Responsible reporting of medical research poses challenges for both journals and journalists to avoid misunderstandings that affect news coverage, say Dartmouth Medical School physicians.
Writing in the July 22 issue of the New England Journal of Medicine, they urge both the research and the media communities to anticipate confusion and consider the impact of their messages on the public. Drs. Lisa Schwartz, Steven Woloshin and H. Gilbert Welch of Dartmouth Medical School and the White River Junction VT, Veterans Affairs Medical Center explore how a medical study published this winter led to misleading reports and suggest ways to improve communications of research findings to the public.
The original study, which claimed to document evidence of sex and race bias in the evaluation of chest pain, appeared in the New England Journal of Medicine (February 25). The study was based on a controlled experiment in which primary care physicians watched video-taped interviews of "patients," portrayed by actors (two races, both sexes, and two ages) playing carefully-scripted hypothetical patient scenarios.
After seeing the "patients" on video and reviewing relevant medical information, doctors were asked whether they would recommend cardiac catheterization, in which a thin catheter is threaded into the heart for examination. The investigators then compared how often "patients" of different gender, race and age with identical symptoms and medical histories were referred for the invasive procedure.
The study's main message was that blacks and women were referred for catheterization 40% less often than whites and men. Responding in the New England Journal, the Dartmouth authors cite three problems with how the data were presented: the magnitude of the finding was overstated, the comparison reported was incorrect, and the implicit assumption -- that cardiac catheterization always represents the best care -- was unwarranted.
"These problems began with the way in which the authors chose to write the article, persisted despite peer and editorial review, and were magnified in the mass media," Schwartz, Woloshin and Welch write. " It is important to expose discriminatory practice in medicine. It is equally important not to exaggerate the prevalence of such practice," they add.
The confusion surrounding the results points to underlying problems in how medical research results are communicated to the public. Misleading interpretations could have been avoided by reporting absolute event rates, presenting risk ratios instead of odds ratios, not assuming that more medical care is always better medical care and being cautious about over generalizing. It is important, according to the authors, to provide realistic expectations.
First, the study summarized the relative chance of referral for blacks compared to whites using a statistical format called an odds ratio. Few people think in terms of odds or encounter them in daily life. Instead, they often mistakenly interpret odds to mean probability and to equate odds ratios with risk ratios. In this case, where the events under discussion are common (i.e., most patients were in fact referred) equating odds ratios with risk ratios leads to a substantial exaggeration of the findings. Thus, while the odds ratio of referral for blacks compared to whites is 0.6, the actual referral rates are similar: 85% of blacks compared to 91% of whites.
Second, the results were reported in terms of whites versus blacks and males versus females. In fact, the referral rates were identical for white men, white women and black men (all 91%); the only difference seen in referral rates was for black women (of whom 78% were referred for catheterization).
Third, the study implied cardiac catheterization was the best treatment and anything less was deemed inappropriate, an assumption the authors say is unfounded.
"It is tempting to place the blame for what went wrong in the coverage of the -- study on reporters, science writers, and media editors and to demand that they improve their reporting. To a large extent, however, the media simply reported what they were told," the authors conclude. "If the research community does not get it right, we cannot expect the media to do better."
Materials provided by Dartmouth Medical School. Note: Content may be edited for style and length.
Cite This Page: