Long-Term Adherence to Evidence-Based Secondary Prevention Therapies in Coronary Artery Disease
Background— Studies have examined the use of evidence-based therapies for coronary artery disease (CAD) in the short term and at hospital discharge, but few have evaluated long-term use.
Methods and Results— Using the Duke Databank for Cardiovascular Disease for the years 1995 to 2002, we determined the annual prevalence and consistency of self-reported use of aspirin, β-blockers, lipid-lowering agents, and their combinations in all CAD patients and of angiotensin-converting enzyme inhibitors (ACEIs) in those with and without heart failure. Logistic-regression models identified characteristics associated with consistent use (reported on ≥2 consecutive follow-up surveys and then through death, withdrawal, or study end), and Cox proportional-hazards models explored the association of consistent use with mortality. Use of all agents and combinations thereof increased yearly. In 2002, 83% reported aspirin use; 61%, β-blocker use; 63%, lipid-lowering therapy use; 54%, aspirin and β-blocker use; and 39%, use of all 3. Consistent use was as follows: For aspirin, 71%; β-blockers, 46%; lipid-lowering therapy, 44%; aspirin and β-blockers, 36%; and all 3, 21%. Among patients without heart failure, 39% reported ACEI use in 2002; consistent use was 20%. Among heart failure patients, ACEI use was 51% in 2002 and consistent use, 39%. Except for ACEIs among patients without heart failure, consistent use was associated with lower adjusted mortality: Aspirin hazard ratio (HR), 0.58 and 95% confidence interval (CI), 0.54 to 0.62; β-blockers, HR, 0.63 and 95% CI, 0.59 to 0.67; lipid-lowering therapy, HR, 0.52 and 95% CI, 0.42 to 0.65; all 3, HR, 0.67 and 95% CI, 0.59 to 0.77; aspirin and β-blockers, HR, 0.61 and 95% CI, 0.57 to 0.65; and ACEIs among heart failure patients, HR, 0.75 and 95% CI, 0.67 to 0.84.
Conclusions— Use of evidence-based therapies for CAD has improved but remains suboptimal. Although improved discharge prescription of these agents is needed, considerable attention must also be focused on understanding and improving long-term adherence.
Received September 8, 2004; revision received August 9, 2005; accepted August 12, 2005.
Evidence from randomized clinical trials supports the use of several pharmacological therapies in the secondary prevention of coronary artery disease (CAD). Yet despite these data, previous studies have documented underprescription of aspirin, β-blockers, lipid-lowering therapy, and angiotensin-converting enzyme inhibitors (ACEIs) among patients with CAD.1–8 It has been estimated that optimal prescription of these agents at hospital discharge could save as many as 80 000 lives per year in the United States.9 Accordingly, several large quality-improvement efforts have been undertaken to improve use of these agents in the short term and at hospital discharge, including the American Heart Association’s (AHA’s) “Get With The Guidelines” program, the American College of Cardiology’s (ACC’s) Guidelines Applied in Practice efforts, and the Can Rapid Risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines (CRUSADE) Quality-Improvement Initiative.
Editorial p 178
Although substantial attention has been focused appropriately on improving prescription of evidence-based therapies by physicians in the hospital and at the time of hospital discharge, much less attention has been paid to understanding and improving long-term adherence to these medications. We examined changes in the prevalence of use of evidence-based secondary prevention medications (aspirin, β-blockers, lipid-lowering therapy, and ACEIs) over time among patients with documented CAD and determined the consistency of use over time among these patients. Furthermore, we sought to identify patient-related factors associated with use of these agents in long-term follow-up and secondarily to explore the association of consistent use with mortality.
