MedNexus
2024年 · 第62卷第09期
MedNexus
randomized controlled trials (RCTs) are regarded as the gold standard for evaluating drug safety and efficacy, but it is sometimes difficult to conduct RCT studies due to high cost, difficult patient recruitment, ethical restrictions and other reasons. New research designs such as externally controlled single-arm trials and simulated target trials with the help of real world data (RWD) are attracting increasing attention. RWD studies especially need to pay attention to the adjustment and control of confounding. Traditional confounding control methods include matching, stratification, and incorporating suspicious confounding factors as covariates into regression analysis. However, when the number of covariates increases, the implementation of these methods becomes extremely difficult. Against the above background, in recent years, the propensity scoring method has gradually become an important method to carry out confounding control in RWD-based research, so as to achieve similar "randomization" effect.
本期目次

