"Statin" Lipid-Lowering Drugs Compared with Guidelines

ORIGINAL INVESTIGATION Use and Monitoring of “Statin” Lipid-Lowering Drugs Compared With Guidelines

Susan A. Abookire, MD, MPH; Andrew S. Karson, MD; Julie Fiskio; David W. Bates, MD, MSc

Background: In patients with high cholesterol, tional Cholesterol Education Program guidelines. Our pri- 3-hydroxy-3-methylglutaryl coenzyme A reductase mary outcome measures included, for all patients tak- inhibitors (or “statins”) have been shown to reduce ing statins, prevalence of appropriateness vs overuse, and overall mortality in primary and secondary prevention. for all patients with coronary heart disease, prevalence The National Cholesterol Education Program expert of appropriateness vs underuse. panel’s guidelines (Adult Treatment Panel II) recommend evaluation and treatment of high cholesterol Results: Overuse of statin therapy was found among based on stratification of patients according to cardio- 69% of patients undergoing primary prevention, and vascular risk. While evidence suggests that many among 47% of patients undergoing secondary preven- patients are undertreated, comparatively few data are tion. In addition, among patients with coronary heart available regarding overtreatment. disease who were not taking statins, 88% were undertreated. Monitoring of liver function varied widely, and Objectives: To assess the appropriateness of statin did not correlate with the risk of adverse events second- therapy compared with national guidelines and to ex- ary to statin use. amine the appropriateness of monitoring for adverse effects. Conclusions: Overtreatment and undertreatment for hy- perlipidemia were frequent. Decision support may help Methods: For all patients at a tertiary medical center, physicians improve their performance compared with electronic medical records were evaluated for presence guidelines. or absence of statin use and for presence of established coronary heart disease or cardiac risk factors. Therapy was compared with the recommendations of the Na- Arch Intern Med. 2001;161:53-58

ANY epidemiologic a survival benefit in patients with estab- studies1,2 over the past lished CHD and significantly elevated se- several decades have rum cholesterol values; these results were established the rela- later extended to patients with CHD whose tion between an el- LDL cholesterol levels were only mod- Mevated serum cholesterol level and the de- estly elevated5 and even to patients with velopment of coronary heart disease CHD who had LDL cholesterol values in (CHD). In 1993, the expert panel of the the average range.6 In patients without es- National Cholesterol Education Program tablished CHD, the benefit of lowering (NCEP) proposed guidelines to stratify pa- LDL cholesterol levels with statin therapy tients according to risk of CHD, based on has been shown in clinical trials to re- cholesterol values and other risk factors. duce the incidence of myocardial infarc- The guidelines recommend drug therapy tion and mortality from coronary events, for individuals at greatest risk.3 supporting their use in primary preven- From the Division of General More recently, controlled trials4-8 have tion.7 These findings were also extended Medicine, Department of demonstrated a conclusive reduction in to show a reduction in coronary events Medicine, Brigham and overall mortality and mortality from CHD among patients with modest LDL choles- Women’s Hospital and Harvard among patients whose low-density lipo- terol elevations, using aggressive LDL cho- Medical School (Drs Abookire, 8 Karson, and Bates), and protein (LDL) cholesterol values were low- lesterol lowering. Partners Information Systems ered with 3-hydroxy-3-methylglutaryl co- Despite the evidence of preventable (Drs Abookire and Bates and enzyme A reductase inhibitor (“statin”) deaths among patients with CHD, sev- Ms Fiskio), Boston, Mass. drug therapy. Initial trials4 demonstrated eral studies have suggested undertreat-

