Figure 1From: Improving clinical trial efficiency by biomarker-guided patient selection Improvement rates in the number of patients to recruit for ARD and BSD as compared to PGD under scenarios IA-C. Positive improvement indicates smaller sample sizes and thus higher efficiency. The sample size requirements for the parallel group design were 1386, 800 and 318 for scenario IA-C, respectively. ARD, active run-in design; BSD, baseline selection design; PGD, parallel group design.Back to article page