The pros and cons of propensity scores

Chest. 2012 Dec;142(6):1380-1382. doi: 10.1378/chest.12-1920.

Abstract

Propensity scores have been proposed as a method of equating groups at baseline, which is a problem, especially in studies that do not use randomization. This article discusses some difficulties with the technique that may jeopardize the findings if users (and readers) are not aware of these problems.

MeSH terms

  • Data Interpretation, Statistical
  • Humans
  • Patient Selection
  • Propensity Score*
  • Randomized Controlled Trials as Topic*
  • Severity of Illness Index