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 53

Downloaded From: https://jamanetwork.com/ on 09/29/2021 PATIENTS AND METHODS and coexisting medications or disease conditions that could increase the risk of adverse effects. We defined patients as meeting criteria for secondary STUDY SITE AND PATIENTS prevention if they had CHD; we did not include a broader definition of athersclerotic disease for this analysis. Pa- This study was performed at a tertiary care center, the tients were identified as having CHD if their computer- Brigham and Women’s Hospital, Boston, Mass, and its af- ized problem lists indicated coronary artery disease, myo- filiated sites. Data were drawn from an electronic outpa- cardial infarction, coronary artery bypass graft, angina, or tient medical record,14 which is used at most sites affili- percutaneous transluminal coronary angioplasty; patients ated with the hospital, including hospital-based practices, without these problems were considered to be taking stat- free-standing community practices, and community health ins for primary prevention. Patients undergoing primary centers. The electronic medical record includes coded prob- prevention were examined for the presence of factors widely lem lists, medication lists, and laboratory data. accepted as conferring risk for heart disease. These in- To assess the accuracy of the electronic medical re- cluded hypertension; current smoking status; diabetes melli- cord,15 we performed manual medical record reviews. In tus; family history of premature heart disease; male sex and an analysis of 670 records, we found that if a specific age older than 45 years; female sex and age older than 55 disease state was on the electronic problem list (coronary years, not taking hormone replacement therapy; and low artery disease, diabetes, or hypertension), then the prob- (Ͻ0.91 mmol/L [Ͻ35 mg/dL]) high-density lipoprotein cho- lem was also found on medical record review 98% of the lesterol values. A high-density lipoprotein cholesterol level time. Conversely, of 177 patient records manually re- greater than 1.55 mmol/L (Ͼ60 mg/dL) was considered a viewed, a disease state found on medical record review on negative risk factor. The total number of risk factors for each average had a 94% likelihood of also being on the elec- patient was summed. Since established guidelines stratify tronic problem list. Similarly, if certain drugs (statins or patients according to whether they have 2 or more risk fac- hormone replacement agents) were on the electronic medi- tors3 or less than 2, we also categorized patients this way. cation list, then more than 95% of the time these agents appeared on medical record review; if these drugs were GUIDELINES found on medical record review, then they were found on the electronic medication list roughly 90% of the time. De- Guidelines for using pharmacological therapy to treat hy- mographic variables (age and sex) and laboratory data percholesterolemia are based on LDL cholesterol values.3 (LDL and high-density lipoprotein cholesterol levels and Thus, we retrieved the most recent LDL cholesterol value the results of liver function tests) were 100% accurate. In- before the initiation of statin therapy. This value, com- formation regarding certain risk factors (smoking and fam- bined with the indication and number of risk factors, was ily history of heart disease) was variably documented in pa- compared with guidelines for initiating statin therapy. Pa- tient medical records and electronic problem lists; however, tients were considered appropriate if their risk factor sta- when the risk factor information appeared, it appeared in tus and LDL cholesterol value before statin initiation were both places. in accordance with guidelines. To determine overuse, we evaluated the cohort of all For patients undergoing primary prevention, those patients taking statins as of January 1, 1996. Among pa- with less than 2 risk factors were considered appropriate tients taking statins, records were further studied to deter- if their LDL cholesterol level before drug initiation was mine the indication for statin use (primary or secondary greater than 4.92 mmol/L (Ͼ190 mg/dL); those with 2 prevention), lipid profiles, and contraindications to statin or more risk factors were considered appropriate if their use. In addition, we evaluated the amount of monitoring prior LDL cholesterol value was greater than 4.14 for liver function abnormalities, the impact of monitoring, mmol/L (Ͼ160 mg/dL).

ment of this group by primary care physicians9-12 and car- to previously established recommendations. Secondary diologists.13 In contrast, relatively little is known about aims included evaluating the impact that monitoring had the frequency of overtreatment, particularly among pa- on clinical outcomes, estimating the safety of statins in tients undergoing primary prevention. Furthermore, little this cohort, and estimating the financial burden of moni- is known about how practice patterns for monitoring the toring. We sought to quantify the occurrence of less ob- safety of these medications vary or compare with recom- vious adverse effects, in addition to hepatotoxicity. We mendations. Since primary care physicians hold a stra- also assessed the potential costs and savings associated tegic position in the detection and management of health with inappropriate and appropriate statin use and liver problems, their adherence to standards such as consen- function monitoring. sus guidelines may have a widespread effect. To address these issues, we performed a study with RESULTS several goals. Our primary aim was to assess the prevalence of appropriateness of statin therapy compared with Among 29543 outpatients who visited their primary care established guidelines. This included the prevalence of physician during 1996, 1575 (5%) were taking statins. overuse among all patients taking statins and the preva- Patients taking statins were 60% female, and their mean lence of underuse among patients with established CHD. age was 63 years (Table 1). Among patients taking stat- Similarly, we wanted to assess how patients were moni- ins, 69% were treated for primary prevention, and 31% tored for adverse effects, and the relation of monitoring had established CHD. Total cholesterol values were mea-

