TRACI L. TOOMEY, PHD * GUDRUN R. KILIAN JOHN P. GEHAN, MPA MA * CHERYL L. PERRY, PHD RHONDA JONES-WEBB, DRPH ALEXANDER C. WAGENAAR, PHD Qualitative Assessment of Training Programs for Servers and Establishment Managers

S Y N O P S I S

Objective. In an attempt to reduce the societal burden associated with alcohol-related problems such as underage drinking and , some local communities and state governments mandate training programs for employees of establishments that serve or sell alcoholic beverages. This study was designed to assess the available training programs for employees and managers and to identify states that either mandate training programs or encourage them by reducing establishments' legal liability. Methods. Training programs were identified through the Internet, key infor- mants, and the research literature. Three researchers independently rated across I0 In the All the authors are with the Division of each of 22 local and national programs categories. addition, authors surveyed alcoholic beverage control agencies and legislative Epidemiology, School of Public Health, research bureaus in the 50 U.S. states. University of Minnesota. Dr. Toomey is a Research Associate, Ms. Kilian is a Com- Results. The results show that training programs are not standardized and munity Program Specialist, Mr. Gehan is vary widely in content, use of behavior change methods, and production the Project Manager of Project ARM, Dr. quality. Most programs targeted waitstaff and . Only one program owners and managers. Jones-Webb is a Professor of Epidemiol- exclusively targeted ogy, and Dr. Wagenaar is a Professor of Conclusions. National standards must be developed for training programs Epidemiology. for servers, managers, and owners.

162 PU6B2LPIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME I 13 Violence and injuries are among the leading Simply having bartenders and waitstaff watch a 45- causes of death in the United States.' Alco- minute training videotape at the start of employment is hol is implicated in one-third to one-half of not adequate. The programs that have been studied have these deaths, which means that some demonstrated success in increasing the frequency with 48,000 deaths per year are due to alcohol- which servers and sellers undertake relatively simple and related injury.2 Because alcohol consumption is influ- straightforward interventions-offering food, soft drinks, enced by a myriad of behavioral, intra- and interpersonal, and water, checking age identification, and making com- and environmental factors that are interactive and ments about quantity consumed or driving-but not dynamic,3 4 changes in people and in the environment are more difficult ones such as cutting off service of alcohol needed to decrease drinking rates and problems related to an intoxicated patron.6'8 to alcohol use. Altering server behaviors that result in refusal of alco- Server and manager training programs are designed to hol sales is clearly a challenge and may require a more teach alcohol establishment staff how to sell alcohol in intensive, theory-based curriculum than that needed to ways that avoid sales to underage people and reduce the reinforce the server behaviors that require less judgment, intoxication levels of patrons. Alcohol servers and sellers such as offering food. can help support a culture in which high risk drinking is not acceptable. Responsible alcohol service may involve COMPONENTS OF A SUCCESSFUL preventive strategies-limiting the number of alcoholic P R O G R A M beverages served per customer per hour, promoting the sale of food and non-alcoholic beverages-as well as To increase the likelihood of servers refusing to sell alco- proactive strategies-preventing alcohol-impaired hol to underage and intoxicated patrons, programs must patrons from driving and refusing to sell alcohol to obvi- first be designed for the proper audience. Then, regard- ously intoxicated and underage patrons (both types of less of the target audience, quality programs should cover sales are illegal in all states). relevant content areas, use behavioral change techniques to increase the likelihood of behavior change, and use EVALUATING TRAINING appropriate communication methods. P R O G R A M S Target audience. According to behavior change theory Server training programs may vary greatly by establish- as well as previous studies of server training pro- ment, by community and by state. Some establishments grams,3'4"2 it is essential that managers and owners of voluntarily participate in training programs, while others alcohol establishments support and model responsible are mandated to do so by local or state governmental alcohol service. Even the best server training programs authorities. Some programs are offered in groups, either will not be effective without management support. Bar- in the establishment or in a local classroom, and are led tenders and waitstaff will be unlikely to serve alcohol by a professional trainer. In other cases, managers may responsibly if owners and managers do not support and conduct training with a staff person one-on-one, show a reinforce their training or, even worse, if management video, or provide reading material for employees to look at threatens to discipline staff members if they do not serve on their own. underage or intoxicated customers. Studies of the effects of server training have shown Studies have shown that training of managers can that training: (a) increases servers' knowledge and lead to changes in policies in alcohol establishments.6'7"0 improves attitudes toward responsible beverage ser- So in addition to training for alcohol servers and sellers, vice;5-9 (b) may increase servers promotion of food and training is needed to help owners and managers identify, non-alcoholic beverages;'0 (c) may reduce the risk of implement, and enforce effective alcohol policies for patron intoxication;54" 1'2 and (d) may decrease alcohol- their establishments. The alcohol establishment's policies related traffic crashes.'3 and the behaviors of the owner and manager(s) must These studies have produced inconsistent findings- be consistent with the messages provided in server in part due to weak research designs but perhaps also due training 15,16 to the poor quality of the programs being evaluated. Some studies did not use control groups,9"'0 and some looked at Content. Effective training programs provide informa- only a small number of establishments.5-7""l'2 tion about why changing a given behavior is important yet

PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME I 13 163 "Alcohol servers and sellers can help support a culture in which high risk drinking is not acceptable."

do not overwhelm an audience with so many details that to determine whether further development is warranted in the audience becomes bored. Key factors that need to be program design, standards, or implementation. In addi- addressed in server training programs include perceptions tion, it is important to identify government jurisdictions of norms, the policies and practices of the establishment, that mandate server training programs to determine the and understanding of the social support for specific types extent to which server training programs are being man- of preventive and proactive interventions.4"14 5 Emphasis dated across the country. must be put on refusing service to underage and intoxi- cated patrons as a norm. In addition, managers, owners, M E T H 0 D S and servers need to understand the legal and social con- text in which they serve alcohol. What are the state and For the present study, we (a) qualitatively evaluated exist- local laws regulating alcohol sales? What are the social ing server and manager training programs and (b) identi- consequences of underage and heavy alcohol use in our fied states that legislatively mandate or encourage society? statewide server training.

Behavioral change methods. In addition to knowledge Review of training packages. We conducted a qualita- of the laws and policies regarding alcohol sales, both man- tive review of server training packages commonly used agers and servers need specific techniques that they can throughout the United States. The authors identified 24 use to refuse illegal sales. Managers and servers need to server training programs, including state and national pro- increase their skills and confidence in refusing service to grams, by searching the Internet, contacting key infor- underage and intoxicated patrons even in emotional situa- mants in the field, and reviewing the relevant research lit- tions such as when a patron is belligerent or when the erature. Because a centralized list of manager and server underage customer is close to the server's own age. Break- training programs does not exist, there may be other local ing down the desired behavior into smaller components and state programs that the authors did not identify. Some and having the servers repeat and practice those compo- programs consisted of either a videotape or written mater- nents-for example, through role playing-leads to ial, some others of a videotape and a trainers' manual. Still greater skill development and self-efficacy.3 7 others included a videotape, a trainers' manual, and writ- ten material for trainees. Trainers' manuals were intended Communication methods. Quality server training pro- for owners and managers training their own staffs or for grams can take lessons from interpersonal communication professional trainers. theory. A training program needs to capture and maintain Eight of the programs were developed by alcohol participants' attention.'8 A training program is not likely to industry groups, six by nonprofit or research groups, seven have any substantive effect if participants believe that the by for-profit training companies, and one by a state alco- program has little relevance to them, is unrealistic, or is hol beverage control agency. boring. An evaluation of one server training program9 We limited our review to programs that specifically showed that some servers found previous programs to be targeted establishments that sell alcohol. We excluded uninteresting. Also, consistent with communication theo- two programs that addressed responsible beverage service ries, the major points of the server training should be but focused on the role of the community and law repeated many times through multiple channels such as enforcement rather than on alcohol establishments. We videos, lectures, discussions, and in booster sessions at a categorized the remaining 22 programs according to the later time.'9 intended audience: (a) bartenders and waitstaff at estab- It is useful to examine the quality of existing programs lishments where alcoholic beverages are sold for on-site

