Adapting by calibration the sample size of a phase III trial on the basis of phase II data

Abstract

The problem of estimating the sample size for a phase III trial on the basis of existing phase II data is considered, where data from phase II cannot be combined with those of the new phase III trial. Focus is on the test for comparing the means of two independent samples. A launching criterion is adopted in order to evaluate the relevance of phase II results: phase III is run if the effect size estimate is higher than a threshold of clinical importance. The variability in sample size estimation is taken into consideration. Then, the frequentist conservative strategies with a fixed amount of conservativeness and Bayesian strategies are compared. A new conservative strategy is introduced and is based on the calibration of the optimal amount of conservativeness - calibrated optimal strategy (COS). To evaluate the results we compute the Overall Power (OP) of the different strategies, as well as the mean and the MSE of sample size estimators. Bayesian strategies have poor characteristics since they show a very high mean and/or MSE of sample size estimators. COS clearly performs better than the other conservative strategies. Indeed, the OP of COS is, on average, the closest to the desired level; it is also the highest. COS sample size is also the closest to the ideal phase III sample size M(I) , showing averages and MSEs lower than those of the other strategies. Costs and experimental times are therefore considerably reduced and standardized. However, if the ideal sample size M(I) is to be estimated the phase II sample size n should be around the ideal phase III sample size, i.e. n ≥2M(I) /3.

Profit Evaluations When Adaptation by Design Is Applied

“…The profit model giving equation 6 can, of course, be improved. For example: safety considerations can be introduced in the analysis of phase II results for the eventual launch of phase III; also, the time on market may continue after the duration of the patent life.…”

“…When these parameters are appropriately set, CSSE can provide high overall success probability (OSP; the success probability of phase II and III trials) and suitable precision in planning phase III. 1,3,6,7 CSSE is based on the assumption that the response variable and the ES are the same in phases II and III. Apart from cancer trials, where the endpoints of phase II and III are usually different, there are no clinical area restrictions to apply CSSE.…”

Robustness and Corrections for Sample Size Adaptation Strategies Based on Effect Size Estimation

“…This strategy is considered because it is the simplest, and also because the results regarding the estimation of the sample size under Scenario 1 were quite good (De Martini, 2010).…”

“…This strategy is based on the fact that if a fixed-γ conservative strategy is adopted, then its OP (say OP n (γ)) turns out to be a concave function of γ. Consequently, we introduced an optimization for OP n (γ) (De Martini, 2010), with the constraint of not exceeding 1 − β. Being that optimum (say γ O,n ) unknown, it can be estimated through the plug-in principle and its estimate is g O,n .…”

“…In a recent work we stated (De Martini, 2010) that phase II sample sizes n between M I /3 and 4M I /3 were the most important from a practical point of view, and that higher values of n were evaluated to look at the asymptotic behavior of SSE strategies only. Now, it should be noted that under Scenarios 2-4 sample size estimation is not…”