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 54

Downloaded From: https://jamanetwork.com/ on 09/29/2021 We considered patients with CHD (secondary pre- abnormal liver function were individually analyzed to de- vention) to be inappropriately taking statin therapy if their termine the impact of these abnormal results on their statin LDL cholesterol value before drug therapy was below 2.59 therapy and to determine whether these patients had any mmol/L (Ͻ100 mg/dL). To estimate underuse of statins, clinically significant abnormality. we reexplored our database for patients who met our cri- Patient problem lists, which included medical conditions teria for CHD and who were not taking statins. When pa- that would increase the risk of using statins (such as hepatitis tients had LDL cholesterol values greater than 2.59 mmol/L or liver disease), were individually reviewed to determine if (Ͼ100 mg/dL) in the presence of CHD and were not tak- monitoring was appropriate. Records were also reviewed to ing statins, we considered this inappropriate underuse. For determine if the problem preceded or resulted from statin use. the sake of this analysis, patients with CHD who were tak- We also evaluated whether monitoring was more fre- ing statins but had not reached the goal LDL cholesterol quent if medications interacting with statins were being level of 2.59 mmol/L (100 mg/dL) were not considered in- taken concurrently. These medications were itraconazole, appropriately treated. nefazodone hydrochloride (Serzone), gemfibrozil, clofi- brate, cyclosporine, and niacin. ANALYSIS Potential adverse reactions to the statin medications were identified by searching patient records for problems Patient demographics, overall lipid values, and the range of that might be related to statin use, including rhabdomy- lipid and liver function monitoring were assessed. The pro- olysis, myositis, sleep disorder or insomnia,16 and throm- portions of inappropriate use were calculated, including over- bocytopenia.17 If any of these problems were present, the use in primary and secondary prevention and underuse in individual patient records were manually reviewed to de- secondary prevention. Patient demographics were com- termine if the problems appeared to have any relation to pared between those meeting and not meeting guidelines. use of the statin medication. Logistic regression modeling was used to identify pre- We estimated potential annual medication and liver dictors of overuse of statin therapy. Indication (primary vs monitoring charges and savings that might be realized if secondary) and patient sex were binary covariates. The 2 guidelines for use and monitoring were followed. To as- continuous variables—patient age and number of risk fac- sess the drug costs of overuse, we used weighted averages tors—were assessed to determine the most appropriate form of the statin drugs in our cohort and their 1996 average to be used in the model. Both variables, when grouped into wholesale prices.18 We estimated the drug cost of correct- categories, showed a nonlinear relation to overuse when ing underuse by choosing a particular statin (atorvastatin used in a logistic regression model. Thus, for the final model, calcium) and calculating the cost using its 1996 average age was categorized into clinically relevant groups, and wholesale price.18 This statin was chosen because its aver- “number of cardiac risk factors” was collapsed into a bi- age wholesale price was the lowest among the statins and nary variable to correlate with guidelines (Ͻ2vsՆ2). would, therefore, result in the most conservative estimate Liver function monitoring was reviewed to assess the of the cost of correcting underuse. range and variability of monitoring frequency compared with The expected cost of liver function monitoring was es- recommendations. To evaluate whether more vigilant moni- timated using an average of 2 episodes of monitoring per toring occurred with increased risk of adverse events due person appropriately undergoing therapy. Excess cost, due to statin use, Spearman rank correlations were used to com- to overmonitoring patients appropriately undergoing therapy pare the degree of abnormality of the laboratory result with or monitoring patients who were inappropriately under- the frequency of monitoring. The effect of monitoring was going therapy, was estimated using prices in the Brigham further assessed to determine whether the abnormalities and Women’s Hospital laboratory. had an impact on therapy, such as discontinuation or re- All data analyses were performed using SAS statisti- placement of the statin drug. Records showing patients with cal software.19