164 PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME 1 13 consumption; (b) owners and managers of establishments We scored each program component (videotape, train- (such as restaurants and bars) where alcoholic beverages ers' manual, managers' manual, and employee materials) are sold for on-site consumption; (c) both owners/man- on a 4-point scale (0=none, 1 =weak, 2=moderate, agers and bartenders/waitstaff at establishments where 3=strong) for each of the 10 categories, based on how well alcoholic beverages are sold for on-site consumption; (d) they addressed each of the 10 categories. Scores for each clerks at establishments (such as , grocery, and con- component represent the means of ratings by three of the venience stores) where alcoholic beverages are sold for co-authors, who rated each program independently. off-site consumption; and (e) servers and other workers Scores were highly consistent across the three raters. who interact with customers at stadiums and arenas Each program component also received an overall score, where alcoholic beverages are sold (including security representing the mean of the scores for each of the 10 cat- staff, parking lot attendants, and ticket takers, among egories, with each category receiving equal weighting. others). Based on the research literature on server training pro- State policies. To identify those states that have grams8"2 and behavior change theory,3" 4"7 we identified statewide server training policies, we surveyed the 50 10 categories in which to rate server training programs. states' alcohol beverage control agencies and legislative research bureaus. We mailed a 69-item questionnaire to Content each beverage control agency, and a longer, 117-item * Legal issues. The legal framework for selling to minors questionnaire to each legislative research bureau. The sur- or intoxicated people (including criminal penalties vey response rate was 80% for the alcohol beverage con- and "dram shop" or civil liability), which identifies trol agencies and 58% for the legislative research bureaus. consequences for servers, managers, and owners. Overall, we received completed questionnaires from * Physiological effects. Brief overview of how alcohol either one or both agencies in 46 (92%) states. High affects the body; skills to recognize cues of intoxication. response rates were ensured through multiple mailings * Policy development. Skills to develop and enforce and follow-up phone calls. establishment policies to increase responsible service Respondents from both types of agencies were asked of alcohol. about changes in their states' alcohol control laws since * Social problems. Health problems associated with 1967. For the present analysis, we looked at answers to the alcohol, including traffic crashes, pedestrian injuries, following question: "Since 1967, has your state enacted drownings, teen pregnancies, sexually transmitted legislation that encourages alcohol establishments to train diseases, suicide, and violence. alcohol servers?" If staff members responded yes, they were then asked whether server training was: (a) mandated Behavior change elements by state law or (b) encouraged as a means by which estab- * Preventing intoxication. Skills to identify obviously lishments could reduce their legal liability for sales to intoxicated patrons; skills to prevent intoxication. intoxicated customers that result in injuries to third par- * Preventing underage drinking. Skills to identify ties. In addition to the surveys, we conducted legal minors; strategies for refusing sales to underage research in each of the 50 states to verify the accuracy of youth. the completed surveys and to identify the training policies * Behavior change methods. Methods to build skill level of states from which we did not receive completed surveys. and self-efficacy, such as peer leadership, role play- ing, and group discussion, instead of a simple lecture RESULTS format. Training programs. The ratings of the 22 programs are Communication methods shown in the Table. * Realism. Use of real life settings, relevant scenarios, and credible role models. Target audience. Existing training packages varied in terms * Respectful ofaudience. Use of a tone that respects the of intended audiences. Some programs had only one audience; use of attention-getting techniques such as intended audience, while others had multiple compo- humor, montage, and musical soundtracks. nents, each directed at a different audience. Eight of the * Production quality. Overall packaging and production 22 programs targeted only bartenders and waitstaff. Nine values; ease of use; layout and editing. included at least one component that addressed owners or

PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME I 13 165 TO::E:0000000-fi;T M EifX0: : ETE ft0 it't 0AL:TfLf4f0ff 0000.X00:000-0fff'0 :S0000000.000 4t?f0:-':0 000

166 PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME I 13 :A::lgS::;E N00STSS;03F IXf;0:(31 ORN CT.WDV;UWUV00 ::70 :':S

PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME I 13 167 managers of alcohol establishments. Only one of these could be improved by incorporating realistic scenarios, programs (consisting of a videotape) exclusively targeted such as a busy and loud bar scene, or by being more owners and managers. Four of the programs had compo- respectful of the audience, not talking down to them. nents that dealt with either off-site alcohol sales or alcohol Thirteen of the 22 programs had at least one component sales at stadiums and arenas. that had poor production quality. Content. Some content areas were more comprehensively Overall scores. Overall scores ranged widely across compo- covered than others. Only six of the programs had at least nents and programs, from 0.83 to 2.87. Only two pro- one component that were rated as "moderately" to grams (numbers 16 and 17) received an overall rating of "strongly" covering legal issues related to alcohol sales. 2.5 or better. However, both of these programs targeted Most programs discussed alcohol-related social problems only bartenders and waitstaff. only in the context of traffic crashes, if they covered social problems at all. Six programs covered several relevant State laws on server training. As of 1994, 13 states had social issues, such as traffic crashes and alcohol-related enacted some form of legislation on alcohol server training. violence. Twelve programs addressed underage youth Eight states (Alaska, Delaware, Maryland, New Mexico, issues, such as ID checking and underage sales, at least Oregon, Utah, Vermont, and Wisconsin) require servers to moderately well, according to the raters. Two programs receive training. Florida, Maine, New Hampshire, and exclusively focused on preventing alcohol sales to under- Texas have legislation that encourages but does not man- age youth and not on preventing sales to intoxicated date server training. Instead, in these states alcohol estab- patrons. Fifteen programs were found to cover issues lishments' potential liability for negligent alcohol sales is related to intoxicated patrons moderately well or better. either diminished or abolished if the establishments pro- Eight of the programs had at least one component that vide server training for their employees. Michigan requires appropriately covered physiological issues, providing nei- insurance companies to give discounts to establishments ther too much nor too little information. Twelve programs that mandate employee server training. Eight states have a were at least moderately successful in covering the impor- legislatively mandated system for reviewing and certifying tance of establishment policies. Three of these programs, server training programs; however, in our review of state however, did not target owners and managers and are laws, we found that the level of monitoring and the certifi- therefore unlikely to result in policy changes that would cation process vary across these states. support responsible alcohol service. DISCUSSION Behavior change methods. Few of the programs used behavior change techniques. Programs consisting of Although several states mandate server training and others audiovisual tapes alone obviously could not use discussion reduce establishments' liability if employees are trained, a groups or role playing. Some videotapes incorporated standardized server training curriculum has not yet been behavior change techniques by using peer leaders-alco- developed. Of the many programs that do exist, some have hol servers or owners and managers-as on-screen been designed by private, for-profit companies, some are spokespeople instead of legal or scientific experts. part of publicly funded research projects, and some have been developed by organizations that have close ties to the Communication methods. The programs varied in terms of alcohol industry. mode of presentation. Eleven consisted of only a video- Our review of 22 existing programs shows great vari- tape, with no supplemental written material to guide ability across programs in terms of coverage of essential trainers or trainees. One program was delivered by an content areas, use of behavior change techniques, and interactive computer program. Six of the programs used communication methods. videotapes and a trainer. The remaining programs used a Many of the programs were rated strongly in one cate- combination of trainers and written material. gory or had one highly rated component, but only a few Fewer than half the programs had one or more compo- programs received high scores across all categories and nents that, in the raters' opinion, adequately used realistic components. No single program was outstanding in all scenarios and recommendations. On the whole, the respects. The two programs that scored highest in our majority of the programs were respectful of the audience; review (numbers 16 and 17) covered all content areas, had however, eight had at least one component that we felt good production values, and employed behavior change

168 PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME 1 13 techniques. These programs, however, like most of the the hospitality industry; representatives of insurance agen- others, focused solely on alcohol bartenders and waitstaff, cies that issue dram shop insurance; and experts in state with no attention given to managers or owners. alcohol laws. The purpose of server training programs is to alter The standards developed at the consensus conference alcohol serving practices to reduce intoxication levels of should include criteria for (a) content areas that must be customers and prevent illegal sales to underage people, all covered, (b) use of scientifically based behavior change with the ultimate goal of reducing alcohol-related prob- techniques, and (c) use of theoretically based communica- lems. Training programs that do not cover essential con- tion methods. State policy makers could then require that tent areas, have few behavior change components, or are training programs in their states meet these standards. A poorly produced are unlikely to achieve these goals. Our national consensus conference could also aid in the devel- results indicate that existing server training programs need opment of guidelines for establishing state systems for further refinement. evaluating and certifying quality training programs. One recommendation may be that states set up stan- Recommendations. Based on this qualitative review, we dard training programs in neutral institutions such as recommend that national standards should be developed technical colleges. Once programs using the identified for server and manager training programs. A national con- standards and systems are developed, researchers should sensus conference should be held, with participants from evaluate the programs to assess their effectiveness in a wide range of organizations and backgrounds, including: reducing sales to intoxicated and underage customers and representatives from the National Institute on Alcohol decreasing alcohol-related problems such as traffic Abuse and , the Centers for Disease Control crashes, homicides, and other forms of violence. and Prevention, the National Highway Transportation These approaches should significantly improve the Safety Association, the Center for Substance Abuse Pre- quality of server training nationally so that good intent vention, and interested foundations such as the Robert can be followed by effective action. Wood Johnson Foundation; public health and behavioral change researchers; specialists in curriculum develop- Preparation of this paper was assisted by grants from the Robert Wood ment; representatives of organizations that offer server Johnson Foundation and the National Institute on and training packages; representatives of alcohol retailers and Alcoholism.