For the years 1995 to 2002, we identified from the Duke Databank for Cardiovascular Disease (DDCD), a clinical database that includes all patients who have undergone a cardiac procedure at Duke University Medical Center since 1969, those patients still living with documented CAD (at least 1 documented coronary stenosis of >50% or coronary bypass surgery) who had at least 1 follow-up survey (process described next) completed during the period 1995 to 2002. Among the total population of 31 750 patients, we then identified 2 subgroups of patients with CAD: (1) Those without heart failure (n=22 539) by excluding patients with a documented ejection fraction <40% or a clinical history of heart failure and (2) those with a clinical heart failure event or ejection fraction <40% identified at baseline or occurring during the study period who had at least 1 follow-up after the diagnosis of heart failure (n=8914).
Follow-Up Process and Ascertainment of Medication Use
After their index procedure, at which time baseline clinical data were collected, at 6 months, and then yearly thereafter, all patients in the DDCD who have documented coronary lesions of >50% are administered follow-up surveys of their clinical status through the mail or secondarily by telephone contact if there is no response to mailed questionnaires. Since 1995, these surveys have collected self-reported medication use by having patients actively list the medications they are currently taking. Patients are prompted to include aspirin if taken regularly. For the purposes of our analyses, the “first follow-up survey” is considered to be the first expected survey since the initiation of medication collection in 1995. For example, for a patient enrolled in 1993, the first survey containing medication information would be expected in 1995, and for a patient entering the DDCD in 1996, it would be 6 months later. Mortality is confirmed by yearly search of the National Death Index. Since the inception of the DDCD, only 2.6% of patients have withdrawn from active follow-up; 0.2% have left the United States, and 2.3% are considered lost to follow-up. At any point, &2% of patients have “alive only” status without complete follow-up survey information; these patients were not included in our analyses. Among patients actively followed up during the 1995 to 2002 period of the current study, the proportion of follow-up surveys reporting medication information was as follows: 86.5% for 1995; 91.9% for 1996; 93.9% for 1997; 93.5% for 1998; 92.8% for 1999; 92.3% for 2000; 93.9% for 2001; and 97.2% for 2002.
Study Definitions and Medication Categories
For purposes of defining consistent use, patients had to have at least 2 consecutive surveys returned during the study period. We initially defined 4 patterns of an individual’s medication use over time: (1) Always using: Patients who had at least 2 surveys and reported use at each follow-up; (2) never using: Patients who had at least 2 surveys and never reported use of a given agent during follow-up; (3) positive converters: Patients who reported no use of an agent initially but subsequently reported use on at least 2 consecutive surveys, continuing through the end of the study period or time of death; and (4) use to no use: Patients who on their early surveys reported use of an agent but on at least 2 consecutive subsequent surveys (and continuing through the end of the study period or death) did not report use. “Consistent use” was then defined as reporting a medication use on at least 2 consecutive occasions and continuing to do so until death, withdrawal from follow-up, or the end of the study period (ie, the sum of groups 1 and 3). Patients were considered inconsistent users if they met criteria for none of these patterns.
We were interested in patterns of use for the following therapies: Aspirin, β-blockers, and lipid-lowering agents; the combination of aspirin, β-blockers, and lipid-lowering agents; and the combination of aspirin plus β-blockers in all patients and of ACEIs in patients with CAD with and without heart failure. Because the supporting evidence and guideline recommendations for prescription of lipid-lowering therapy and ACEI use among patients without heart failure have been in flux over the time period studied and have not been recommended for all patients with CAD, we also evaluated use patterns for these agents that included only patients who had reported use of 1 of these medications on at least 1 occasion. In this case, the initial report of use was considered evidence of an indication for the drug.
Baseline characteristics were summarized by percentages for discrete variables and medians with 25th and 75th percentiles for continuous variables. Clinical characteristics and physical examination information were taken from the most recent procedure before the start of the study period. Medication use reflects that reported on the first follow-up survey. Prevalence of use of the specified agents or combinations of agents was summarized by percentages for each year under study. In the case of patients who reported no medications on their follow-up survey, our primary assessment of the prevalence of use included them in the population but counted them as not using the medication of interest. From all patients who had at least 2 follow-up surveys completed between 1995 and 2002, the proportion reporting consistent use of the same medication(s) as assessed in the annual prevalence analysis was also determined.
To better understand the patient factors associated with use, we developed logistic-regression models predicting the consistent use of each agent. For the purposes of this modeling, consistent users were defined as the combined categories of “always using” and “positive converters”; “never using” was the reference category. For combinations of agents, “never using” refers to failure to use all components of the combination. For the analyses that included ACEIs in patients without heart failure or lipid-lowering therapy, we restricted the population to those reporting at least 1 use of ACEIs or lipid-lowering therapy. After this 1 documented use, the subsequent pattern of use was classified by our predetermined categories of use. Variables considered in the stepwise modeling approach for each medication or combination of medications were as follows: Year of entry into the DDCD (2000 or later, between 1995 and 1999, between 1990 and 1994; reference was before 1990), demographic characteristics (age at entry into the DDCD, interaction of age with year of entry, sex, and race), cardiac risk factors (diabetes, hyperlipidemia, smoking, or hypertension), prior documented cardiovascular events or procedures before the study period (myocardial infarction, stroke, percutaneous coronary intervention, or bypass surgery), medical history (cerebrovascular disease, peripheral vascular disease, chronic obstructive pulmonary disease, renal disease, peptic ulcer disease, or heart failure), systolic and diastolic blood pressures at entry, events during the study period (myocardial infarction, stroke, percutaneous coronary intervention, or bypass surgery) findings at angiography (ejection fraction, missing ejection fraction, number of diseased vessels, or other vascular disease), and medications at first follow-up. For patients with >1 record in the DDCD, the data for these variables were obtained from the record most immediately before the first post-1995 clinical follow-up. Variables were considered significantly associated with use of a given medication when P<0.05. In exploratory analyses, Cox proportional-hazards modeling was used to examine the adjusted association with mortality of consistent use versus never use, with the same starting populations and considering these same variables according to the stepwise procedure. The starting time for the survival analysis was entry into the DDCD, and analyses were first stratified by year of entry into the DDCD. In addition, we then examined the effect on mortality of the interaction of age with year of entry into the DDCD. For the subset with CAD and heart failure, analyses were stratified by year of diagnosis of heart failure. We also performed a sensitivity analysis in which the starting time for the analysis was 1995, the year that collection of medication use was initiated and follow-up for the current study began. Although all variables listed earlier were considered in the Cox modeling, the variables included in the final Cox model for each agent and combination of agents are displayed in the online-only Appendix. Finally, we conducted a propensity analysis for the association of consistent use of each medication with mortality with a quintiles approach, in which patients were grouped according to propensity of consistent use into 5 equal subgroups. The relation between consistent use and mortality was assessed within each group.
We performed 2 additional exploratory analyses for both the association of covariates with consistent use and with mortality. In the first, we examined the effect of including patients in the consistent-use group who were initially classified as having an inconsistent pattern of use but who had at least 2 terminal consecutive “yes” answers. Second, we assessed the effect of changing the comparator group to a broad category of “not consistent use,” which included patients who were in the “use to no use” category, had inconsistent use, or were never users. Logistic-regression models with the variables described and Cox proportional-hazards models with the variables in the Appendix were performed for these analyses. All analyses were performed with SAS statistical software, version 8.2 (SAS Institute).
Displayed in Table 1⇓ are the baseline characteristics for overall CAD (n=31 750) and for CAD with (n=8914) and without (n=22 539) heart failure populations. Medications are those reported on the first follow-up survey. The yearly prevalences of use for each agent and combination of agents are displayed in Figure 1. Among all patients with CAD, the proportion of patients reporting use of each agent increased over time, with peak rates of use occurring in the year 2002. In 2002, the use of aspirin was 83%; β-blockers, 61%; and lipid-lowering agents, 63%. Use of all 3 drugs was reported by 39% of patients in 2002 and of aspirin and β-blockers, 54%. Among the subset of patients with CAD without heart failure, use of ACEIs also increased over time, to 39% in 2002. Among the subset of patients with CAD and heart failure, ACEI use also increased, from 32% in 1995% to 51% in 2002.