sured on average 2.8 times during the 12-month period. older, being treated for primary prevention, and having There was no correlation between monitoring fre- fewer than 2 risk factors were significant (PϽ.001) pre- quency and the amount of total cholesterol elevation. dictors of statin overuse, but patient sex was not For LDL cholesterol levels, the correlation was modest (Table 4). Reintroducing the covariate of sex to our (Table 2). model did not reveal confounding. Based on LDL cholesterol levels and risk factor sta- Among patients with CHD who were not receiving tus, only 336 (31%) of 1080 patients undergoing pri- statin therapy (n=1459), 88% met the criteria for being mary prevention met NCEP guidelines. Of patients un- able to receive a statin and were thus undertreated. Among dergoing secondary prevention therapy, 260 (53%) of 495 patients with CHD who were taking statins (n=544), 47% met guidelines (Figure 1). Among patients undergo- did not meet the criteria and were, therefore, being over- ing primary prevention who did not meet NCEP guide- treated according to guidelines (Figure 2). lines, 69% had fewer than 2 risk factors; this group had We found that liver function monitoring during a a mean total cholesterol level of 6.40 mmol/L (247 mg/ 1-year period varied widely (Table 5). Among approxi- dL) before beginning therapy. The remaining 31% had mately 5000 liver function tests performed on the co- 2 or more risk factors, and the mean total cholesterol level hort of 1575 patients taking statins, 37 (2%) of the pa- for this group was 6.14 mmol/L (237 mg/dL) (Table 3). tients had values greater than 3 times normal. Thirty- In a logistic regression analysis using appropriate- five patients had their statin medication changed or ness as a binary outcome, we found that age 70 years or discontinued within 90 days of an abnormal result; 10

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 55

Downloaded From: https://jamanetwork.com/ on 09/29/2021 Table 1. Characteristics of 1575 Patients Table 3. Characteristics of 744 Patients Taking Statins Taking Statin Drugs During 1996* for Primary Prevention Who Did Not Meet NCEP Guidelines, by Number of Risk Factors* Characteristic Variable† No. of Risk Factors Female sex 948 (60.2) Age, mean (SD), y 63 (11.3) Characteristic Ͻ2 Ն2 Statin Lovastatin 1106 (70.2) Did not meet NCEP guidelines 513 (69) 231 (31) Simvastatin 280 (17.8) Values before statin initiation Pravastatin sodium 183 (11.6) Cholesterol level, mean (SD), mmol/L 6.40 (1.19) 6.14 (1.20) Fluvastatin sodium 6 (0.4) LDL cholesterol level, mmol/L Ͻ Atorvastatin calcium 0 (0) 4.92 334 NA Ͻ Cholesterol level, mean (SD), mmol/L‡ 5.96 (1.26) 4.14 NA 124 Ͻ HDL cholesterol level, mean (SD), mmol/L‡ 1.22 (0.38) 2.59 18 13 LDL cholesterol level, mean (SD), mmol/L‡ 3.65 (1.29) Not tested 179 107 Primary prevention 1080 (68.6) Cholesterol level, mean (SD), 6.68 (1.62) 6.45 (1.43) Secondary prevention 495 (31.4) mmol/L, if LDL cholesterol level not tested Neither LDL cholesterol nor cholesterol 50 26 Statin drugs are 3-hydroxy-3-methylglutaryl coenzyme A reductase * level tested inhibitors. HDL indicates high-density lipoprotein; LDL, low-density lipoprotein. †Data are given as number (percentage) of patients unless otherwise *Data are given as number of patients unless otherwise indicated. Statins indicated. are 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. NCEP ‡To convert cholesterol, HDL cholesterol, and LDL cholesterol from indicates National Cholesterol Education Program; LDL, low-density millimoles per liter to milligrams per deciliter, divide millimoles per liter by lipoprotein; and NA, data not applicable. 0.0259.

Table 4. Results of Logistic Regression* Table 2. Range of Monitoring Frequencies Wald Times Monitored Spearman Rank Correlation Predictor Odds Ratio Confidence Limits Value Monitored* During 1996† With Magnitude of Result Primary prevention† 1.6‡ 1.3-2.0 Cholesterol 2.8 (1-58) −0.02 Ͻ2 Risk factors§ 1.5‡ 1.2-1.8 HDL cholesterol 1.6 (1-6) −0.09 Age Ն70 y࿣ è.5‡ 1.2-1.9 LDL cholesterol 1.5 (1-6) 0.13 *The outcome is the odds of inappropriate prescribing of a statin *HDL indicates high-density lipoprotein; LDL, low-density lipoprotein. (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor). †Data are given as mean (range). †Compared with patients undergoing secondary prevention. ‡PϽ.001. §Compared with patients with 2 or fewer cardiac risk factors. 800 ࿣Compared with patients younger than 70 years.

