Optimizing Dose Selection With Respect to Multiple Safety/Efficacy Endpoints Using Clinical Utility Concepts
Roy, A. and M. Pfister
Bristol-Myers Squibb, Princeton, NJ
Objectives: An important aspect of dose selection for late phase studies is an evaluation of benefit/risk profiles of candidate dosing regimens. This evaluation involves the consideration of multiple safety and efficacy endpoints, the relative importance of which is subjective, and depends upon the opinion of experts who do not always concur. The objective of this presentation is to introduce the concept of Clinical Utility (CU) as a means of identifying optimal doses relative to a reference treatment, given explicitly stated relative importance weightings of all the efficacy and safety endpoints under consideration. CU concepts are illustrated by two anonymized case studies.
Methods: Conceptually, clinical utility is a summary measure of the net benefit of a given treatment. Quantifying the benefit as probability of efficacious response, and the risk as the probability of adverse effect, transforms efficacy and safety data into measures that facilitate the assessment of net benefit. As the efficacy and safety endpoints are unlikely to be of equal clinical importance, the probability of efficacy/safety events are weighted by their relative importance to determine clinical utility. Alternative methods of constructing clinical utility functions are discussed.
Results: CU concepts were applied to facilitate dose selection for 2 compounds. In each case, the implication of alternative importance weightings on the selection of an optimal dose was evaluated. The CU for compound A is based on the dose-response for 2 efficacy and 3 safety endpoints; the CU for compound B is based on 1 efficacy and 1 safety exposure-response relationships.
Conclusion: CU represents a convenient metric for assessing the net benefit of a treatment relative to a reference treatment, by combining disparate safety and efficacy endpoint data. An advantage of the CU approach is that it requires explicit statement of clinical expert judgements on the relative importance of safety and efficacy endpoints.