol-Related Problems: An consumption, its health and social ties and policy development and International Review of Preventive Measures, Policies

248 Alcohol Research: Current Reviews and Programmes. toronto, Canada: alcoholism and for 2005: a contribution to the Comparative Risk Who. Public Health Problems Caused by Harmful Use Drug addiction Research Foundation, 1980. assessment for the 2010 global burden of Disease of Alcohol. geneva, switzerland: Who, 2005. study. Addiction, 108:912-922, 2013. PmiD: mosER, J. Alcohol Problems, Policies and Programmes 23347092 Who. Engaging for Health: 11th General Programme in Europe. Copenhagen, Denmark: World health of Work, 2006-2015. A Global Health Agenda. organization Regional office for Europe, 1992. united nations . Political declaration of the High-level geneva, switzerland: Who, 2006. available at: http:// Meeting of the General Assembly on the Prevention REhm, J., anD gaDEnnE, V. Intuitive Predictions and whqlibdoc.who.int/publications/2006/gPW_eng.pdf Professional Forecasts: Cognitive Processes and and Control of Non-communicable Diseases. Who. WHO Expert Committee on Problems Related to Social Consequences. oxford: Pergamon Press, (Document a/66/l1) new york: united nations, 2011. 1990. available at: http://www.un.org/ga/search/ Alcohol Consumption. Second Report. Who technical view_doc.asp?symbol=a/66/l.1 Report series (944):1-55; 55-57, 2007. PmiD: REhm, J.; KantEREs, F.; anD laChEnmEiER, D.W. 17970166 unRECoRDED consumption, quality of alcohol and World health organization (Who). Constitution of the health consequences. Drug and Alcohol Review World Health Organization [online]. Who. available Who. Strategies to Reduce the Harmful Use of Alcohol 29:426–436, 2010. PmiD: 20636660 at: http://apps.who.int/gb/bd/PDF/bd47/En/constitu- Sixty-first World Health Assembly: Eighth Plenary tion-en.pdf. (accessed 29 august 2013). Meeting, 2008. geneva, switzerland: Who, 2008. REhm, J.; KlotsChE, J.; anD PatRa, J. Comparative quan- tification of alcohol exposure as risk factor for global Who. Global Status Report on Alcohol. geneva, Who. Global Health Risks. Mortality and Burden of burden of disease. International Journal of Methods switzerland: Who, 1999. available at: http://www. Disease Attributable to Major Risks. geneva, in Psychiatric Research 16:66–76, 2007. PmiD: who.int/substance_abuse/publications/en/globalalco switzerland: Who, 2009. available at: http://www. 17623386 hol_overview.pdf who.int/healthinfo/global_burden_disease/globalhea REhm, J.; REhn, n.; Room, R., Et al. the global distribu- Who. International Guide for Monitoring Alcohol lthRisks_report_full.pdf tion of average volume of alcohol consumption and Consumption and Related Harm. geneva, Who. Global Strategy to Reduce the Harmful Use of patterns of drinking. European Addiction Research switzerland: Who, 2000. available at: http://whqlib- Alcohol. geneva, switzerland: Who, 2010. available 9:147–156, 2003. PmiD: 12970583 doc.who.int/hq/2000/Who_msD_msb_00.4.pdf at: http://www.who.int/substance_abuse/ msbalc- REhm, J.; Room, R.; montEiRo, m., Et al.. alcohol use. Who. global status Report. Alcohol and Young stragegy.pdf in: Ezzati, m.; loPEz, a. D.; RoDgERs, a.; and murray, People. geneva, switzerland: Who, 2001. available C.J.l., Eds. Comparative Quantification of Health at: http://whqlibdoc.who.int/hq/2001/Who_ Who. WHO Global Status Report on Alcohol and Risks: Global and Regional Burden of Disease msD_msb_01.1.pdf Health. geneva, switzerland: Who, 2011. available Attribution to Selected Major Risk Factors. geneva, at: http://libdoc.who.int/publications/2011/ switzerland: World health organization, 2004. Who. The World Health Report 2002: Reducing Risks, 9789241564151_eng.pdf Promoting Healthy Life. geneva, switzerland: Who, REhm, J., anD sCaFato, E. indicators of alcohol con- sumption and attributable harm for monitoring and 2002. available at: http://www.who.int/whr/2002/ Who. Report of the Formal Meeting of Member States surveillance in European union countries. Addiction en/whr02_en.pdf to Conclude the Work on the Comprehensive Global Monitoring Framework, Including Indicators, and a 106(suppl. 1):4–10, 2011. PmiD: 21324016 Who. Global Status Report on Alcohol 2004. geneva, Set of Voluntary Global Targets for the Prevention and shiElD, K., anD REhm, J. Difficulties with telephone- switzerland: Who, 2004a. available at: http://www.who. Control of Noncommunicable Diseases, Nov 5–7, based surveys on alcohol consumption in high- int/substance_abuse/publications/global_status_repo 2012. geneva, switzerland: Who, 2012. income countries: the Canadian example. rt_2004_overview.pdf International Journal of Methods in Psychiatric Who. Global Status Report: Alcohol Policy. geneva, Who. Progress Report by the Secretariat to Sixty-Sixth Research 21:17–28, 2012. PmiD: 22337654 switzerland: Who, 2004b. available at: http://www. World Health Assembly “Global Strategy to Reduce shiElD, K.; RylEtt, m.; gmEl, g., Et al. global alcohol who.int/substance_abuse/publications/en/alcohol%2 the Harmful Use of Alcohol (Resolution WHA63.13).” exposure estimates by country, territory and region 0Policy%20Report.pdf geneva, switzerland: Who, 2013.

the World Health organization’s Global Monitoring System on Alcohol and Health 249 of the design and implementation of clinical trials and the Measuring the Burden— implication of technological advances for research also were considered. Other topics included the links between genetics and other lifestyle factors, such as eating behavior, and the Current and Future relationship between drinking and various chronic diseases. Taken together, these summaries provide unique insight into Research trends the current state of research on alcohol’s role in chronic dis- ease and the direction these investigations may take in the future. (For more information on the epidemiological chal- lenges of elucidating the effects of alcohol consumption and Results From the nIAAA Expert drinking as they relate to the initiation/ exacerbation and treatment of chronic diseases, see the article by Shield and Panel on Alcohol and Chronic colleagues [pp. 155–173]). Panel members also were asked what research they would most strongly support if funds disease Epidemiology were unlimited and how they might scale back that research if funding were limited (see Future Ideas textbox). Highlights from this panel are presented below and specific recommen- dations are listed in the accompanying sidebar.