Consistence of Use of Evidence-Based Therapies
Table 2 shows the numbers of patients in each use category. In general, consistent use (always use and positive converters) of all medications and their combinations was low; only aspirin was used consistently in more than half of patients (71%). Consistent use of β-blockers was 46% and for lipid-lowering agents was 43%. Use of the combination aspirin, β-blockers, and lipid-lowering therapy was consistently reported by only 21% of patients and of aspirin plus β-blocker, 36%. Considering the effect of fluctuating guidelines and new clinical trial results during the period of observation, we examined consistent use of lipid-lowering therapy only among patients who had at least 1 reported use. Consistent use among those patients was 70%, and consistent use of the combination aspirin, β-blockers, and lipid-lowering therapy, 35%. Among patients with CAD without heart failure who reported at least 1 use of an ACEI, 61% consistently reported ACEI use subsequently. For the entire cohort with CAD without heart failure, consistent use of ACEIs was only 26%. Among patients with CAD with heart failure, consistent use of ACEIs was 39%.
Table 3 and Table 4 display factors that discriminated consistent users of each medication or combination of medications from nonusers in multivariable modeling. Among the strongest predictors, increasing age at entry into the DDCD was associated with lower consistent use of all agents, except ACEIs in the population without heart failure. Later year of entry into the DDCD (or year of diagnosis of heart failure among the heart failure subset) was associated with greater consistent use, except for lipid-lowering agents, in which entry after 1990 was associated with lower use than before 1990. There was no association of year of entry with use of ACEIs in patients without heart failure. Diabetes and smoking history were associated with lower consistent use of aspirin, β-blockers, and their combination, and a clinical history of heart failure was associated with lower consistent use of β-blockers. Revascularization during or before the observation period was associated with greater use of all therapies except ACEIs, and use of 1 or more other evidence-based therapies was associated with greater use of all therapies except ACEIs among patients without heart failure. As expected, use of warfarin and other nonaspirin antiplatelet agents was associated with lower use of aspirin, β-agonist use with lower use of β-blockers, and angiotensin II receptor blockers with lower use of ACEIs. Female sex was associated with lower use of aspirin, ACEIs, and lipid-lowering therapy. Factors associated with the use of combinations of aspirin and β-blockers and of aspirin, β-blockers, and lipid-lowering agents paralleled the observations for the individual agents.
Patterns of association with consistent use were not different when the comparison group was “not consistent users,” ie, never users plus the use to no use group and inconsistent users. Inclusion of individuals in the consistent-use group who were initially classified as inconsistent users by our primary definition but who had terminal-use reports that were “yes” on at least 2 consecutive surveys (for aspirin, n=1384; β-blockers, n=740; aspirin plus β-blockers, n=960; lipid-lowering agents, n=610; aspirin plus β-blockers plus lipid-lowering agents, n=457; and ACEIs, n=130) also did not change the factors associated with consistent use.
