Accounting for Interim Analyses in the Assessment of Program Expected Net Present Value
- PDF / 9,990,914 Bytes
- 10 Pages / 612 x 792 pts (letter) Page_size
- 99 Downloads / 193 Views
Jeffrey Helterbrand, PhD Senior Director, Clinical Biostatistics and Epidemiology, Genentech. Inc.. South San Francisco, California
Hal Borron, MD, FACC Senior Vice President. Development, and Chief Medical Officer, Genentech. Inc., South San Francisco, California
The calculation of the expected net present value of a clinical development program invdves the integration of future outcomes with their timing, cash flows, and associated probabilities, both subjectbe and relative fiequency. As applied to clinical trials, the outcomes are the clinical trial results, typically timed at the completion of the trial. When the trial design includes interim analyses with the possibility of early trial stopping with an efficacyor futility conclusion, we show that the expected net present value of the program can be mean-
Key Words Interim analysis; Expected net present value; Portfolio management
Correspondence Address Lee Kaiser, Genentech. Inc..
I DNA Way, M S 452a, South San Francisco, CA 94080 (email: [email protected]).
INTRODUCTION Portfolio management is commonly applied in the biopharmaceutical industry to manage the complex drug development decisions required to balance costs, risks, timing, options, revenue, and fit with corporate strategy (1-3). Typically, at the end of each clinical phase of a drug development program, future program timing, costs, risks, revenue, and ultimately the expected net present value (NPV) of the program are updated to facilitate decision making regarding the value of continued development (Figure 1A). Phase 3 clinical trials can be lengthy and may include planned interim analyses. Especially in indications with a survival endpoint, such as oncology, early positive trial results can lead to a marketing application, and this possibility can be added to the decision tree. In addition, many such trials also include a futility bound in the interim analysis plan, and the possibility of early negative results can also be added (Figure 1B). Further, as the trial progresses and interim analyses are passed without stopping the trial, information about the study drug’s treatment effect is gained, even without knowledge of the actual accrued data, and the probability of technical success (PTS) of the trial can be updated,
i n & & influenced through the inclusion of early stopping times and probabilities in the expected net present value calculations. This approach can be applied prior to the start of the trial, and it can also be applied during the course of the trial. In the latter case, given that interim analyses have been passed, the trialstopping probabilities change, with resulting eflects on the expected net present value. In both cases, the improved expected net present value figures should lead to better portfdio decisions.
leading to a revised expected NPV of the program. Many, if not most, of the inputs into portfolio management calculations are subjective and associated with a reasonable amount of uncertainty. By contrast, the probability of stopping a trial at an interim analysis
Data Loading...
