Propensity scores: Methods, considerations, and applications in the Journal of Thoracic and Cardiovascular Surgery.

Abstract

To review the published literature using propensity scoring, describe shortcomings in the use of this technique, and provide conceptual background for understanding and correctly implementing studies that use propensity matching. We survey the published statistical literature and make recommendations for a set of standard criteria for studies that use propensity matching. We evaluated adherence to these criteria in recent publications in the Journal of Thoracic and Cardiovascular Surgery and determined how well the standards were applied. We found that studies that use propensity matching are rarely documented well enough to be convincing in their results. When documentation is available, statistical shortcomings are common. Improved statistical practice is needed when using propensity scoring. This article suggests standard criteria for using this method in Journal publications.