Rosalind A. Breslow, Ph.D., M.P.H., R.D., Clinical trials and kenneth J. Mukamal, M.D. Clinical studies include clinical nutrition studies, controlled alcohol has a significant impact on health and well-being, from feeding studies, and metabolic studies. This type of research the beneficial aspects of moderate drinking to the detrimental has numerous strengths for studying alcohol and chronic effects of alcoholism. the broad implications of alcohol use on disease, including the ability to control alcohol dose and public health have been addressed through a wide range of diet, collect abundant biologic samples from a variety of tis- epidemiological and clinical studies, many of which are sues, assess cause and effect, and examine mechanisms—all described in this issue of Alcohol Research: Current Reviews. with a relatively small number of participants enrolled for a short period of time. Where chronic disease is involved, alcohol use can be a risk Clinical study end points typically are surrogate markers factor that not only affects the onset of various chronic diseases for chronic diseases because the disease itself may take years but also exacerbates the ongoing extent and severity of those or even decades to develop. For example, lipoproteins and diseases. lifestyle choices and genetic influences also markers of inflammation have been used as surrogates for contribute to, or help to alleviate, that risk. kEY WoRDS: NiAAA cardiovascular disease, insulin sensitivity for diabetes, and Expert Panel on Alcohol and Chronic Disease Epidemiology; DNA damage for cancer. alcohol consumption; alcohol burden; chronic disease; risk According to Dr. David J. Baer, considerable need for factors; epidemiology; research; diabetes; cardiovascular controlled clinical studies on alcohol and chronic disease still disease; cancer; stroke; liver disease; genetic factors; eating exists. There have been few clinical studies, even on cardio- behaviors; clinical trials vascular disease (Brien et al. 2011), which is the focus of most alcohol-related chronic disease research. He also noted the relatively few controlled clinical studies of alcohol and obesity (Sayon-Orea et al. 2011) that were advocated by the esearch is continuing to investigate how alcohol impacts chronic disease. The National Institute on Alcohol RAbuse and Alcoholism (NIAAA) hosted a 2-day Expert Rosalind A. Breslow, Ph.D., M.P.H., R.D., is an epidemi- Panel on Alcohol and Chronic Disease Epidemiology in August 2011 to review the state of the field on alcohol and ologist, Division of Epidemiology and Prevention Research, chronic disease. The panel was chaired by Kenneth J. Mukamal, National Institute on Alcohol Abuse and Alcoholism, M.D., and Rosalind A. Breslow, Ph.D., M.P.H., R.D., and National Institutes of Health, Bethesda, Maryland. was convened by NIAAA’s Division of Epidemiology and Prevention Research. kenneth J. Mukamal, M.D., is associate professor of Panel members (see textbox) represented a wide range of medicine, Harvard Medical School, Division of General backgrounds and expertise, ranging from alcohol-related chronic Medicine & Primary Care, Beth Israel Deaconess Medical diseases and risk factors to methods and technology. Among Center, Boston, Massachusetts. the chronic diseases addressed were diabetes, cardiovascular disease, cancer, stroke, and liver disease. The broader aspects

250 Alcohol Research: Current Reviews Report of the Dietary Guidelines Advisory Committee on Although the exact mechanisms involved in these cardio- the Dietary Guidelines for Americans (U.S. Department of protective effects still are under investigation, the putative Agriculture 2010). benefits on cardiovascular disease likely are the result of alcohol’s Dr. Baer suggested the following future opportunities effects on lipids and insulin sensitivity (Dijousse et al. 2009). for alcohol and chronic disease research: In his presentation, Dr. Kenneth J. Mukamal noted that standard epidemiologic studies of alcohol consumption and • Drinking patterns; coronary heart disease incidence or mortality are no longer useful, as virtually all prospective studies performed since • Effects on metabolism and disease risk; 1980 have shown that moderate drinking reduces risk (Corrao et al. 2000; Mukamal et al. 2010; Ronksley et al. • Non-ethanol components of alcoholic beverages; 2011). Recent analytic strategies have resulted in more pre- cise statistical estimates, but the conclusion is unchanged. In • Possible effects on cardiovascular disease, diabetes (insulin essence, he stated, “We’ve been doing the same epidemiology sensitivity), cancer, and bone metabolism; since 1992.” Dr. Mukamal suggested the following future opportuni- • Gender and age differences (pre- and postmenopausal ties for alcohol and cardiovascular disease research: women, men); • Effects of heavy and binge drinking; • Genetic basis for response of chronic disease surrogate markers to alcohol; • Effects of changes in alcohol consumption over time;

• Energy metabolism, body weight regulation, and insulin • Differences in effect of gender-specific drinking patterns; sensitivity; • Genetic interactions; • Interaction of alcohol with lower-fat or higher-protein diets; and • Studies of new mechanisms directly related to alcohol’s effects (for example, cholesterol efflux capacity) (Khera et • Bone metabolism. al. 2011); • Pooling projects for questions that require large samples; and