References I. Baker SP, O'Neill B, Ginsburg MJ, Li G. The injury fact book. 2nd ed. I1. Russ NW, Geller ES. Training bar personnel to prevent drunken dri- New York: Oxford University Press; 1992. ving: a field evaluation. Am J Public Health 1 987;77:952-4. 2. Alcohol-related mortality and years of potential life lost-United 12. Saltz RF. The roles of bars and restaurants in preventing alcohol- States, 1987. MMWR Morb Mortal Wkly Rep 1990;39:173-188. impaired driving: an evaluation of server intervention. Eval Health Pro- 3. Bandura A, editor. Social learning theory. Englewood Cliffs (NJ): Pren- fessions 1987; 10:5-27. tice Hall; 1977. 13. Holder HD, Wagenaar AC. Mandated server training and reduced 4. Perry CL, Jessor R. The concept of health promotion and the preven- alcohol-involved traffic crashes: a time series analysis of the Oregon tion of adolescent drug abuse. Health Educ Q 1985; 12:169-84. experience. Accid Anal Prev 1994;26:89-97. 5. Gliksman L, Single E. A field evaluation of a server intervention pro- 14. Roski J, Perry CL, McGovern PG, Williams CL, Farbakhsh K, Veblen- gram: accommodating reality. Paper presented at the Canadian Evalua- Mortenson S. School and community influences on adolescent alcohol tion Society Meetings; 1988 May; Montreal, Canada. and drug use. Health Educ Res 1997; 12:25566. 6. Howard-Pitney B, Johnson MD, Altman DG, Hopkins R, Hammond, N. 15. Wolfson M, Toomey TL, Forster JL, Murray DM, Wagenaar AC. Alco- Responsible alcohol service: a study of server, manager, and environ- hol outlet policies and practices concerning sales to underage people. mental impact. Am J Public Health 1991;8 1:197-9. Addiction 1996;91:589-601. 7. McKnight JA. An evaluation of a host responsibility program. Washing- 16. McKnight JA. Server intervention: accomplishments and needs. Alcohol ton: Department of Transportation, National Highway Traffic Safety Health Res World 1993; 17:76-83. Administration (US); 1987. Final report: DOT HS 807 380. 17. Perry CL, Baranowski T, Parcel G. How individuals, environments, and 8. McKnight J. Factors influencing the effectiveness of server-intervention health behavior interact. social learning theory. In: Glanz K, Lewis FM, education. J Stud Alcohol 1991;52:389-97. Rimer BK, editors. Health behavior and health education. San Fran- 9. Molof MJ, Kimball C. A study of the implementation and effects of Ore- cisco: Jossey-Bass; 1990. p. 1 61-86. gon's mandatory alcohol server training program. Eugene (OR): Inte- 18. Mendelsohn H. Some reasons why information campaigns can succeed. grated Research Services, Inc.; 1994. Public Opinion Q 1973;37:50-61. 10. National Highway Traffic Safety Administration (US). TEAM: Tech- 19. Robertson D. A review of factors influencing mass media health com- niques of Effective Alcohol Management: findings from the first year. munications and a model for program development. Ottawa: Canadian Washington: NHTSA; 1986. Health Education Society; 1982. U

PUBLIC HEALTH REPORTS * MARCH/APRIL 1998 * VOLUME 1 13 169