Still up in the Air: Bias in Today's Commercially Supported CME

Way back in June, education expert Robert Cervero and graduate research assistant Jiang He, both of the University of Georgia, published their ACCME-funded examination of professional articles that addressed the relationship between industry support of CME and CME bias. Surprisingly enough, they found no directly relevant, published studies. However, they discovered exactly 2 original research articles, both of which predate the ACCME's 1992 Standards for Commercial Support, that examined the effect of commercially supported CME on physicians' prescribing practices.

In 1988, Bowman and Pearle published before-and-after prescribing patterns of 374 physicians who attended 3 industry-sponsored educational courses on beta blockers or calcium-channel blockers (CCBs). Physicians' prescription data were limited in that they were obtained by mailed self-reports, and before-and-after responses were not matched for 2 of the 3 courses. Also rates of survey returns (6-month "after" responses) were less than robust—from 40% to 60%.

Bowman and Pearle concluded that prescribing patterns generally favored the sponsoring company's drug. However, Cervero and He interpreted these results as mixed. They found that prescriptions for the CME sponsor's drug increased after course 1 (up 12%), but not as much as those for a competitor drug (up 17%). After course 2 (beta blockers), prescriptions for the sponsor's drug increased nonsignificantly, and after course 3, the sponsor's drug became the most frequently prescribed CCB (when compared with 2 other drugs).

Over at The Carlat Psychiatry Blog, Daniel Carlat—fervent critic of industry-supported CME—examined the Bowman-Peale study and, not surprisingly, maligns Cervero and He for not coming to a more emphatic conclusion that industry sponsorship of CME unduly influences prescribing patterns. Curiously enough, though, Carlat proposes that the mixed results seen after course 1—specifically an increase in prescriptions for a competitor drug—are due to the more effective CME efforts of the competitor (eg, course 3 and other CME activities).

This interesting proposal from Carlat actually supports the fact that industry-sponsored CME is not consumed in a vacuum. Carlat (perhaps unwittingly) bolsters the argument that a wide range of industry-sponsored CME, in conjunction with non-industry-sponsored CME and other countless sources of medical information, should be considered (if at all possible) when assessing influences on physicians' prescribing practices.

The other CME study cited by Cervero and He was published in 1992, and the CME activity in question featured a number of perks—for instance, an all-expenses-paid trip to a resort—that would not be compatible with the current standards of the ACCME. The study authors, by assessing hospital pharmacy data, determined that this more blatant marketing activity increased prescriptions for the sponsor's drug.

Cervero and He conclude their report by posing a number of questions to be addressed in the investigation of bias in commercially supported CME. The most important, paraphrased here, is How does the adoption of a sponsor's product (if influenced by CME) affect patient care?

ACCME = Accreditation Council for Continuing Medical Education; CME = continuing medical education.