Traditional statistical methods used to address issues with external validity all have limitations. Subgroup analyses may demonstrate that study results are upheld, but they are often underpowered to make such claims and are prone to type II error. Alternatively, sensitivity analyses may show a large subgroup effect size, needed to alter the inference from the study population, but these are estimates and not based on actual patient data. Poststratification, or adjusted analysis, requires careful a priori selection of covariates or it risks introducing bias. Follow-up observational studies are sometimes advocated, but they are time-consuming, expensive, and can delay dissemination of effective therapies or recognition of harm.