Consistent Use of Evidence-Based Therapies and Clinical Outcomes
Consistent use of aspirin, β-blockers, lipid-lowering therapy, and combinations of aspirin plus β-blockers and of aspirin, β-blockers, and lipid-lowering therapy and of ACEI use in patients with heart failure was associated with substantially better long-term survival in multivariable modeling after adjusting for the potential confounders listed in the Appendix (Table 5). After adjusting for potential confounders, survival in consistent users of ACEIs among patients without heart failure was not significantly different from that of nonusers. Propensity analysis showed that these findings were consistent across propensity quintiles. Using the year 1995 as the start time for the Cox proportional-hazards modeling rather than year of entry into the DDCD did not alter the patterns of association of consistent use of any agent or combination with mortality. When we repeated our analyses with “not consistent users” as the comparison group, ie, never users plus the use to no use group and inconsistent users, associations of consistent use with mortality were unchanged. However, inclusion of individuals in the consistent-use group who were initially classified as inconsistent users by our primary definition but who had terminal-use reports that were “yes” on at least 2 consecutive surveys resulted in a decrease in the strength of association of consistent use with lower mortality. For aspirin (hazard ratio [HR], 0.96; 95% confidence interval [CI], 0.90 to 1.02) and lipid-lowering agent (HR, 0.90; 95% CI, 0.81 to 1.00), the associations were no longer significant, and for ACEIs among patients without heart failure, the HR indicated statistically worse survival (HR, 1.28; 95% CI, 1.10 to 1.48).
In this study of long-term adherence to the use of evidence-based medications for CAD secondary prevention, we found that patient-reported annual outpatient use of evidence-based therapies has steadily improved since 1995 but remains suboptimal. More concerning, nearly 30% of patients were not consistently using aspirin, and fewer than half reported consistent long-term use of β-blockers, lipid-lowering therapy, or combinations of these life-saving drugs. Given that outpatient consistent use, as defined in our study, was associated with improved long-term outcomes, substantial improvements in the care of CAD patients and in their long-term outcomes could result from efforts focused on improving long-term adherence to proven therapies.
Improving the Use of Evidence-Based Medications
To date, efforts to improve the use of evidence-based medications among patients with CAD have focused primarily on inpatient administration of appropriate drugs during the acute phase of illness and prescription at hospital discharge. The link between improving adherence to guideline recommendations for management in these settings and clinical outcomes has now been demonstrated among patients hospitalized with acute coronary syndromes.10,11 Importantly, recent reports from several quality-improvement initiatives suggest that adoption of quality-improvement programs is having modest effects on improving adherence to the use of these lifesaving therapies.12–14
It is estimated that 13.2 million Americans have a history of coronary heart disease and therefore are at risk for a new or recurrent ischemic event each year.15 Thus, we focused our attention on the long-term outpatient use of evidence-based medications in secondary prevention as a powerful opportunity to improve clinical outcomes. Although previous studies have shown that discharge prescription of evidence-based medications is associated with better 1-year use of these agents6,16,17 and estimate that continued long-term use would not only be clinically beneficial but also cost-effective,18 little is known of the relation between consistent outpatient use over longer periods of follow-up and clinical outcomes or the factors that are associated with better longer-term adherence. In a single-center study of 600 patients, Muhlestein and colleagues19 found that 23% of patients with CAD documented by coronary angiography and prescribed a statin at discharge were not using a statin at an average of 3 years of follow-up. Our findings in >30 000 patients across multiple categories of evidence-based medications and a 7-year period of observation extend these findings and highlight the potential for programs designed to improve consistent use of proven therapies.
Despite marked increases in the yearly prevalence of use in our study population from 1995 to 2002, consistent use lagged substantially behind, suggesting that in addition to the importance of prescribing, additional factors are important in determining long-term adherence. This is further evidenced by our observation that when we selected only those patients with at least 1 report of use in follow-up, indicating that a prescription had been given and filled at least once, only 70% of patients reported consistent statin use in long-term follow-up, and only 61% without heart failure, an ACEI.
Characteristics Associated With Use of Proven Therapies
The strong association of consistent use with better long-term outcome underscores the importance of understanding characteristics that are associated with consistent use. Although in part improved outcome with consistent use may reflect the effect of the therapy itself, analyses from placebo-controlled studies such as the Coronary Drug Project study of clofibrate versus placebo suggest that unmeasured patterns of behavior that are associated with adherence may contribute substantially to improved outcomes as well.20 Whereas 5-year mortality values in the clofibrate (20.0%) and placebo (20.9%) groups overall were similar, in both groups, patients who took >80% of their protocol medication had substantially lower 5-year mortality than did those who were less compliant: Placebo group mortality, 15.1% for good adherers versus 28.3% for poor adherers; clofibrate group mortality, 15.0% for good adherers versus 24.6%.
As demonstrated in Table 2 and Table 3⇑ and in our previous work from the DDCD,21 a number of measured factors were associated either positively or negatively with consistent, long-term use of these proven therapies. However, it is concerning that consistent use of evidence-based medications was paradoxically lower among groups with the highest risk of poor outcomes and therefore who could potentially benefit the most from sustained therapy. In a pattern that has been observed in other settings, in our analysis, the elderly and patients with diabetes and evidence of heart failure were less likely to consistently use the evidence-based therapies under study. These findings suggest that it may be possible to design educational and compliance intervention programs targeted to groups of patients at high risk for both underuse of medications in secondary prevention and adverse clinical outcomes. These could include direct, patient-focused interventions as well as those implemented through pharmacists and medical care providers.
Importantly, we also observed that for each agent studied, among the strongest factors associated with consistent use was baseline use of other evidence-based medications. For each use of aspirin, β-blockers, and lipid-lowering therapy, the use of others within this group was associated with a greater likelihood of consistent use of the studied therapies. This most likely reflects characteristics involved in both physician prescription of evidence-based medications in general and patient compliance with a program of therapy. Conversely and not unexpected, use of coumarin and other antiplatelet agents such as ticlopidine, clopidogrel, and dipyridamole was strongly associated with lower use of aspirin. Because use of these agents represents only a relative contraindication to use of aspirin, this highlights an area of concern not only that may reflect perceived risk but also that persistent myths about concurrent use of these agents must be addressed to improve adherence to proven therapies such as aspirin.
Our study was based on patients predominantly from communities in central North Carolina and southern Virginia and therefore may not be representative of the entire US population. In addition, although the association of consistent use with lower mortality was particularly strong, the observational nature of our analyses and the presence of unmeasured confounders that may have contributed to the strength of association and survival bias introduced by requiring 2 consecutive surveys to determine use pattern must be considered. However, the direction of the association is consistent with that in randomized clinical trials of these agents, with the exception of ACEI use among patients without heart failure, which may reflect the inability to adjust for unmeasured confounders that affect both use of these agents and clinical outcomes. The changes in the strengths of association of consistent use with mortality when initially inconsistent users who had a terminal pattern of ≥2 consecutive reports of use were considered consistent users highlights the potential for confounding by unmeasured factors in these analyses. In this case, the reason for the new or renewed use may confound the analysis.
Medication use was collected by patient self-report. In an overview of 86 studies that reported both self-report (telephone interview, questionnaire, or diary) and non–self-report (administrative claims, pill count, plasma drug concentration, electronic monitors, or clinical opinion) measures of adherence, concordance varied widely by the type of measure used.22 Overall, concordance was high (κ>0.6 or Pearson correlation >0.8) in only 17% of comparisons of self-report measures with electronic non–self-report measures but in 58% of comparisons with other non–self-report measures. Questionnaires and diaries had significantly higher concordance with non–self-report measures than did telephone interviews. In a recent study in elderly patients, there was 81% congruence between self-reported use of specific cardiovascular or central nervous system medications and pharmacy records.23
Although patients may have reported a medication on the yearly follow-up forms, we did not assess actual day-to-day compliance with the reported therapy. However, in a previous survey of aspirin use among patients in the DDCD, we found that among patients reporting aspirin use, 85% reported daily use.24 Finally, we did not define “ideal” use populations. However, even best estimates of ideal use in general populations are much higher than those observed, suggesting that better understanding of differences between consistent users and those who are not may provide insights into improving long-term adherence.
