Trends in the use of pharmacotherapies for acute myocardial infarction among physicians who design and/or implement randomized trials versus physicians in routine clinical practice: The MILIS-TIMI experience

Background: Although studies have documented that randomized, controlled trials (RCTs) have a measurable influence on clinical practice, investigators have uncovered important deficiencies in the application of RCT findings to the management of acute myocardial infarction (AMI). Little is known about the extent to which physicians who design and/or implement clinical trials differ from physicians in routine practice in their translation of the literature. Methods: Our aims were to (1) evaluate recent trends in selected treatments of AMI in relation to the publication of RCTs, statistical overviews, and task-force guidelines, and (2) compare prescribing practices in AMI management between physicians in routine clinical practice and physicians who design and/or implement RCTs. We reviewed the use of aspirin, β-blockers, angiotensin-converting enzyme inhibitors, and calcium channel blockers on entry and at discharge n patients enrolled in the MILIS, TIM 1, 2, 4, 5, 6, and 9B trials with ST-elevation (and depression in MILIS) myocardial infarction for a period approaching 2 decades (August 1978 to September 1995). We hypothesized that physicians who participate in RCTs apply the findings of the published literature more promptly and thoroughly than physicians in routine practice. Results: Use of aspirin, β-blockers, and angiotensin converting enzyme inhibitors exhibited a statistically significant time-related increase at discharge and, excepting β-blockers, at enrollment across the trials. Prescription of calcium channel blockers showed a statistically significant decrease at discharge only. For all medications under study, increases and earlier and more decreases in use associated with publication of clinical data occurred earlier and more steeply for the discharge cohort (prescriptions by physicians participating in RCTs) than for the enrollment cohort (prescriptions by physicians in routine practice). Recent prescribing practices (1994 to 1995) among RCT investigators and their colleagues have higher concordance with published findings than those of physicians in routine practice. Conclusions: Physicians who design and/or implement RCTs translate the results of the medical literature more promptly and to a greater extent than physicians in routine clinical practice. Differences between different physician classes need to be studied further amid efforts to reconfigure health care delivery that currently favor more dominant roles for primary care physicians.