Randomized controlled trials (RCTs) are widely considered the gold standard for inference on the comparative effects of 2 or more treatment strategies. Randomization is the most powerful technique available in causal inference, and the controlled setting of carefully designed experiments is a necessity in evaluating most treatments. Still, as currently designed, conducted, and analyzed, clinical trials are not without limitations and might be susceptible to some of the biases present in observational studies (1, 2). Strict inclusion and exclusion criteria, as well as trials often being conducted “in parallel” with regular clinical care, have led to concerns that the enrolled ...

References