Cardiovascular Disease • Use of case crossover designs to account for both triggering events and chronic use (Mostofsky 2011). Studies on alcohol and cardiovascular disease have yielded important findings with regard to public health. For example, we now know that the association of alcohol use within rec- ommended limits with lower risk of heart disease depends Cancer more on the frequency with which alcohol is consumed and Alcohol consumption increases the risk for several cancers, not on the type (Cleophas 1999). Wine, beer, and spirits all including breast, colon, liver, and upper aero-digestive cancers have been associated with reduced risk of myocardial infarc- (oral, pharynx, larynx, and esophagus) (Schutze et al. 2011; tion. Modest differences in the effects of those different types World Cancer Research Fund 2007). The potential mecha- of alcohol are thought to be more a result of lifestyle differ- nisms underlying alcohol’s effects include the carcinogenicity ences among drinkers rather than a direct link to a specific of acetaldehyde (for colorectal cancer and upper aero-digestive type of alcohol. How often people drink alcohol has a larger tract cancers), which is an intermediate product of alcohol impact on cardiovascular disease. Among men, drinking metabolism; impairment of the one-carbon nutrient more frequently seems to have a greater impact than the metabolism (for colorectal cancer); alteration of hormone actual amount consumed (Mukamal et al. 2003); effects are levels (for breast cancer); and oxidative stress resulting from less clear among women. The beneficial effects of alcohol alcohol metabolism. also have been shown to be similar for people with existing Dr. Edward Giovannucci noted the paucity of research cardiovascular disease or diabetes (Costanzo et al. 2010; on drinking patterns and cancer. He acknowledged too that Koppes et al. 2006) and those in the general population. studies can yield disparate findings, describing a study that In addition to its beneficial effects on coronary heart disease, initially showed no relationship between average alcohol moderate drinking has been found to reduce the risk of consumption and prostate cancer but which in a posteriori ischemic stroke but at a lesser magnitude and with lower lev- analyses hinted at a possible relationship with high-quantity/ els of consumption (Klatsky et al. 2001). low-frequency drinking (Platz et al. 2004).

Current and Future Research trends 251 In identifying areas for future research, Dr. Giovannucci genes involved in the one-carbon metabolism pathway) are discussed the importance of studying cancer–nutrient interac- other areas that warrant additional study. Determining the tions, particularly for colon cancer. For example, the epidemi- molecular characteristics of tumors, such as tumor subtypes ologic literature has consistently shown an interaction classified by level of methylation, which might reflect defects between alcohol and folate, a nutrient that seems to be in one-carbon metabolism (Schernhammer et al. 2010), is protective at higher levels of drinking (Ferrari et al. 2007; another area that requires further investigation. In addition, Jiang et al. 2003). This suggests that the excess risk of cancer resulting from alcohol use potentially could be modified by little research has been conducted with cancer survivors, a a nutrient or combination of nutrients. group that may be especially willing to modify their drink- Further study also is needed to better understand the ing habits. role of genetics and family history in cancer risk. The genes Finally, as noted by Dr. Giovannucci, alcohol increases involved in alcohol metabolism (Yokoyama et al. 2001) the risk for many cancers, but not all. Recent studies have and nutrient metabolism (for example, the gene methylenete- found that alcohol is associated with a lower risk of kidney trahydrofolate reductase [MTHFR] for folate as well as other cancer (Lee et al. 2007) and non-Hodgkins lymphoma

Future Research ideas, Large and Small, for Consideration

n addition to the full panel discussions, panelists were fat and fibroscan for fibrosis. Such a cohort could asked to consider directions for future studies—both additionally fold in cardiovascular disease risk factors ilarge and small. Specifically, the panelists described and clinical and subclinical cardiovascular disease. what studies they would suggest for future research and Among other things, this study would help to how they would refine those visions when funds are lim- address the simultaneous associations of alcohol con- ited. Selected noteworthy examples are described below. sumption with lower risk of cardiovascular disease but higher risk of fatty liver, which is associated with • A randomized trial to evaluate alcohol consumption a higher risk for cardiovascular disease. Although of and risk of multiple clinical outcomes with sufficient more limited utility, a cross-sectional study with the power to evaluate prespecified genetic environmental interactions would be ideal. However, with limited same measures would also be of clear import. resources, it might be more realistic to use a hybrid design, with a prospective cohort study and a smaller • Studies to verify estimates of drinking patterns. This nested trial. For example, a trial might evaluate if rec- is particularly important as self-reported estimates ommending moderate alcohol consumption, versus form the basis for epidemiological studies but have no recommendation, had an effect on cardiovascular yet to be validated, particularly in the context of eat- and stroke outcomes among patients with a high risk ing patterns, portion sizes, and health beliefs. for vascular problems. • Studies of how alcohol ingestion impacts energy • Clinical trials to establish the effects of alcohol balance in both moderate and binge drinkers. consumption on clinical cardiovascular and cancer outcomes. A large-scale trial using high-risk popula- • Studies to better understand the risk factors underly- tions with standardized exposure to alcohol would be ing alcohol-related chronic disease. These factors ideal. A more practical approach would be to con- range from fixed characteristics, such as genetics and duct shorter trials with subclinical measures of both ethnic background, to broader modifiable behaviors, cardiovascular disease and, to a lesser degree, cancer, such as diet, exercise, or smoking. An ideal study using such techniques as serial computed tomogra- would be multifaceted and include both disease- phy angiography and colonography. specific and composite global endpoints, such as healthy aging or survival free of chronic disease. A • Studies to identify factors that influence the risk for more limited study could simply compile data from liver disease among moderate drinkers. A large, the dozens of cohort studies worldwide where much prospective study would be ideal and would include of this data already have been collected. A more serial measures of genomic, dietary, anthropometric, comprehensive effort would use ongoing studies and behavioral risk factors obtained as objectively as prospectively to incorporate novel measures of drink- possible, coupled with serial noninvasive measures of ing patterns, biomarkers of health status, or greater liver disease using magnetic resonance imaging for assessment of quality of life and mental health.

252 Alcohol Research: Current Reviews (Kroll et al. 2012). Understanding how these two cancers • Pool data from large cohort studies with genetic informa- differ from others is another area requiring additional research. tion on alcohol metabolizing and diabetes-related genes Dr. Giovannucci suggested the following future oppor- to examine the interactions between alcohol, genetic tunities for alcohol and cancer research: predisposition, and diabetes risk.