Inappropriate Statin Use 700

Appropriate Statin Use We also evaluated whether monitoring was more 600 intensive among patients receiving a drug that in- teracted with statins. Ninety-eight patients (6%) 500 were found to be taking other medications with important drug-drug interactions, including niacin, 400 gemfibrozil, cyclosporine, and itraconazole. However,

No. of Patients 300 these patients were not monitored more frequently. None of these patients had clinically significant 200 adverse events. Nine patients had other documented problems that may be considered adverse reactions to 100 the statin drugs, including sleep disorder and throm- bocytopenia, but none required discontinuation of the 0 Primary Prevention Secondary Prevention drug. One impact of inappropriate overuse and liver Figure 1. Appropriateness of 3-hydroxy-3-methylglutaryl coenzyme A function monitoring is cost. Based on the average reductase inhibitor (statin) use in primary and secondary prevention for 1575 patients taking statins. wholesale price with weighted averages of specific statin drugs, from the payer perspective an estimated $1338449 in annual cost savings might have been real- of these were confirmed, on individual record review, to ized if statin use in primary and secondary prevention be related to the abnormal laboratory result. Two pa- were restricted to NCEP guidelines. If recommended tients remained off statins as a result of abnormal liver liver function monitoring were followed, we estimated a values, and none had clinical manifestations of hepati- further potential annual cost savings of $26620. The tis. The frequency of monitoring did not correlate with cost of correcting underuse among patients with CHD the level of test abnormality. would be $841020.

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 56

Downloaded From: https://jamanetwork.com/ on 09/29/2021 29 543 Eligible Patients

2003 With CAD 27 540 Without CAD

544 Taking Statins 1459 Not Taking Statins 1080 Taking Statins 26 460 Not Taking Statins

53% 47% 12% 88% 31% 69% Appropriate Inappropriate Appropriate Inappropriate Appropriate Inappropriate (Overuse) (Underuse) (Overuse)

Figure 2. Overview of patients eligible for 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor (statin) therapy. CAD indicates coronary artery disease.

COMMENT Table 5. Range of Monitoring Liver Function for 1575 Patients Taking Statins These data suggest that, despite widely available guide- * lines for the use of drug therapy in primary and second- Laboratory Value Checked Variable ary cardiovascular disease prevention, use of statin lipid- lowering therapy is often inappropriate. Overuse of statin ALT therapy was found among 69% of patients undergoing No. of times checked 2545 No. (%) of times abnormal 145 (6) primary prevention, and among 47% of patients under- No. of patients with an abnormal result 57 going secondary prevention. Overuse was more preva- Frequency of monitoring, mean (SD) 3 (3.9) lent among patients who were being treated for primary Frequency of monitoring correlated with level 0.08† prevention, who were older than 70 years, or who had AST fewer than 2 cardiac risk factors. We also found an 88% No. of times checked 2509 rate of underuse among patients undergoing secondary No. (%) of times abnormal 343 (14) No. of patients with an abnormal result 165 prevention who were not taking statins. Furthermore, Frequency of monitoring, mean (SD) 3 (3.9) monitoring for safety varied widely, and was not inten- Frequency of monitoring correlated with level 0.09† sified for patients at highest risk. The potential pharmacy and laboratory savings that *Statins are 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. would occur by eliminating overuse are substantial. Also, ALT indicates alanine aminotransferase; AST, aspartate aminotransferase. the cost of correcting underuse would be more than off- †Spearman coefficients. set by the savings of eliminating overuse, and might also reduce the morbidity of CHD. These cost and savings es- orders, can modify ordering behavior.32 In addition, de- timates represent drug and laboratory charges; they do cision support is effective for helping physicians remem- not include social costs of treatment, such as lost work ber to implement an order that follows from another or- days, or costs of treating adverse drug events. der, such as ordering laboratory tests to monitor liver These results are consistent with those of several function after the initiation of statin therapy.32 other studies20-22 that demonstrate lack of adherence to This study also illustrates the power of the elec- available guidelines. Our findings extend prior evi- tronic medical record for measuring quality. Although dence of undertreatment to show that overtreatment is such records are not yet widely used, they have many ben- also a significant concern, and that a substantial finan- efits,30 and facilitation of quality measurement is high on cial burden is associated with overtreatment. Moreover, the list. we assessed adherence to the NCEP guidelines, which Our study has several limitations. While some mis- tend to be aggressive regarding therapy; use of other guide- classifications undoubtedly occurred because of inaccu- lines23,24 might have suggested that overtreatment is even racies in the database, they could not account for these more frequent. figures; guideline adherence could clearly be improved. Several studies25-27 have shown that publication of We may have overestimated the amount of underuse guidelines without more intensive accompanying infor- among patients undergoing secondary prevention, since mation has little impact on clinical practice. This sug- our manual medical record review revealed 1 patient of gests that research evidence and consensus statements 20 to be actually taking a statin. If this is the actual pro- are not primary determinants of physician behavior. Prac- portion of underascertainment, then the rate of under- tice-based interventions may be more effective at hav- use would fall to 84%, which is still a formidable figure. ing an impact on practice behavior.27-29 Also, some people consider that other populations should Tools for improving compliance, and reducing the be included in the secondary prevention group, such as number of errors, include reminders and computerized patients with a history of cerebrovascular accident or pe- alerts.30,31 A growing body of evidence suggests that such ripheral vascular disease, but we followed a strict inter- computerized decision support, especially when pre- pretation of the NCEP guidelines. In addition, we did not sented at key times such as when physicians are writing address whether patients undergoing secondary preven-

