Skip to main content
Fig. 2 | BMC Infectious Diseases

Fig. 2

From: A comparison of clinical development pathways to advance tuberculosis regimen development

Fig. 2

Phase IIc MAMS and BAR optimization and comparison. A Fraction of simulations graduated for Phase IIC MAMS trials with interim criteria fixed at TCC HR > 1.7 while changing interim timing from 10 to 100 patients recruited into the control arm, at 10 patients/week interim timing is approximately study week 10–100. B Fraction of simulations graduated for Phase IIC MAMS trials with interim timing fixed to 50 patients and changes TCC HR criteria from 1.1 to 2.3. Dotted lines show the chosen optimized conditions, where an interim timing of 50 patients per arm is the earliest timing in which the risk of stopping the desirable regimen is negligible and an interim criteria of TCC HR > 1.7 is the strictest criteria in which the risk of stopping the desirable regimen is negligible. The control in grey, represents the proportion of simulations in which the trial was not stopped prematurely due to all investigational arms being stopped. C Comparison of graduation and stopping rates of optimized Phase IIc MAMS and BAR designs. The continuous nature of the BAR recruitment was translated into a semi-discrete outcome for comparison to the MAMS design, graduation of an arm was defined as greater than 80 patients recruited into the arm, stopping of an arm defined as less than 50 patients recruited (analogous to stopping at MAMS interim) and in-between defined as 50–80 patients recruited into the arm. Both designs meet our target criteria graduating > 95% of desirable regimens and < 10% of suboptimal regimens. D Heatmap quantifying the aggressiveness of Bayesian adaptive randomization as the ratio of patients allocated to the desirable regimen/suboptimal regimen across a range of reasonable ɣ and η values. E Heatmap of the variability expressed as %CV in the ratio shown in D across 1000 simulations. The optimal condition outlined in black was chosen for its aggressiveness and limited variability while meeting our target critiera

Back to article page