[Skip to Navigation]
Sign In
Users' Guides to the Medical Literature
January 22/29, 2014

How to Use a Subgroup Analysis: Users’ Guide to the Medical Literature

Author Affiliations
  • 1Chinese Evidence-Based Medicine Center, West China Hospital, Sichuan University, Chengdu, Sichuan, Chinaj
  • 2Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada
  • 3Stanford Prevention Research Center, Department of Medicine, Stanford University School of Medicine, Stanford, California
  • 4Stanford Prevention Research Center, Department of Health Research and Policy, Stanford University School of Medicine, Stanford, California
  • 5Department of Statistics, Stanford University School of Humanities and Sciences, Meta-Research Innovation Center at Stanford (METRICS), Stanford, California
  • 6Heart Failure and Transplantation Program, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada
JAMA. 2014;311(4):405-411. doi:10.1001/jama.2013.285063

Clinicians, when trying to apply trial results to patient care, need to individualize patient care and, potentially, manage patients based on results of subgroup analyses. Apparently compelling subgroup effects often prove spurious, and guidance is needed to differentiate credible from less credible subgroup claims. We therefore provide 5 criteria to use when assessing the validity of subgroup analyses: (1) Can chance explain the apparent subgroup effect; (2) Is the effect consistent across studies; (3) Was the subgroup hypothesis one of a small number of hypotheses developed a priori with direction specified; (4) Is there strong preexisting biological support; and (5) Is the evidence supporting the effect based on within- or between-study comparisons. The first 4 criteria are applicable to individual studies or systematic reviews, the last only to systematic reviews of multiple studies. These criteria will help clinicians deciding whether to use subgroup analyses to guide their patient care.

Add or change institution