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 57

Downloaded From: https://jamanetwork.com/ on 09/29/2021 tion actually achieved the recommended LDL choles- 8. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute coronary events terol values; many undoubtedly did not. Another limi- with lovastatin in men and women with average cholesterol levels: results of AFCAPS/TexCAPS. JAMA. 1998;279:1615-1622. tation is that this study was done at one site, so our results 9. McBride P, Schrott H, Plane M, Underbakke G, Brown RL. Primary care practice may not be generalizable to other populations. How- adherence to National Cholesterol Education Program guidelines for patients with ever, poor adherence to guidelines has been found in other coronary heart disease. Arch Intern Med. 1998;158:1238-1244. studies, and is extended herein to show overtreatment 10. Schrott HG, Bittner V, Vittinghoff E, Herrington DM, Hulley S, for the HERS Re- search Group. Adherence to National Cholesterol Education Program treatment in primary prevention and to show inappropriate safety goals in postmenopausal women with heart disease: the Heart and Estrogen/ monitoring. Progestin Replacement Study (HERS). JAMA. 1997;277:1281-1286. One possible explanation of our findings of statin 11. Stafford R, Blumenthal D, Pasternak R. Variations in cholesterol management overuse is that physicians may be extrapolating from practices of US physicians. J Am Coll Cardiol. 1997;29:139-146. recent trials supporting a more aggressive approach to 12. Eaton C, McQuade W, Glupczynski D. A comparison of primary vs secondary cardiovascular disease prevention in an academic family practice. Fam Med. 1994; lipid lowering instead of following the NCEP guide- 26:587-592. lines. However, our data are drawn from physician 13. Cohen M, Byrne M, Levine B, Gutowski T, Adelson R. Low rate of treatment of behavior as of January 1, 1996, before many of the clini- hypercholesterolemia by cardiologists in patients with suspected and proven cal trials, particularly those in primary prevention. An coronary artery disease. Circulation. 1991;83:1294-1304. 14. Teich JM, Glaser JP, Beckley RF, et al. Toward cost-effective, quality care: the additional implication may be that the guidelines are Brigham Integrated Computing System. In: Steen EB, ed. Proceedings of the outdated, and should be revised to reflect more current Second Nicholas E. Daives CPR Recognition Symposium. Chicago, Ill: Computer- evidence. Based Patient Record Institute; 1996:3-34. We conclude that, taken together, these results sug- 15. Wagner MM, Hogan WR. The accuracy of medication data in an outpatient elec- gest a substantial and costly burden of statin overtreat- tronic medical record. J Am Med Inform Assoc. 1996;3:234-244. 16. Partinen M, Pihl S, Strandberg T, et al. Comparison of effects on sleep of lova- ment and undertreatment, and widely varying liver func- statin and pravastatin in hypercholesterolemia. Am J Cardiol. 1994;73:876-880. tion monitoring for adverse effects. Decision support, 17. Mantell G, Burke T, Staggers J. Extended clinical safety profile of lovastatin. Am offered during the prescribing and laboratory test– J Cardiol. 1990;66:11B-15B. ordering processes, may help physicians optimize use of 18. Choice of lipid-lowering drugs. Med Lett Drugs Ther. 1996;38:67-70. 19. SAS Institute Inc. SAS, Version 6.12. Cary, NC: SAS Institute Inc; 1997. these medications from the population perspective. 20. Thorndike AN, Rigotti NA, Stafford RS, Singer DE. National patterns in the treatment of smokers by physicians. JAMA. 1998;279:604-608. Accepted for publication June 30, 2000. 