A new study by researchers at the Agency for Healthcare Research and Quality, Yale University, and Johns Hopkins University suggests that when medical trial results are released prior to journal publication, doctors can change their practice dramatically. However, the changes may not be in line with the detailed research results later published in a peer-reviewed journal and may be harmful to some patients.
Yale University's Cary Gross, M.D., who led the study, points out that a real balance must be struck between the public's demand for rapid information and publication in medical journals, an inherently slower process. Dr. Gross notes that health policymakers, medical journal editors, scientists, and the press have long debated whether it is in the public's best interest to release the results of clinical trials prior to publication in medical journals.
According to AHRQ researcher Claudia Steiner, M.D., a coauthor of the study, physicians may be able to judge how to use a new medical treatment better after they have read all the details in the full report published in the medical literature. Future research could focus on how clinical alerts—the method used for alerting clinicians early to trial findings—might be structured to preserve their advantage while avoiding any potential downside.
The researchers tracked the use of carotid endarterectomy (CEA)—a surgical procedure for clearing a diseased carotid artery in stroke-threatened patients—after the National Institutes of Health (NIH) disseminated the results of two clinical trials prior to journal publication. Each of the trials had been halted early because potentially life-saving benefits of CEA were found. Rather than waiting for the studies to be published in the medical literature, the NIH expedited dissemination of the results to physicians by means of clinical alerts. These alerts explained the findings of the studies, but also cautioned doctors that the trials included only patients under 80 years of age. NIH also cautioned that the trials had been conducted at medical centers with documented expertise in CEA, a highly complex procedure.
Immediately after the alerts were released, there was a substantial increase in CEA use in the States studied. The adjusted CEA use rate increased roughly 3 percent per month over the 6-month period following the issuance of the first alert in 1991. The rate of increase declined to only 0.5 percent per month after the clinical trial's findings were published later that year. Following the release in late 1994 of the clinical alert on the second CEA trial, the procedure's rate of use increased 7.3 percent per month over the next 7 months, but after the findings were published in May 1995, there was a decrease in the rate of use of 0.44 percent per month.
The study also found that the use of CEA following the second clinical alert was greater among patients over 80 years of age, despite the fact that these patients would not have been eligible for either trial because of their higher risk of complications and death. In contrast, following publication of the results in a medical journal, there was a greater decrease among those 80 and older, compared with patients under 80 years of age. The researchers also found that many patients were referred to hospitals with less experience in the use of CEA, despite the warnings of the clinical alerts.
The study used data from the State Inpatient Databases (SID) of California, Colorado, Florida, Illinois, New York, Pennsylvania, and Wisconsin. SID is part of the Healthcare Cost and Utilization Project (HCUP), a family of powerful State and national hospital databases built in partnership with 22 States and AHRQ.
For more information, see "Relation between prepublication release of clinical trial results and the practice of carotid endarterectomy," by Drs. Gross and Steiner, Eric Bass, M.D., and Neil Powe, M.D., in the December 13, 2000 Journal of the American Medical Association 284(22), pp. 2886-2893.
Reprints (AHRQ Publication No. 01-R017) are available from the AHRQ Publications Clearinghouse.
Return to Contents
Proceed to Next Article
540 Gaither Road Rockville, MD 20850