All Times EDT
Keywords: Go/No-go, Precision, ANCOVA, Unequal Covariances
Many accounts exist on using quantitative approaches for decision in proof-of-concept (PoC) trials. However, these accounts are silent on impact of unequal covariances and post-treatment variances for continuous endpoints. Further, power, precision, and probability of indeterminate decision—these are some of the approaches for sample size computation in PoC trials. The choice of approach is driven by preference, experience, or precedence among other factors, e.g., ongoing effervescent debate on using p-value for decision. In this session, the following question will be addressed: 1) Are the approaches for sample size computation for PoC trials equivalent? 2) Are there advantages of selecting an approach versus the other for sample size computation? 3) Does unequal covariance and post-treatment variances affect quantitative decision? 4) Does ignoring correlation between a baseline covariate and an outcome rob the study team of ability to minimize uncertain decision?