Relation between prepublication release of clinical trial results and the practice of carotid endarterectomy

Context: Little is known about how clinical practice is affected by disseminating results of clinical trials prior to publication in peer-reviewed journals. Objective: To determine whether prepublication release of carotid endarterectomy (CEA) trial results via National Institutes of Health Clinical Alerts was associated with prompt changes in patient care that were consistent with the new medical evidence. Design, Setting, and Patients: Longitudinal data series analysis using acute care hospital discharge data from the Healthcare Cost and Utilization Project for patients who had CEA performed in acute care hospitals in 7 states (New York, California, Pennsylvania, Florida, Colorado, Illinois, and Wisconsin). The trials were the North American Symptomatic Carotid Endarterectomy Trial (NASCET clinical alert released February 1991) and the Asymptomatic Carotid Atherosclerosis Study (ACAS clinical alert released September 1994). Main Outcome Measure: Carotid endarterectomy rate during each month from 1989 (2 years before the NASCET clinical alert) to 1996 (2 years after the ACAS clinical alert), adjusted for age and sex. Because both trials were limited to patients 80 years or younger in hospitals with low mortality, we also stratified CEA rates by patient age and hospital mortality rate. Results: From 1989 through 1996, 272849 CEAs were performed in the acute care hospitals in these 7 states, with the annual number increasing from 22300 to 51495. After the NASCET clinical alert, the adjusted CEA rate increased 3.4% per month (95 % confidence interval [CI], 1.6%-5.3%) during the following 6 months and then increased 0.5% per month (95% CI, 0.2%-0.8%; P<.04) after journal publication of the NASCET study. After the ACAS clinical alert, the CEA rate increased 7.3 % per month (95% CI, 6.0%-8.5%) during the following 7 months and then decreased by 0.44% per month (95% CI,-0.86% to -0.0002%; P<.04)after journal publication of the ACAS study. After the ACAS clinical alert, the CEA rate increased more in patients aged 80 years or older than in younger patients; whereas, after journal publication of ACAS, the CEA rate decreased more rapidly in the older population. The overall proportion of CEAs performed in low-mortality hospitals did not change substantially after release of the clinical alerts or after journal publication. Conclusion: In this study, prepublication dissemination of CEA trial results with clinical alerts was associated with prompt and substantial changes in medical practice, but the observed changes suggest that the results were extrapolated to patients and settings not directly supported by the trials.