AirTalk for October 24, 2012

Many medical study results really are too good (or too bad) to be true

Justin Sullivan/Getty Images

A pharmacist pours Truvada pills back into the bottle at Jack's Pharmacy on November 23, 2010 in San Anselmo, California. A study published by the New England Journal of Medicine showed that men who took the daily antiretroviral pill Truvada significantly reduced their risk of contracting HIV.

If you happen to come across medical study results that claim a treatment has a “very large effect,” those results are likely either exaggerated or flat out wrong, according to researchers at Stanford University’s School of Medicine.

A statistical analysis of nearly 230,000 trials led by Dr. John Ioannidis, published today in the Journal of the American Medical Association, found that 90 percent of studies showing "very large effects” in initial reports on medical treatments are less effective or nonexistent when additional trials are conducted. Dr. Ioannidis attributes this misleading trend to a variety of causes including the fact that many studies’ sample sizes are too small and that the results are often based on intermediate effects only.

How surprising is it that the allegedly dramatic effects achieved by many medical treatments are exaggerated or false? What can medical professionals do differently to avoid making false claims about treatments?

Guest:

Dr. John Ioannidis, MD, Professor of Medicine and Health Research & Policy, Stanford University's School of Medicine; Co-author, "Empirical Evaluation of Very Large Treatment Effects of Medical Interventions," published in The Journal of the American Medical Association this week

Dr. Ivan Oransky, MD, Executive Editor, Reuters Health; teaches medical journalism at New York University’s Science, Health, and Environmental Reporting Program; co-creator of the blog Retraction Watch focused on retractions of studies in science journals


blog comments powered by Disqus