[Skip to Content]
[Skip to Content Landing]
Views 496
Citations 0
Comment & Response
June 2017

When to Adjust for Potentially Confounding Variables

Author Affiliations
  • 1Institute of Organization and Global Management Studies, Johannes Kepler University Linz, Linz, Austria
JAMA Intern Med. 2017;177(6):891-892. doi:10.1001/jamainternmed.2017.1071

Both researchers and publishers deserve to celebrate when an article attracts such widespread public attention as “Association of Long-term, Low-Intensity Smoking With All-Cause and Cause-Specific Mortality in the National Institutes of Health–AARP Diet and Health Study”1 about the impact of low-intensity smoking on mortality by Inoue-Choi et al and published in a recent issue of JAMA Internal Medicine. On the other hand, flaws of high-profile studies are more consequential under such wide circulation, read by scholars from other fields, and, especially, if the conclusions of the study may influence the behavior of readers. I suspect a serious deficiency in the regression analysis, more specifically the Cox hazard-rate estimation, performed by Inoue-Choi and colleagues,1 that may threaten their main conclusions.

×