A recent issue of The American Journal of Obstetrics and Gynecology reports that doctors have an inflated perception of the risks of hormone replacement therapy (HRT). The survey of Florida physicians found that primary care doctors were more likely to overestimate the risk than Ob-Gyns. The authors felt the respondents were confused between relative risk difference and absolute risk difference.
This is illustrated by data from the Women’s Health Initiative (WHI) study of estrogen and progesterone in which the relative risk numbers, when considered alone, exaggerate the perception of harm from HRT. This may have resulted in distorted media reporting of the study. For example, considering only the relative risks, there was a 29% increase in coronary heart disease events, a 26% increase in breast cancer, a 41% increase in stroke, and a doubling of pulmonary embolism. Unfortunately these figures are deceptive because they ignore the low baseline rate of events. Translated into absolute risk, HRT was associated with only 7 excess coronary heart disease events per 10,000 patient years, 8 more breast cancers, 8 more strokes, and 8 more PEs. These numbers represented the true magnitude of harm but seemed to be buried in the fine print of much of the popular reporting. An editorial in the same issue of JAMA cautioned against taking relative risk out of context, correctly pointing out that absolute risk data provided the true measure of effect. Although the editorial put the study in perspective it was given little notice in popular reporting.
Undue emphasis on relative risk difference is deceptive. Such emphasis can be used not only to exaggerate harm but also to inflate the apparent effectiveness of treatment, as is sometimes done in pharmaceutical company advertisements.
Why aren’t doctors more savvy about this issue? Maybe we should pay less attention to the popular news reports and concentrate more on the primary sources.