Power of 'phase 0' chronobiologic trials at different signal-to-noise ratios and sample sizes.

Clinical trials would gain from incorporating 'Phase 0' chronobiologic pilot designs both from the viewpoint of (statistical) power and cost-effectiveness. Herein, this statement is documented by power computations and is further illustrated by clinical examples answering specific questions. Power computations show the merits both of chronobiologic designs (that assign samples at equidistant intervals to cover one full cycle of anticipated pertinent rhythms) and of chronobiologic analyses (the cosinor versus the analysis of variance). Randomized clinical trials would gain from incorporating a concern for timing as well as dosing in all three stages of clinical trials (Phase I, II and III focusing on toxicity, efficacy and a comparison with the current best treatment, respectively) and could be cost-effectively preceded by 'Phase 0' trials so as to detect, sooner and with smaller sample sizes, desired or undesired effects that may otherwise be missed.