• Effects of drinking patterns on cancer risk; • Conduct metabolic studies specifically within subgroups to examine how alcohol modifies risk based on lifestyle • Nutrient interactions; characteristics, such as body mass index, diet, and physi- cal activity. • Genetic susceptibility (genes related to alcohol metabolism, genes related to one-carbon metabolism); Stroke and Cognition • Tumor subtypes; Several important findings on the effects of alcohol con- • Cancer survivors; and sumption on the incidence of stroke have emerged from the Northern Manhattan Study, a prospective, multiethnic • Pathways that might explain the limited protective cohort study (Elkind et al. 2006; Sacco et al. 1999). In that aspects of alcohol consumption. study, subjects with the lowest risk for ischemic stroke con- sumed, on average, two drinks per day. Those effects were similar among drinkers of wine, beer, and liquor. In contrast, no protective effect was found for hemorrhagic stroke. Diabetes The study’s principal investigator, Dr. Ralph Sacco, presented the results of two meta-analyses. One found the Evidence that alcohol can impact diabetes has been consis- greatest protection against all strokes combined was most tent over several studies. Results from the Nurses’ Health evident at a lower level of drinking, less than or equal to Study (Stampfer et al. 1988), the Health Professionals one drink per day (Ronksley et al. 2011). Other analyses Follow-up Study (Conigrave et al. 2001), a systematic review compared results from ischemic with hemorrhagic strokes (Howard et al. 2004), and two meta-analyses (Baliunas et al. (Reynolds et al. 2003). For ischemic stroke, moderate drink- 2009; Koppes et al. 2005) all show that moderate drinking is ing was protective, whereas heavy drinking was associated associated with a lower risk of diabetes. Heavy drinking, on with an increased risk; for hemorrhagic stroke, heavy drink- the other hand, seems to lead to an increased risk of diabetes, ing increased risk (although sample size was insufficient to although sample sizes generally have been too small to draw study the effects of moderate drinking on hemorrhagic stroke). firm conclusions. The heterogeneity of strokes underscores the impor- Dr. Eric Rimm described specific areas of research that tance of studying stroke subtypes. Both ischemic strokes (the warrant further study. For example, only about 30 to 50 percent majority of all strokes) and hemorrhagic strokes (about 17 percent of all strokes) have subtypes with differing etiologies of alcohol’s beneficial effects on diabetes can be linked to that may respond differently to alcohol consumption. Little biomarkers studied to date. In addition to its overall effect research has been conducted on these subtypes, partly on insulin sensitivity (Davies et al. 2002), moderate alcohol because of the small numbers of each that occur within most consumption improves adiponectin, a fat-tissue hormone studies and the need for relatively large samples to obtain associated with insulin sensitivity; inflammatory status sufficiently precise estimates of risk. Numerous subclinical (Joosten et al. 2008); and HDL cholesterol. With regard markers of stroke, such as endothelial function, currently are to metabolic studies, he noted the value of using short-term being pursued by researchers (Suzuki et al. 2009). feeding studies because they provide an opportunity to control and simultaneously examine drinking (for example, Cognition with meals or without) and diet (for example, high versus The prevalence of cognitive impairment is growing rapidly low glycemic load) (Mekary et al. 2011). He also discussed as the population ages, and, like stroke, cognitive impair- the importance of studying genetic predisposition (Beulens ment is not a single disease or condition. Studies of alcohol et al. 2007). use and cognition have examined a variety of outcomes, In addition to these areas, Dr. Rimm suggested several including Alzheimer’s disease, cognitive function, dementia, future opportunities for alcohol and type 2 diabetes research: and mild cognitive impairment (Lee et al. 2010). Studies and meta-analyses generally show that moderate drinking is • Pool large cohort studies to maximize power to look at associated with a decreased risk of dementia (Mukamal et al. subpopulations where alcohol may be most detrimental 2003b; Peters et al. 2008), Alzheimer’s disease (Peters et al. or most beneficial. 2008), vascular dementia (Peters et al. 2008), and cognitive

Current and Future Research trends 253 Recommendations for Strengthening Studies

n addition to offering ideas for future studies, the rics, dietary intake/nutritional status, smoking status, Expert Panel also made recommendations for strength- physical activity/fitness, cancer survivorship, and age iening research in the field. Specific suggestions include: of drinking onset. Pooled data also may facilitate studies of rare or understudied outcomes such as 1. Standardize alcohol consumption measurement in liver disease. prospective and retrospective studies of alcohol and chronic disease to the greatest degree possible. a. Standardized alcohol questions should be used Standardized measures: where possible.

a. Should include consumption quantity, frequency, b. Confounding and interaction should be consid- and binge drinking (i.e., basic drinking patterns). ered to ensure robust estimates and define suscep- tible subgroups. b. Should consider drinking over the lifespan (for example, during youth, middle age, menopause, c. Targeted sub-studies within large cohorts should and during time of heaviest drinking) as the criti- be considered as a cost-efficient way to better cal time periods for effects of alcohol on chronic understand and explain results in the full cohort. disease development are uncertain. For example, when data on alcohol consumption are not gathered in enough detail in the original c. Are available from NIAAA and from the study, targeted follow-up studies may be used among NIH/National Human Genome Research Institute stratified subsets of subjects to collect biological Phenx Toolkit: http://www.niaaa.nih.gov/Resources/ samples and to obtain more detailed data on con- Research Resources/Pages/TaskForce.aspx; https://www.phenx.org/Default.aspx?tabid=36 sumption for extrapolating to the parent study.

2. Strongly encourage collection of biological material 7. Include associations between alcohol dependence/ and broad consent for genetic studies in all clinical abuse and chronic disease outcomes. Studies using trials and in as many population studies as possible. pooled data or sub-studies within large cohorts may have the power to address these drinking problems. Data on period of maximum drinking could be 3. Objectively validate standardized alcohol measures important, particularly given the marked variation using novel technologies as they become available. in alcohol intake during the lifespan. Examples may include implantable biosensors and point-of-care devices with wireless transmission capability. 8. Perform studies in understudied areas, including but not limited to the effects of alcohol on diabetes, 4. Develop new biomarkers for moderate alcohol con- obesity, cognition, healthy aging, and food intake. sumption to complement those used for heavy drinking. 9. Focus on relationships between drinking patterns 5. Identify surrogate markers for chronic disease and chronic disease. Drinking patterns include but (including measures of subclinical disease) that will are not limited to basic patterns such as usual quan- have utility in small-scale studies and for elucidating tity, frequency, and binge drinking as well as when, mechanisms and pathways linking alcohol to chronic where, and with whom alcohol was consumed and disease. whether it was consumed with a meal.

6. Pool data from existing cohort studies to facilitate 10. Encourage clinical trials across the spectrum of examination by population subgroups, including chronic disease from studies that examine key physi- but not limited to age, lifespan phase, race/ethnicity, ological parameters and intermediate studies such as menopausal status, body mass index/anthropomet- feeding studies that examine surrogates or subclinical

254 Alcohol Research: Current Reviews phenotypes to practical trials that examine chronic 13. Encourage studies on moderate drinking patterns disease outcomes. and metabolism ranging from total energy and macronutrient metabolism to specific metabolic a. Physiologic studies are preferred when epidemio- pathways for small molecules such as vitamins, logic evidence is relatively limited. amino acids, sugars, and steroids and their products b. Practical trials are preferred when there is extensive and precursors. evidence from physiological and epidemiological studies. 14. Examine the effectiveness of communication mes- sages about drinking. Studies may include, but are 11. Encourage studies examining the interactions not limited to, how to disseminate cost-benefit mes- between the genetics that predispose individuals sages, individualized messages based on patient to drink and the genetics that modify how alcohol demographic and clinical history, and guidance for affects chronic disease. health care professionals on how to advise patients.

12. Encourage studies of carefully defined homogeneous phenotypes. For example, studies are needed to 15. Encourage the use of natural experiments to exam- clarify the effects of alcohol on thrombotic versus ine whether policy interventions or alcohol inter- embolic ischemic stroke, Alzheimer’s disease versus vention studies might change the relationship other dementias, specific eye diseases, etc. between alcohol and chronic disease. decline (Peters et al. 2008). According to Dr. Sacco, there ation, however, there is a lack of strong clinical measures to currently is great interest in vascular risk factors for demen- describe and predict the progression of chronic liver disease. tia, yet little alcohol research has been done in that area. Dr. James Everhart noted that the course of alcoholic liver Other future opportunities for research into alcohol and disease is several decades in duration and begins as simple chronic neurological disease noted by Dr. Sacco include the steatosis (fatty liver) before progressing to more advanced following: stages including steatohepatitis, alcoholic cirrhosis, and, eventually, liver failure. • Cohort studies with careful end point adjudication to Dr. Everhart noted that alcoholic liver disease may be separate ischemic stroke subtypes and different etiologies overrepresented in terms of mortality because of the current of dementia and cognitive impairment; classification system. Histologically, alcoholic fatty liver and nonalcoholic fatty liver look similar (Scaglioni et al. 2011), • Examination of interactions with race and ethnicity and and patients with otherwise similar multiple risk factors and other neurological risk factors; histology may be classified as having alcoholic liver disease rather than nonalcoholic steatohepatitis simply because they • Comparison of associations across beverage types for do or do not drink. According to Dr. Everhart, the current neurological outcomes; and strict separation of alcoholic and nonalcoholic fatty liver disease limits epidemiology, public health, and clinical • Understanding protective alcohol mechanisms including understanding. inflammatory relationships, subclinical measures and In examining the effects of drinking amounts on liver biomarkers, and gene–environment interactions. disease, little association has been found between moderate drinking and alcoholic liver disease, and only a minority of very heavy drinkers develops alcoholic liver disease, although the reason is not clear. It is possible that drinking patterns Chronic Liver Disease and diet each play a role in risk. More information also is needed to determine if drinking at times other than during Chronic liver disease has long been associated with alcohol meals could increase risk. consumption and includes alcoholic liver disease, hepatitis Other factors that put people at higher risk for liver dis- C, and nonalcoholic steatohepatitis. Despite this clear associ- ease include being obese, using cannabis, having diabetes,

Current and Future Research trends 255 and being female (Hart et al. 2010). Conversely, coffee con- studies, where appropriate. Existing studies also could be sumption seems to lower risk and smoking seems to have no enhanced through targeted ancillary studies in which key effect on the development of chronic liver disease. Genetic subsets of subjects are re-contacted to provide more detailed susceptibility is another important risk factor for liver dis- or standardized information. The payoffs from such steps ease. For example, a variant in one gene, PNPLA3, originally could lead to the discovery of key genes and pathways that associated with fatty liver, has been strongly associated with reveal mechanisms and potential targets for therapy. Even if alcoholic liver disease. Again, additional research is needed the effect of a variant is small, the pathway it leads to could to determine how these factors influence alcohol’s effects. be of major importance. Dr. Everhart suggested several future opportunities for Dr. Edenberg suggested several future opportunities for alcohol and chronic liver disease research: the genetics of alcohol and chronic disease research:

• Improve the current chronic liver disease classification • Design and incorporate more detailed alcohol exposure scheme; measures that include patterns of consumption and drinking problems; • Develop reliable and accurate measures of progressive liver disease that can be applied serially; • Search out ongoing and planned studies to;

• Implement better measures of alcohol consumption and – Partner to incorporate exposure measures as early as its patterns to study drinking patterns and interactions possible; between drinking and diet; – Target follow-up and additional studies to gather more • Evaluate how genetics may influence the link between detailed exposure information and genetic samples; and alcohol consumption and the risk of liver disease; and – Encourage collection of samples with consent for • Identify determinants of chronic liver disease among genetic studies. heavy drinkers.

