[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Views 473
Citations 0
Comment & Response
January 22, 2019

Reforming Nutritional Epidemiologic Research—Reply

Author Affiliations
  • 1Meta-Research Innovation Center at Stanford, Stanford University, Stanford, California
JAMA. 2019;321(3):310. doi:10.1001/jama.2018.18676

In Reply The calculation of 12 years of life gained with the consumption of 12 hazelnuts daily (and similar calculations with other foods) was an argumentum ad absurdum.1 The estimation combined several prevalent assumptions and common practices to demonstrate how absurd inferences can arise. These assumptions are that nutritional epidemiology–derived hazard ratios estimate causal effects; and these causal estimates can be extrapolated across the life span. These assumptions may be false, but nevertheless they are held by several scientists and widely disseminated by media. Moreover, a common practice in multiple fields is to report almost exclusively RRs without providing absolute risks.2 Almost all nutritional epidemiology studies, and even the best meta-analyses such as the one done by Schwingshackl et al,3 report RRs, not the far more informative absolute risks. It is not uncommon to read in scientific articles (and in the media) that some factor reduced all-cause mortality by 15% or improved survival by 15%. A naive calculation by a nonexpert, based on a baseline life expectancy of 80 years, might produce an estimate of (80 × 15%) 12 years of life gained. False assumptions combined with poor reporting and misleading dissemination propagate absurdity.