Answers are grounded in this paper and cite their sources.
For late-stage clinical development, we propose combining phase 2 and 3 trials via a two-stage adaptive design. In the first stage, short-term safety and efficacy are examined, after which low doses that lack efficacy and high doses that cause safety concerns are eliminated from further evaluation. The trial continues to the second stage with doses that are not eliminated. For the second stage, the required sample size is adjusted to maintain power. All patients, including those enrolled in the first stage, are evaluated using a clinical endpoint requiring a longer follow-up. For the second stage, trend statistics are adaptively chosen based on the estimated dose-response curve in the clinical endpoint of the first-stage patients. At the end of the trial, pairwise statistics for the first stage and adaptive trend statistics for the second stage of the clinical endpoint are combined to establish dose-response and to identify the lowest effective dose. A notable feature of our proposed approach is that the adaptation rule governing dose selection, sample size calculation, and derivation of test statistics for the second stage need not be specified in advance to maintain the validity of the trial. The phase 2/3 combination design is effective in achieving robust statistical power and is also efficient, because the number of patients and time needed are substantially reduced.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2005) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: