Integrated Planning of Drug Development Programs
Software Engineering and Programming Languages
Final Report Abstract
In the previous research project, methods to design the integrated planning of pilot and subsequent confirmatory studies in clinical research in an optimal way were developed. This was done within a utility-based framework using the example of phase II/III drug development programs, consisting of a phase II trial which is, in case of promising results, followed by (at least) one phase III trial. The utility function takes costs of the program (fixed and variable per-patient costs for phase II and III), expected benefit when launching the product successfully on the market, and the development risk (expected probability of a successful program) into account. By optimizing over the threshold value for the go/no-go decision rule and the phase II sample size, optimal decision rules and sample size allocations can be found. The achievements of the research project include methods for optimal phase II/III drug development designs for different endpoints (time-to-event, binary and normally distributed). A wide range of possible drug development program scenarios is covered by including methods for programs with 1.) several phase III trials, 2.) multiple arms, and 3.) multiple endpoints. To counteract the selection bias, that is the overestimation of the treatment effect after a go decision, different adjustment methods have been incorporated into the framework (4.). The planning of optimized phase II/III drug development programs with utility functions requires a number of input parameters from many different areas (medical, commercial, resources, ...). In order to facilitate the practical implementation of the previous introduced approaches, the framework was implemented in the programming language R. The aim of the current research project was to ensure the quality of the research software developed by making it usable in the long term. By implementing the above mentioned extensions 1.-4. for time-to-event, binary, and normally distributed outcomes and merging all software aspects into a single modular program package (“drugdevelopR”), utilization is facilitated and the methods can be further adapted for specific scenarios. Furthermore, as regulatory authorities require the validation of any software used to produce records within clinical research, a sophisticated quality assurance concept consisting of measures for archiving, versioning, bug reporting, and code documentation was followed, guaranteeing reliable results.
Publications
-
drugdevelopR: Utility-Based Optimal Phase II/III Drug Development Planning. CRAN: Contributed Packages. The R Foundation.
Erdmann, Stella; Cepicka, Johannes; Kirchner, Marietta; Kieser, Meinhard & Sauer, Lukas D.