Eating Behaviors Genetics The link between alcohol intake and eating behaviors is not Chronic diseases tend to run in families yet do not follow a well known. Studies generally show that alcohol calories, simple genetic pattern; that is, they are complex and poly- when added to the diet, increase total energy intake (Yeomans genic. Identifying the genes that affect chronic disease risk 2010). Yet despite the fact that alcohol is an energy source, can be hampered by multiple factors, including phenotypic is largely uncompensated (i.e., supplements rather than complexity, multiple genes with small effects, environmental replaces other calories), may weaken feeding controls, and variability, gene–gene interactions, and gene–environment spares fat for storage, little evidence exists that moderate interactions. Alcohol’s role in chronic disease likely reflects a drinking is associated with increased body mass index or gene–environment interaction in which risk is influenced by weight gain (Liangpunsakul 2010; Liu et al. 1994; Wang et genes, by lifestyle choices, and by a combination of both. In al. 2010) (although a French study did show such an effect addition, as noted by Dr. Howard J. Edenberg, most of the [Lukasiewicz et al. 2005]). On the other hand, certain drink- variations in genes related to alcohol and chronic disease ing patterns, particularly binge drinking, have been associ- likely have only small effects, making those genetic influ- ated with higher body mass index (Arif and Rohrer 2005; ences especially difficult to identify. Breslow and Smothers 2005), although impulsivity related One way of overcoming these difficulties, as proposed to both eating and drinking could be an alternative explana- by Dr. Edenberg, is to obtain large sample sizes by combining tion. According to Dr. Richard Mattes, determining alcohol’s data from multiple epidemiologic studies. This enables effects on eating behaviors is further confounded by beverage investigators to examine gene–environmental associations consumption itself and the fact that energy compensation using secondary data analyses. The drawback is that studies for fluids is less than for semisolid or solid foods (Mattes typically ask different questions about alcohol use and often 1996; Mourao et al. 2007). include different time frames, often collect no data on drink- He also suggested that what people think they are eating ing problems, and may not obtain appropriate consent for may be more important in terms of appetitive sensations genetic testing. Dr. Edenberg suggested a number of strategies than its true energy value, noting current research showing to manage these obstacles. For example, investigators could that manipulating food form (liquid or solid) can alter a be encouraged to incorporate standardized alcohol consump- person’s expectation of how filling that food will be. tion questions, particularly for patterns of consumption, and Dr. Mattes suggested several research opportunities for to obtain DNA samples using proper consent for genetic future studies on ingestive behavior and alcohol-related

256 Alcohol Research: Current Reviews chronic disease research, particularly in controlled experi- raphy and single-photon emission computed tomography). mental designs: Magnetic resonance spectroscopy can image relative chemi- cal composition. MRI diffusion tensor imaging can image • Clarify the role of moderate alcohol consumption on white matter tracts (connectivity), and functional MRI can energy balance; image relative blood flow, a marker of neural activity. These structural and functional neuroimaging methods currently • Assess which properties of alcohol contribute to hunger are being used in alcohol research (Buhler and Mann 2011). and satiety; Dr. Haller noted that informatics (data modeling, simulation, and analysis) also will have a significant role in making sense • Ascertain the true biological energy value of alcohol; of the large amounts of high-dimension data now available. Dr. Haller had the following suggestions regarding alcohol- • Test the role of drinking patterns on energy balance; and related chronic disease research:

• Determine the effects of different levels of alcohol con- • Among the variety of technologies and medical devices sumption on body composition and energy balance. that exist for the study of individuals and populations, those of particular interest might include sensors, POC diagnostic devices, imaging technologies, and bioinfor- technology matics tools; A number of promising technologies and medical devices • A better alternative to the “hammer-in-search-of-a-nail” currently are under development by the National Institute approach in imaging is to define the clinical problem of of Biomedical Imaging and Bioengineering and others that interest first, then find the appropriate tools to address may enhance alcohol-related chronic disease research in the the problem or chronic disease under study; future. Dr. John Haller reviewed the research on three areas: sensors, point-of-care (POC) diagnostic devices, and imag- • Alcohol and chronic disease epidemiology could be ing technologies and bioinformatics tools. improved through the use of new sensors (including Sensors are used to detect and quantitate clinically rele- POC diagnostics, sensors embedded in the home or vant analytes. Examples include BioMEMs, microfluidics implanted in the body) to enhance alcohol measure- (Chin et al. 2011), and nanoscale technologies, including ment and by techniques that can image physiological micro-total analysis systems, arrays, and biochips. These mul- function early in the course of chronic disease; and tifunctional devices can measure multiple analytes across a variety of diseases using a platform the size of a credit card. • Technological advances will inevitably produce vast Such technologies then can be combined into POC tests, amounts of data about individuals and populations, but which are defined as diagnostic testing at or near the site of they require new informatics tools that enable meaningful patient care (rather than at centralized laboratories). Benefits use of the data in wide varieties of research settings. include earlier diagnosis of disease and the ability to monitor patients at home. For example, POC tests for alcohol include a breath test and saliva-testing devices (http://www.aacc. Summary org/events/online_progs/documents/AlcoholTesting1.2.pdf); This NIAAA workshop provided an excellent forum for SpectRx, a wristwatch-type device; and Giner, a WrisTas trans- summarizing the current state of the field and for identifying dermal sensor for measuring alcohol consumption (Marques future research opportunities. Although by no means and McKnight 2009). Dr. Haller also reviewed implantable exhaustive, the ideas provided here highlight areas in need monitors and a tattoo using nanosensors that reside under the of additional study and offer a roadmap for moving forward skin. By shining a light on the tattoo the subject enables track- across a variety of methodological approaches and content ing of sodium and glucose levels by portable digital devices, areas. NIAAA would like to thank all of the presenters for including smartphones. In the future, such a technology could their insight and for taking the time to participate in this be used to track alcohol unique workshop. Our hope is that the ideas presented here consumption. will stimulate additional research and further advance our Biomedical imaging of the brain is another area where understanding of the role of alcohol in chronic disease. ■ advances could be applied to the study of alcohol and chronic disease. Most radiology images (e.g., magnetic reso- nance imaging [MRI], computerized tomography) show Additional Resources anatomy/morphology. These images generally capture the late stages of chronic disease. An alternative approach would The agenda, roster of speakers, and speaker’s abstracts can be be to examine the physiological function (e.g., neurorecep- obtained from the author. A copy of the meeting transcript tors) using nuclear imaging (e.g., positron emission tomog- also is available from the author, upon request.