From this database, we were unable to differentiate patient nonadherence from physician nonprescription. From previous studies, it is clear that an important factor in long-term use of a drug is prescription at the time of discharge after an acute event.6,16,17,19 In addition, we have shown that subsequent physician support for use of a drug is an important factor to patients in their long-term continuing use of aspirin.24 Our database also did not allow us to examine the influence of changes in risk factors over time or medication cost on adherence. Out-of-pocket expense has been shown to be an important predictor of long-term adherence to statin therapy,25 but among 11 health plans in the Council for Affordable Quality Healthcare, despite prescription coverage, only 45% of patients were adherent to β-blockers in the first year after myocardial infarction.26 Thus, areas for future focus include ongoing in-hospital and discharge prescription quality-improvement programs, patient education, and improved communication with patients and between hospital teams (nurses, physicians, and pharmacists) and outpatient providers (physicians and pharmacists) in the transition and maintenance of care and attention to understanding economic influences on long-term adherence to treatment for CAD.
Use of evidence-based therapies for CAD has improved but remains suboptimal. Although continued improvement in short-term use and discharge prescription of these agents is needed, considerable attention must also be focused on understanding and improving long-term adherence to achieve the full potential of these treatments to improve clinical outcomes.
This study was supported in part by the Agency for Healthcare Research and Quality, Centers for Education and Research on Therapeutics, grant No. U18HS10548.
Dr Newby received research grants from Bristol Myers Squibb, Sanofi, Schering Plough, and Novartis. Dr Allen LaPointe has received research grants from Eli Lilly, AstraZeneca, Reliant, Pfizer, GlaxoSmithKline, and Merck. Dr Kramer has received unrestricted educational grants from GlaxoSmithKline and AstraZeneca to increase β-blocker use in congestive heart failure and research grants from Pfizer for a systematic review of medical adherence and to study a method of increasing long-term adherence to evidence-based medicine for coronary artery disease. Drs DeLong and Mulhlbaier received unrestricted educational grants from GlaxoSmithKline and AstraZeneca to study β-blocker use in congestive heart failure. As Director of the Duke Clinical Research Institute, Dr Califf benefits indirectly from institutional grants and contracts from the following: Actelion, Ajinomoto, Alsius, Amgen, AstraZeneca, Aventis, Berlex, Biomarin, Boston Scientific, Bristol Myers Squibb, Cambridge Heart, Cardiodynamics, Centocor, Chase Medical, Corgentech, Daiichi, Eli Lilly, Enzon, Esai, Geneceutics, Genentech, GlaxoSmithKline, Guidant, Guilford Pharmaceuticals, InfraReDx, Medicure, Medivance, Medtronic, Merck, Millennium Pharmaceuticals, NABI, Novartis, NITROX, Ortho Biotech, Pfizer, Pharmanetics, Proctor and Gamble, Promethesus, Roche Holdings, Ltd, Salix, Sanofi, Schering Plough, St Jude Medical, Synaptic, The Medicines Company, Theravance, VDDI Pharmaceuticals, Vesicor, Vicuron, Wyeth-Ayerst, ASSLD, Abbott Labs, ACC, Acusphere, AHA, AHRQ, ANGES, ArgiNOx, Ark, BI, Biogen, Cardiac Science, CardioKinetix, Conor Medsystems, Corautus, Critical Therapeutics, Cubist Pharma, DoD, Dyax, Edwards Lifescience, Emory University, First Circle Medical, Flow Cardia, Inc, GE Healthcare, GE Medical Systems, HGS, HLR, IARS, IDB, Inhibitex, IONS, IU, KAI, KPS, Liebert, Lumen, McMaster Core Lab, Mycosol, NCCAM, NCI, NCSU, Neuron, NHLBI, BIA, NIAID, NIAMS, NICHD, NIDA, NIDDK, NIH, NIMH, Organon, Palo Alto CSCC, Pathway, Paul Tierstein, Rama Fdt, Reliant, SAMHSA, Scios, Sicel, SPRI, Terumo, UCSF, UTSW, Velocimed, Veridex, Vicuron, XOMA, Xsira, and XTL. Dr Califf also has equity in NITROX. The other authors report no conflicts.
