Authors' objectives:

The overall objective of this Special Report is to examine how the publication of new evidence influences physician prescribing behavior. For the purpose of this Special Report, new evidence will refer primarily to publication of clinical trial results, and/or composite evidence reports such as technology assessments or evidence-based guidelines. There are several ways in which physicians are exposed to new evidence. They may read medical journals, attend scientific meetings at which results are presented, or hear about this information through the mass media. This Special Report will seek studies that examine changes in physician prescribing over time in response to the publication of evidence in any of these formats.

Authors' results and conclusions: The evidence demonstrates that clinical trial results can have an effect on prescribing patterns, but that this effect is not universal, and there is a large amount of variability in the impact of new evidence. Some studies show a large impact on prescribing patterns, some show a modest impact, and others show no impact at all. This body of evidence builds on prior research in hypothesizing certain factors that are associated with the degree of impact of new evidence: - results that suggest harm; - clinical trials that are definitive, i.e., large, well-conducted, and free of major bias; - results that have clear implications for clinical practice; - results in which dissemination is facilitated, due to high-profile nature of the study, media attention, or dissemination by the scientific community; - the strength and direction of underlying secular trends prior to publication of new evidence. Among the studies evaluated, the largest impact was demonstrated for hormonal replacement therapy prescriptions following publication of results from the Womens Health Initiative (WHI) in 2002. In this case, there was a definite reversal of a trend toward increasing hormone-replacement therapy (HRT) prescriptions, and a substantial reduction in overall prescribing rates. In this situation, the publication of new evidence included all of the factors that may be associated with a larger impact. The WHI results suggested that there was net harm associated with prescription of HRT for postmenopausal women. The trial was a large, rigorous, and randomized, controlled trial (RCT) that represents the most definitive evidence to date on this issue. As a result, the results of this trial had clear implications for clinical practice. Also, the trial was very high-profile, the results were widely anticipated for a long time prior to publication, and release of results was associated with a high degree of media coverage. These factors aided in the widespread, rapid dissemination of results. Of the 3 trials that evaluated hypertensive medications, the publication of ALLHAT results for alpha-blockers had the greatest impact. These results also suggested harm, with an increase in congestive heart failure for patients prescribed alpha-blockers. This was also the largest trial of antihypertensive medications performed to date, and it was performed with rigorous methodology. For the other trials of hypertension, the effect was modest or not significant. Although these trials were also rigorous RCTs, they had a lower profile and less clear implications for clinical practice when compared with ALLHAT. For example, 1 study tested whether results of 3 hypertension trials were associated with greater prescribing of diuretics and beta-blockers and less prescribing of calcium-channel blockers and angiotensin-converting-enzyme (ACE) inhibitors. While the trials demonstrated a benefit of diuretics and beta-blockers, these medications were not clearly demonstrated to be superior to the other medications, nor was there any suggestion of harm associated with the other antihypertensive agents. As a result, the results of these trials did not have clear implications for clinical practice. For the study that evaluated statin use, the clinical trials were large, rigorous, and the results had clear implications for clinical practice. However, the prevailing secular trend was for rapid increases in the prescribing rates for statins, indicating that physicians were adopting these medications prior to the clinical trial results. The trial results confirmed that the use of these medications was beneficial, and the trend toward increasing use continued after publication of the evidence. The final 2 studies evaluated the use of other cardiovascular medications. One study demonstrated a modest increase in aspirin use following the publication of ISIS-2. This trial was a large, definitive RCT confirming the benefits of aspirin following myocardial infarction (MI). The impact of the trial may have been limited by a lack of dissemination efforts through the media or through the pharmaceutical industry. The other study demonstrated a positive impact of 2 large, rigorous RCTs on prescribing of ramipril and spironolactone. Both RCTs demonstrated a mortality benefit for these drugs, and, therefore, had clear implications for clinical practice. The ramipril trial was associated with aggressive dissemination efforts by the pharmaceutical industry that may have facilitated awareness of the results and promoted adoption of ramipril use. In contrast, there was little promotion of spironolactone use following the publication of the clinical trial, and the increase in spironolactone use was less. The positive change in spironolactone use was evident, however, given that there was low use of this drug prior to the clinical trial and only a very weak secular trend toward increasing use.

Authors' recommendations: Health plans and other organizations may be interested in the findings of this research to better direct their efforts toward promotion of evidence-based prescribing. The evidence suggests that facilitation of dissemination of evidence may be effective in changing prescribing behavior. Assessment of the level of interest in research questions and anticipation of publication of evidence that addresses these questions can help health plans decide where to focus their efforts. Trials that demonstrate harm associated with certain medications appear to be the type of evidence that has the greatest impact on prescribing. Also, knowledge of prevailing secular trends in medication prescribing is important in estimating the potential impact of new evidence. Finally, it is likely prudent to avoid promotion of evidence that does not have clear implications for prescribing in clinical practice.

Authors' methods: Review

Project Status: Completed

URL for project: http://www.bcbs.com/blueresources/tec/contact-tec.html

Year Published: 2004

English language abstract: An English language summary is available

Publication Type: Not Assigned

Country: United States

Organisation Name: BlueCross BlueShield Association

Contact Address: BlueCross BlueShield Association, Technology Evaluation Center, 225 North Michigan Ave, Chicago, Illinois, USA. Tel: 888 832 4321

Contact Name: tec@bcbsa.com

Contact Email: tec@bcbsa.com

Copyright: BlueCross BlueShield Association (BCBS)