Current and Future Research trends 257 Financial Disclosure Jiang, R.; hu, F.b.; gioVannuCCi, E.l.; Et al. Joint association of alcohol and folate intake with risk of major chronic disease in women. American Journal of Epidemiology The authors declare that they have no competing financial 158:760–771, 2003. PmiD: 14561666 interests. JoostEn, m.m.; bEulEns, J.W.; KERstEn, s.; anD hEnDRiKs, h.F. moderate alcohol consump- tion increases insulin sensitivity and aDiPoQ expression in postmenopausal women: a randomised, crossover trial. Diabetologia 51:1375–1381, 2008. PmiD: 18504547 References KhERa, a.V.; CuChEl, m.; llERa-moya, m.; Et al. Cholesterol efflux capacity, high-density lipoprotein function, and atherosclerosis. New England Journal of Medicine 364:127–135, 2011. PmiD: 21226578 aRiF, a.a., anD RohRER, J.E. Patterns of alcohol drinking and its association with obesity: Data from the third national health and nutrition Examination survey, 1988–1994. KlatsKy, a.l.; aRmstRong, m.a.; FRiEDman, g.D.; anD siDnEy, s. alcohol drinking and risk of BMC Public Health 5:126, 2005. PmiD: 163297757 hospitalization for ischemic stroke. American Journal of Cardiology 88:703–706, 2001. PmiD: 11564405 baliunas, D.o.; tayloR, b.J.; iRVing, h.; Et al. alcohol as a risk factor for type 2 diabetes: a systematic review and meta-analysis. Diabetes Care 32:2123–2132, 2009. PmiD: KoPPEs, l.l.; DEKKER, J.m.; hEnDRiKs, h.F.; Et al. meta-analysis of the relationship between 19875607 alcohol consumption and coronary heart disease and mortality in type 2 diabetic patients. Diabetologia 49:648–652, 2006. PmiD: 16463045 bEulEns, J.W.; Rimm, E.b.; hEnDRiKs, h.F.; Et al. alcohol consumption and type 2 diabetes: influence of genetic variation in alcohol dehydrogenase. Diabetes 56:2388–2394, KoPPEs, l.l.; DEKKER, J.m.; hEnDRiKs, h.F.; Et al. moderate alcohol consumption lowers the 2007. PmiD: 17563066 risk of type 2 diabetes: a meta-analysis of prospective observational studies. Diabetes Care 28:719–725, 2005. PmiD: 15735217 bREsloW, R.a.,anD smothERs, b.a. Drinking patterns and body mass index in never smok- ers: national health interview survey, 1997-2001. American Journal of Epidemiology KRoll, m.E.; muRPhy, F.; PiRiE, K.; Et al. alcohol drinking, tobacco smoking and subtypes 161:368–376, 2005. PmiD: 15692081 of haematological malignancy in the uK million Women study. British Journal of Cancer 107:879–887, 2012. PmiD: 22878373 bRiEn, s.E.; RonKslEy, P.E.; tuRnER, b.J.; Et al. Effect of alcohol consumption on biological lEE, J.E.; huntER, D.J.; sPiEgElman, D.; Et al. alcohol intake and renal cell cancer in a markers associated with risk of coronary heart disease: systematic review and meta- pooled analysis of 12 prospective studies. Journal of the National Cancer Institute analysis of interventional studies. BMJ 342:d636, 2011. PmiD: 21343206 99:801–810, 2007. PmiD: 17505075 buhlER, m., anD mann, K. alcohol and the human brain: a systematic review of different lEE, y.; baCK, J.h.; Kim, J.; Et al. systematic review of health behavioral risks and cogni- neuroimaging methods. Alcoholism: Clinical and Experimental Research tive health in older adults. International Psychogeriatrics 22:174–187, 2010. PmiD: 35:1771–1793, 2011. PmiD: 21777260 19883522 Chin, C.D.; laKsanasoPin, t.; ChEung, y.K.; Et al. microfluidics-based diagnostics of infec- liangPunsaKul, s. Relationship between alcohol intake and dietary pattern: Findings from tious diseases in the developing world. Nature Medicine 17:1015–1019, 2011. PmiD: nhanEs iii. World Journal of Gastroenterology 16:4055–4060, 2010. PmiD: 20731019 21804541 liu, s.; sERDula, m.K.; Williamson, D.F.; Et al. a prospective study of alcohol intake and ClEoPhas, t.J. Wine, beer and spirits and the risk of myocardial infarction: a systematic change in body weight among us adults. American Journal of Epidemiology review. Biomedicine & Pharmacotherapy 53:417-423. PmiD: 10554677 140:912–920, 1994. PmiD: 7977278

ConigRaVE, K.m.; hu b.F.; CamaRgo C.a., JR.; Et al. a prospective study of drinking patterns luKasiEWiCz, E.; mEnnEn, l.i.; bERtRais, s.; Et al. alcohol intake in relation to body mass in relation to risk of type 2 diabetes among men. Diabetes 50:2390–2395, 2001. index and waist-to-hip ratio: the importance of type of alcoholic beverage. Public PmiD: 1157424 Health Nutrition 8:315–320, 2005. PmiD: 15918929

