Quantitative decision making in early phases of drug development with correlated endpoints
*Shyla Jagannatha, Janssen Pharmaceuticals, Inc.  Cristiana Mayer, Johnson & Johnson 

Keywords: No/No Go, Correlated Endpoints, Simulation

One of the primary objectives of a Phase 2 study is to establish proof of concept (PoC) of efficacy for the desired endpoint(s). Often times, the design involves testing a high dose versus placebo to establish PoC and then proceeding to a dose ranging study. An interim analysis in the PoC study can be used to accelerate the decision to start the dose ranging study and thus avoid a pause in enrollment between the 2 stages. Different decision criteria to accelerate development may depend on a set of design parameters and assumptions including the timing of the IA, the amount of information at interim, and the expected efficacy response at the end of the study. Evaluations of scenarios can become complicated when more than one endpoint (e.g., 2 endpoints, one continuous and one binary) are to be taken into account for decision purposes at interim. This case study with a 2-stage operationally seamless design will illustrate the probability of “Go/No Go” into the second stage of dose selection at the interim time point using two correlated endpoints and a confidence interval approach. Simulations are critical in supporting a quantitative decision making process to quantify the risk and the benefits of a 2-stage approach in view of different potential outcomes. A simulation study was used to determine the appropriate IA time point and the endpoint(s) decision rules that would trigger, or not, the acceleration into the dose-ranging study. The Go/No Go decision rules, operating characteristics, and risk quantification of the design features will be presented with details on the simulations.