Propensity score methods in the surgical oncology literature

The use of propensity score methods in the surgical literature is increasing. Randomized, controlled clinical trials are the gold standard of medical research, allowing for accurate measurement and analysis of treatment effects. Use of propensity score methods allows researchers to mimic randomization when true randomization may not be possible. When used properly, these methods are a powerful tool for the medical researcher, allowing more rigorous conclusions to be drawn from retrospective data. With the increasing prevalence of propensity methods, it is important that these methods are used correctly, lest researchers be led to misleading conclusions based on poor statistical study design and analysis. The objective of this review is to analyze and evaluate the use of propensity score methods in the surgical oncology literature. We critique the current state of the use of propensity scores in the surgical oncology literature and offer recommendations to assure appropriate usage of propensity score methods. • Propensity score (PS) methods allows researchers to mimic randomization. • PS methods should only be used when both control and treatment groups are similar. • PS usage can increase the reliability of conclusions drawn from retrospective data. • PS methods must be used correctly, lest researchers be led to false conclusions.