Claims of differential treatment effects were largely rated as “low credibility” or “very low credibility.” Subgroup analyses in cancer trials may not be credible, and their results should be ...