Designing a Bayesian adaptive clinical trial to evaluate novel mechanical ventilation strategies in acute respiratory failure using Integrated Nested Laplace Approximations
Authors:
Reyhaneh Hosseini,
Ziming Chen,
Ewan Goligher,
Eddy Fan,
Niall D. Ferguson,
Michael O. Harhay,
Sarina Sahetya,
Martin Urner,
Christopher J. Yarnell,
Anna Heath
Abstract:
Background: We aimed to design a Bayesian adaption trial through extensive simulations to determine values for key design parameters, demonstrate error rates, and establish the expected sample size. The complexity of the proposed outcome and analysis meant that Markov Chain Monte Carlo methods were required, resulting in an infeasible computational burden. Thus, we leveraged the Integrated Nested…
▽ More
Background: We aimed to design a Bayesian adaption trial through extensive simulations to determine values for key design parameters, demonstrate error rates, and establish the expected sample size. The complexity of the proposed outcome and analysis meant that Markov Chain Monte Carlo methods were required, resulting in an infeasible computational burden. Thus, we leveraged the Integrated Nested Laplace Approximations (INLA) algorithm, a fast approximation method, to ensure the feasibility of these simulations. Methods: We simulated Bayesian adaptive two-arm superiority trials that stratified participants into two disease severity states. The outcome was analyzed with proportional odds logistic regression. Trials were stopped for superiority or futility, separately for each state. We calculated the type I error and power across 64 scenarios that varied the stopping thresholds and the minimum sample size before commencing adaptive analyses. We incorporated dynamic borrowing and used INLA to compute the posterior distributions at each adaptive analysis. Designs that maintained a type I error below 5%, a power above 80%, and a feasible mean sample size were then evaluated across 22 scenarios that varied the odds ratios for the two severity states. Results: Power generally increased as the initial sample size and the threshold for declaring futility increased. Two designs were selected for further analysis. In the comprehensive simulations, the one design had a higher chance of reaching a trial conclusion before the maximum sample size and higher probability of declaring superiority when appropriate without a substantial increase in sample size for the more realistic scenarios and was selected as the trial design. Conclusions: We designed a Bayesian adaptive trial to evaluate novel strategies for ventilation using the INLA algorithm to and optimize the trial design through simulation.
△ Less
Submitted 31 March, 2023;
originally announced March 2023.