21. Cohen SJ, Robinson D, Dugan E, et al. Communication between older adults and This study was supported in part by a grant from their physicians about urinary incontinence. J Gerontol A Biol Sci Med Sci. 1999; 54:M34-M37. Aetna–US Healthcare, Hartford, Conn. 22. McBride P, Schrott H, Plane M, Underbakke G, Brown R. Primary care practice Corresponding author and reprints: Susan A. Abookire, adherence to National Cholesterol Education Program guidelines for patients with MD, MPH, Division of General Medicine, Department of coronary heart disease. Arch Intern Med. 1998;158:1238-1244. Medicine, Brigham and Women’s Hospital, 75 Francis St, 23. Garber AM, Browner WS, Hulley SB. Cholesterol screening in asymptomatic adults, revisited: part 2. Ann Intern Med. 1996;124:518-531. Boston, MA 02115 (e-mail: [email protected]). 24. Canadian Consensus Conference on Cholesterol: final report: the Canadian Con- sensus Conference on the Prevention of Heart and Vascular Disease by Altering REFERENCES Serum Cholesterol and Lipoprotein Risk Factors. CMAJ. 1998;139:1-8. 25. Basinski AS. Evaluation of clinical practice guidelines. CMAJ. 1995;153:1575- 1581. 1. Kannel WB, Castelli WP, Gordon T, McNamara PM. Serum cholesterol, lipopro- 26. Lomas J, Anderson GM, Domnick-Pierre K, Vayda E, Enkin MW, Hannah WJ. Do teins, and the risk of coronary heart disease: the Framingham study. Ann Intern practice guidelines guide practice? the effects of a consensus statement on the Med. 1971;74:1-12. practice of physicians. N Engl J Med. 1989;321:1306-1311. 2. Multiple Risk Factor Intervention Trial Research Group. Multiple Risk Factor In- 27. Davis DA, Thomson MA, Oxman AD, Haynes RB. Changing physician perfor- tervention Trial: risk factor changes and mortality results. JAMA. 1982;248:1465- mance: a systematic review of the effect of continuing medical education strat- 1477. egies. JAMA. 1995;274:700-705. 3. Summary of the second report of the National Cholesterol Education Program 28. Cohen SJ, Halvorson HW, Gosselink CA. Changing physician behavior to im- (NCEP) expert panel on detection, evaluation, and treatment of high blood cho- prove disease prevention. Prev Med. 1994;21:909-914. lesterol in adults (Adult Treatment Panel II). JAMA. 1993;269:3015-3023. 29. Dugan E, Cohen SJ. Changing physician behavior to increase guideline imple- 4. The Scandinavian Simvastatin Survival Study Group. Randomised trial of cho- mentation. In: Shumaker SA, Schron E, Ockene J, Mcbee W, eds. Handbook for lesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Health Behavior Change. 2nd ed. New York, NY: Springer Publishing Co Inc; 1998: Simvastatin Survival Study (4S). Lancet. 1994;334:1383-1389. 283-304. 5. Cholesterol and Recurrent Events (CARE) Trial Investigators. The effect of prava- 30. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized order entry and a statin on coronary events after myocardial infarction in patients with average cho- team intervention on prevention of serious medication errors. JAMA. 1998;280: lesterol levels. N Engl J Med. 1996;335:1001-1009. 1311-1316. 6. The LIPID Study Group. Prevention of cardiovascular events with pravastatin in 31. Raschke RA, Gollihare B, Wunderlich TA, et al. A computerized alert system to patients with coronary heart disease and a broad range of initial cholesterol lev- prevent injury from adverse drug events: development and evaluation in a com- els. N Engl J Med. 1998;339:1349-1357. munity teaching hospital. JAMA. 1998;280:1317-1320. 7. West of Scotland Coronary Prevention Study Group. Prevention of coronary heart 32. Overhage JM, Tierney WM, Zhou XA, McDonald CJ. A randomized trial of “cor- disease with pravastatin in men with hypercholesterolemia. N Engl J Med. 1995; ollary orders” to prevent errors of omission. J Am Med Inform Assoc. 1997;4: 333:1301-1307. 364-375.

(REPRINTED) ARCH INTERN MED/ VOL 161, JAN 8, 2001 WWW.ARCHINTERNMED.COM 58

Downloaded From: https://jamanetwork.com/ on 09/29/2021