The authors of this web content have examined a large number of randomized clinical trials and found that the majority of trials have much lower statistical power than what is commonly assumed. This means that “statistically significant” results often overestimate treatment effects, while “nonsignificant” results may actually correspond to important effects. To address this issue, the authors propose reinterpreting the P value in terms of a reference population of studies. They also discuss the use of a shrinkage estimator, which they find to be more efficient than the unbiased estimator typically used. The authors emphasize the importance of incorporating prior information and the need to learn from previous trials.
https://statmodeling.stat.columbia.edu/2023/12/23/bayesians-moving-from-defense-to-offense-i-really-think-its-kind-of-irresponsible-now-not-to-use-the-information-from-all-those-thousands-of-medical-trials-that-came-before-is-that-very/