2021/04/21 by Alejandro Schuler, Schuler, Alejandro · 1 citation
Mathematics · #Advanced Causal Inference Techniques #Applications (stat.AP) #FOS: Computer and information sciences #Methodology (stat.ME) #Statistical Methods and Bayesian Inference #Statistical Methods in Clinical Trials
paper · pdf · doi:10.48550/arxiv.2104.10784
openalex publication_date 2021/04/21 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Trials enroll a large number of subjects in order to attain power, making\nthem expensive and time-consuming. Sample size calculations are often performed\nwith the assumption of an unadjusted analysis, even if the trial analysis plan\nspecifies a more efficient estimator (e.g. ANCOVA). This leads to conservative\nestimates of required sample sizes and an opportunity for savings. Here we show\nthat a relatively simple formula can be used to estimate the power of any\ntwo-arm, single-timepoint trial analyzed with a semiparametric efficient\nestimator, regardless of the domain of the outcome or kind of treatment effect\n(e.g. odds ratio, mean difference). Since an efficient estimator attains the\nminimum possible asymptotic variance, this allows for the design of trials that\nare as small as possible while still attaining design power and control of type\nI error. The required sample size calculation is parsimonious and requires the\nanalyst to provide only a small number of population parameters. We verify in\nsimulation that the large-sample properties of trials designed this way attain\ntheir nominal values. Lastly, we demonstrate how to use this formula in the\n"design" (and subsequent reanalysis) of a real clinical trial and show that\nfewer subjects are required to attain the same design power when a\nsemiparametric efficient estimator is accounted for at the design stage.\n