CoRRao, g.; Rubbiati, l.; bagnaRDi, V.; Et al. alcohol and coronary heart disease: a meta- maRQuEs, P.R. anD mCKnight, a.s.; Field and laboratory alcohol detection with 2 types of analysis. Addiction 95:1505–1523, 2000. PmiD: 11070527 transdermal devices. Alcoholism: Clinical and Experimental Research 33:703–711, 2009. PmiD: 19170663 Costanzo, s.; Di, CastElnuoVo, a.; Donati, m.b.; Et al. alcohol consumption and mortality in patients with cardiovascular disease: a meta-analysis. Journal of the American mattEs, R.D. Dietary compensation by humans for supplemental energy provided as College of Cardiology 55:1339–1347, 2010. PmiD: 20338495 ethanol or carbohydrate in fluids. Physiology & Behavior 59:179–187, 1996. PmiD: 8848479 DaViEs, m.J.; baER, D.J.; JuDD, J.t.; Et al. Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a mEKaRy, R.a.; Rimm, E.b.; gioVannuCCi, E.; Et al. Joint association of glycemic load and randomized controlled trial. JAMA 287:2559–2562, 2002. PmiD: 12020337 alcohol intake with type 2 diabetes incidence in women. American Journal of Clinical Nutrition 94:1525–1532, 2011. PmiD: 22049168 DJoussE, l.; lEE, i.m.; buRing, J.E.; anD gaziano, J.m. alcohol consumption and risk of car- mostoFsKy, E.; laiER, E.; lEVitan, E.b.; Et al. Physical activity and onset of acute ischemic diovascular disease and death in women: Potential mediating mechanisms. Circulation stroke: the stroke onset study. American Journal of Epidemiology 173:330–336, 2011. 120:237–244, 2009. PmiD: 19597054 PmiD: 21159732 ElKinD, m.s.; sCiaCCa, R.; boDEn-albala, b.; Et al. moderate alcohol consumption reduces mouRao, D.m.; bREssan, J.; CamPbEll, W.W.; anD mattEs, R.D. Effects of food form on risk of ischemic stroke: the northern manhattan study. Stroke 37:13-19, 2006. PmiD: appetite and energy intake in lean and obese young adults. International Journal of 16306464 Obesity (London) 31:1688–1695, 2007. PmiD: 17579632 FERRaRi, P.; JEnab, m.; noRat, t.; Et al. lifetime and baseline alcohol intake and risk of muKamal, K.J.; ChEn, C.m.; Rao, s.R.; anD bREsloW, R.a. alcohol consumption and cardio- colon and rectal cancers in the European prospective investigation into cancer and vascular mortality among u.s. adults, 1987 to 2002. Journal of the American College nutrition (EPiC). International Journal of Cancer 121:2065–2072, 2007. PmiD: 17640039 of Cardiology 55:1328–1335, 2010. PmiD: 20338493

haRt, C.l.; moRRison, D.s.; batty, g.D.; Et al. Effect of body mass index and alcohol con- muKamal, K.J.; ConigRaVE, K.m.; mittlEman, m.a.; Et al. Roles of drinking pattern and type sumption on liver disease: analysis of data from two prospective cohort studies. BMJ of alcohol consumed in coronary heart disease in men. New England Journal of 340:c1240, 2010. PmiD: 20223873 Medicine 348:109–118, 2003a. PmiD: 12519921

hoWaRD, a.a.; aRnstEn, J.h.; anD gouREVitCh, m.n. Effect of alcohol consumption on dia- muKamal, K.J.; KullER, l.h.; FitzPatRiCK, a.l.; Et al. Prospective study of alcohol consump- betes mellitus: a systematic review. Annals of Internal Medicine 140:211–219, 2004. tion and risk of dementia in older adults. JAMA 289:1405–1413, 2003b. PmiD: PmiD: 14757619 12636463

258 Alcohol Research: Current Reviews PEtERs, R.; PEtERs, J.; WaRnER, J.; Et al. alcohol, dementia and cognitive decline in the sChutzE, m.; boEing, h.; PisChon, t.; Et al. alcohol attributable burden of incidence of elderly: a systematic review. Age and Ageing 37:505–512, 2008. PmiD: 18487267 cancer in eight European countries based on results from prospective cohort study. BMJ 342:d1584, 2011. PmiD: 21474525 Platz, E.a.; lEitzmann, m.F.; Rimm, E.b.; Et al. alcohol intake, drinking patterns, and risk of stamPFER, m.J.; ColDitz, g.a.; WillEtt, W.C.; Et al. a prospective study of moderate alco- prostate cancer in a large prospective cohort study. American Journal of Epidemiology hol drinking and risk of diabetes in women. American Journal of Epidemiology 159:444–453, 2004. PmiD: 14977640 128:549–558, 1988. PmiD: 3414660

REynolDs, K.; lEWis, b.; nolEn, J.D.; Et al. alcohol consumption and risk of stroke: a meta- suzuKi, K.; ElKinD, m.s.; boDEn-albala, b.; Et al. moderate alcohol consumption is associ- analysis. JAMA 289:579–588, 2003. PmiD: 12578491 ated with better endothelial function: a cross sectional study. BMC Cardiovascular Disorders 9:8, 2009. PmiD: 19228434 RonKslEy, P.E.; bRiEn, s.E.; tuRnER, b.J.; Et al. association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ u.s. Department of agriculture. Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans, 2010. us Department of agriculture, agricultural 342:d671, 2011. PmiD: 21343207 Research service, 2010. saCCo, R.l.; ElKinD, m.; boDEn-albala, b.; Et al. the protective effect of moderate alcohol Wang, l.; lEE, i.m.; manson, J.E.; Et al. alcohol consumption, weight gain, and risk of consumption on ischemic stroke. JAMA 281:53–60, 1999. PmiD: 9892451 becoming overweight in middle-aged and older women. Archives of Internal Medicine 170:453–461, 2010. PmiD: 20212182 sayon-oREa, C.; maRtinEz-gonzalEz, m.a.; anD bEs-RastRollo, m. alcohol consumption and body weight: a systematic review. Nutrition Reviews 69:419–431, 2011. PmiD: World Cancer Research Fund. Food, Nutrition, Physical Activity, and the Prevention of 212790610 Cancer: A Global Perspective. Washington, D.C.: american institute for Cancer Research, 2007. sCaglioni, F.; CiCCia, s.; maRino, m.; Et al. ash and nash. Digestive Diseases yEomans, m.R. alcohol, appetite and energy balance: is alcohol intake a risk factor for 29:202–210, 2011. PmiD: 21734385 obesity? Physiology & Behavior 100:82–89, 2010. PmiD: 20096714 sChERnhammER, E.s.; gioVannuCCi, E.; KaWasaKi, t.; Et al. Dietary folate, alcohol and b vita- yoKoyama, a.; muRamatsu, t.; omoRi, t.; Et al. alcohol and aldehyde dehydrogenase gene mins in relation to linE-1 hypomethylation in colon cancer. Gut 59:794–799, 2010. polymorphisms and oropharyngolaryngeal, esophageal and stomach cancers in PmiD: 19828464 Japanese alcoholics. Carcinogenesis 22:433–439, 2001. PmiD: 11238183

Current and Future Research trends 259