Key points are not available for this paper at this time.
Propensity scores are one useful tool for accounting for observed differences between treated and comparison groups. Careful testing of propensity scores is required before using them to estimate treatment effects.
Garrido et al. (Wed,) studied this question.