Centrum—Spinning the Vitamins?
Scott K. Aberegg, MD, MPH, has written an amusing and interesting blog about a recently published randomized controlled trial (RCT) on vitamins and cancer outcomes[1]. In the blog, he critiques the Physicians’ Health Study II and points out the following:
- Aberegg wonders why, with a trial of 14,000 people, you would adjust the baseline variables.
- The lay press reported a statistically significant 8% reduction in subjects taking Centrum multivitamins; the unadjusted Crude Log Rank p-value, however, was 0.05—not statistically significant.
- The adjusted p-value was 0.04 for the hazard ratio which means that the 8% was a relative risk reduction.
- His own calculations reveals an absolute risk reduction of 1.2% and, by performing a simple sensitivity analysis—by adding 5 cancers and then 10 cancers to the placebo group—the p-value changes to 0.0768 and 0.0967, demonstrating that small changes have a big impact on the p-value.
He concludes that, “…without spin, we see that multivitamins (and other supplements) create both expensive urine and expensive studies – and both just go right down the drain.”
A reminder that, if the results had indeed been clinically meaningful, then the next step would be to perform a critical appraisal to determine if the study were valid or not.
Reference
[1] http://medicalevidence.blogspot.com/2012/10/a-centrum-day-keeps-cancer-at-bay.html accessed 10/25/12.
[2] Gaziano JM et al. Multivitamins in the Prevention of Cancer in Men The Physicians’ Health Study II Randomized Controlled Trial. JAMA. 2012;308(18):doi:10.1001/jama.2012.14641.