UCLA study: Nearly a third of drug ads in medical journals contain no references for medical claims

UCLA investigators reviewed pharmaceutical ads in American medical journals and found that nearly one-third contained no references for medical claims; while the majority of references to published material were available, only a minority of company data-on-file documents were provided upon request; and the majority of original research cited in the ads was funded by or had authors affiliated with the product’s manufacturer.

In one of the largest studies of its kind, UCLA investigators reviewed pharmaceutical ads in American medical journals to determine what materials are cited in support of medical claims and if those references are available to physicians. Researchers also determined the funding sources of research cited in the ads. The study is published in the Feb. 15 issue of the Canadian Medical Association Journal.

“Previous studies have shown that physicians’ drug prescribing is influenced by pharmaceutical ads. We wanted to see what documents were being used to substantiate the claims and how accessible these were to physicians who may want to verify the research findings,” said lead author Dr. Richelle Cooper, assistant clinical professor, Division of Emergency Medicine, David Geffen School of Medicine at UCLA.

Two UCLA investigators independently reviewed 438 ads from 10 American medical journals from 1999 and also reviewed a random sample of 400 references in journal articles from the same publications for comparison.

Researchers found that out of 438 pharmaceutical ads, 126 or 29 percent, offered no references in support of medical claims. The most commonly cited references in the ads included journal articles (55 percent) and data-on-file (19 percent), which is a reference to an unpublished company document. Other sources included books; prescribing information such as the Physician’s Desk Reference — which annually provides information on drug usage, warnings and drug interactions; government documents; or an internet site.

Reviewers were able to obtain 84 percent of the referenced documents in the ads. However, researchers could only obtain 20 percent of the data-on-file documents, despite repeated requests to the pharmaceutical companies. The companies responded to 37 out of 88 requests for data-on-file, but over half of the responses indicated that the information would not be provided due to company policy or because it was proprietary.

“We found that almost one-third of the pharmaceutical ads offered no references at all to support medical claims. In addition, most of the data-on-file documents in support of the medical claims were not available from the drug companies.” said Dr. David Schriger, study author and professor of Emergency Medicine at UCLA.

In comparison, out of the 400 journal articles, the most commonly cited sources of references included journal articles (88 percent) and books (8 percent). When reviewers tried to obtain the reference documents, 99 percent were available.

UCLA researchers also checked the funding sources of the original research cited in all the ads and of a random sample of 100 journal articles. Out of 294 ad references citing original research, 58 percent had research sponsored by a pharmaceutical company or had affiliated authors, 19 percent was funded by government or charitable sources and 23 percent had no funding statement at all.

In comparison, 8 percent of the original research reported in these journal issues was funded by a pharmaceutical company or had an author affiliated with the company, and 44 percent was funded by a government or charitable organization.

“We found that the majority of the original research cited in the ads to support medical claims was sponsored by pharmaceutical companies,” said Cooper.

Schriger notes that improved monitoring and access to references in pharmaceutical ads may be helpful to physicians and policy makers in making more informed prescribing decisions.

According to the authors, the Federal Drug Administration screens the drug ads for validity, but the sheer volume makes it difficult to adequately check many of the ads. For example, in 2002, the FDA screened over 34,000 ads.

Journals included in the study: American Journal of Psychiatry, Annals of Emergency Medicine, Annals of Internal Medicine, Annals of Surgery, Hospital Practice, Journal of the American Medical Association, New England Journal of Medicine, Neurology, Obstetrics and Gynecology and Pediatrics.