An improved approach for confirmatory phase III population pharmacokinetic analysis

In contrast to the traditional extensive exploratory approach, the authors propose a confirmatory approach to phase III population pharmacokinetics for regulatory submissions. In their approach, they recommend a pre-specified primary analysis based on phase I/II data and phase III study design. Their approach also incorporates several specific sensitivity analyses to evaluate the robustness of the conclusions. According to statistical rationale, this approach eliminates certain biases that may occur in the estimated covariate effects, thereby precluding potentially unnecessary dosing adjustments and allowing for more accurate assessments of ambiguity in the results. Because exploration is vastly reduced, the analysis time is also substantially shortened. The proposed analyses are relatively easy to implement, although careful and prospective planning is required. Applications of the approach are provided, including 2 phase III analyses for regulatory submissions. Differences between the proposed approach and the commonly used extensive exploratory analyses submitted to regulatory agencies were small and consistent with practical expectations.