The online-only Data Supplement, which contains an Appendix, can be found with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.105.505636/DC1.
Alexander KP, Peterson ED, Granger CB, Casas AC, Van de Werf F, Armstrong PW, Guerci A, Topol EJ, Califf RM, for the GUSTO IIb Investigators. Potential impact of evidence-based medicine in acute coronary syndromes: insights from GUSTO-IIb. J Am Coll Cardiol. 1998; 32: 2023–2030.
Rogers WJ, Canto JG, Lambrew CT, Tiefenbrunn AJ, Kinkaid B, Shoultz DA, Frederick PD, Every N. Temporal trends in the treatment of over 1.5 million patients with myocardial infarction in the US from 1990 through 1999: the National Registry of Myocardial Infarction 1, 2, 3. J Am Coll Cardiol. 2000; 36: 2056–2063.
Fonarow GC, French WJ, Parsons LS, Sun H, Malmgren JA, for the National Registry of Myocardial Infarction 3 Participants. Use of lipid-lowering medications at discharge in patients with acute myocardial infarction: data from the National Registry of Myocardial Infarction. Circulation. 2001; 103: 38–44.
Sanborn TA, Jacobs AK, Frederick PD, Every NR, French WJ, for the National Registry of Myocardial Infarction 3 and 4 Investigators. Comparability of quality-or-care indicator for emergency coronary angioplasty in patients with acute myocardial infarction regardless of on-site cardiac surgery (report from the National Registry of Myocardial Infarction). Am J Cardiol. 2004; 93: 1335–1339.
Bahit MC, Granger CB, Alexander KP, Kramer J, Allen-LaPointe NM, Califf RM. Applying the evidence: opportunity in US for 80,000 additional lives saved per year. Circulation. 2000; 102 (suppl II): II-873. Abstract.
Peterson ED, Parsons LS, Pollack CV, Newby LK, Littrell KA. Variation in AMI quality care across 1,085 US hospitals and its association with hospital mortality rates. Circulation. 2002; 106 (suppl II): II-722. Abstract.
Mukherjee D, Fang J, Chetcuti S, Moscucci M, Kline-Rogers E, Eagle KA. Impact of combination evidence-based medical therapy on mortality in patients with acute coronary syndromes. Circulation. 2004; 109: 745–749.
Mehta RH, Montoye CK, Gallogly M, Baker P, Blount A, Faul J, Roychoudhury C, Borzak S, Fox S, Franklin M, Freundl M, Kline-Rogers E, LaLonde T, Orza M, Parrish R, Satwicz M, Smith MJ, Sobotka P, Winston S, Riba AA, Eagle KA, for the GAP Steering Committee of the American College of Cardiology. Improving quality of care for acute myocardial infarction: the Guidelines Applied in Practice (GAP) Initiative. JAMA. 2002; 287: 1269–1276.
CRUSADE National Report. Available at http://www.crusadeqi.com/main/ecab/National_Report_Nov03.pdf. Accessed August 21, 2004.
American Heart Association. Heart Disease and Stroke Statistics–2004 Update. Available at http://www.americanheart.org/downloadable/heart/1079736729696HDSStats2004UpdateREV3–19–04.pdf. Accessed August 21, 2004.
Caskie GIL, Willis SL. Congruence of self-reported medications with pharmacy prescription records in low-income older adults. Gerontologist. 2004; 44: 176–185.
Kramer JM, Fetterolf D, Charde JP, Snyder R, Hammill B, DeLong E, Allen LaPointe N, Hoffman BS, Arrington RL, Peterson E. National evaluation of adherence to beta-blocker therapy for one year after acute myocardial infarction in patients with commercial health insurance. Circulation. 2004; 109: e291. Abstract.