arXiv subjects

Zhangyi He

Publications and source records attributed to Zhangyi He.

5 recordsLinked to original sources

arXiv2026

adabay: an R package for rapid evaluation and calibration of Bayesian group sequential designs across common endpoint types

Bayesian group sequential designs (GSDs) combine the efficiency of frequentist GSDs with clinically interpretable probability statements and principled external evidence incorporation. However, their uptake in confirmatory trials has been held back by the cost of evaluating frequentist operating characteristics at the design stage, which often nests Markov chain Monte Carlo or another approximate posterior inference within a Monte Carlo trial-simulation loop. We introduce adabay, an open-source R package implementing a semi-simulation framework for the rapid evaluation and calibration of Bayesian GSDs. Trial data paths are simulated by Monte Carlo, while per-look posteriors and posterior tail probabilities are computed analytically or by low-dimensional deterministic quadrature. Flexible prior specification is achieved by approximating any user-specified prior with a finite mixture of conjugate components, with tail-probability diagnostics that flag inadequate approximations at the decision thresholds. The package offers a unified application programming interface for continuous, binary, count and time-to-event endpoints, supports posterior-probability decision rules with one or more efficacy and futility criteria under binding or non-binding regimes, and includes a precomputation strategy that decouples threshold calibration and look-time selection from the simulation pass. adabay reproduces the operating characteristics of BATSS and adaptr in the continuous and binary case studies, the analytic gsbDesign values in the continuous case, and the BATSS results in the count case, all within Monte Carlo error, while running approximately four to five orders of magnitude faster than BATSS and one to over two orders of magnitude faster than adaptr per virtual trial on eight cores. The package is distributed under an MIT licence.

stat.ME↗

arXiv2026

Rapid evaluation and calibration of Bayesian group sequential designs via conjugate-mixture semi-simulation

Bayesian group sequential designs (GSDs) extend frequentist GSDs with interpretable decision-making and external evidence borrowing, but their use is limited by the computational burden of design-stage operating-characteristic evaluation. Conventional methods simulate virtual trials with Markov chain Monte Carlo or approximate analytical posterior updates at each interim look, making joint calibration of decision thresholds and design skeletons impractical on commodity hardware. Here we introduce a semi-simulation framework with two innovations. First, finite conjugate-mixture priors replace the posterior computation for each look (``per-look'') with closed-form conjugate updates and low-dimensional numerical integration for decision-rule tail probabilities. Second, a precomputation strategy caches per-look posterior tail probabilities from a single Monte Carlo pass at the union of all candidate analysis times, and each design in the calibration grid is evaluated against the same cache by a sub-second sweep, with no further simulation cost. The framework supports posterior-probability decision rules with multiple efficacy and futility criteria under either binding or non-binding futility, and derives closed-form per-look updates for binary, continuous, count and time-to-event endpoints, with benchmarking here focused on the binary endpoint. When applied to re-design the ADRENAL trial, with up to nine analyses, the framework reproduces the operating characteristics of BATSS and adaptr within Monte Carlo error while running, per GSD, approximately $7\times$ to $16\times$ faster than adaptr at a matched budget (several hundredfold at the million-trial calibration budget) and $3{,}700\times$ to $6{,}600\times$ faster than BATSS. This brings routine Bayesian GSD calibration within computational reach for confirmatory trials.

stat.ME↗

arXiv2026

From aggressive to conservative early stopping in Bayesian group sequential designs

Group sequential designs (GSDs) are widely used in confirmatory trials to allow interim monitoring while preserving control of the type I error rate. In the frequentist framework, O'Brien-Fleming-type stopping boundaries dominate practice because they impose highly conservative early stopping while allowing more liberal decisions as information accumulates. Bayesian GSDs, in contrast, are most often implemented using fixed posterior probability thresholds applied uniformly at all analyses. While such designs can be calibrated to control the overall type I error rate, they do not penalise early analyses and can therefore lead to substantially more aggressive early stopping. Such behaviour can risk premature conclusions and inflation of treatment effect estimates, raising concerns for confirmatory trials. We introduce two practically implementable refinements that restore conservative early stopping in Bayesian GSDs. The first introduces a two-phase structure for posterior probability thresholds, applying more stringent criteria in the early phase of the trial and relaxing them later to preserve power. The second replaces posterior probability monitoring at interim looks with predictive probability criteria, which naturally account for uncertainty in future data and therefore suppress premature stopping. Both strategies require only one additional tuning parameter and can be efficiently calibrated. In the HYPRESS setting, both approaches achieve higher power than the conventional Bayesian design while producing alpha-spending profiles closely aligned with O'Brien-Fleming-type behaviour at early looks. These refinements provide a principled and tractable way to align Bayesian GSDs with accepted frequentist practice and regulatory expectations, supporting their robust application in confirmatory trials.

stat.ME↗

arXiv2025

A Bayesian model-free dose-finding approach in Phase II clinical trials with the flexibility of historical borrowing

Accurate dose selection in Phase II trials is critical to the success of subsequent Phase III trials, but suboptimal choices remain a leading cause of trial failure and regulatory rejection. Although MCP-Mod is widely adopted and endorsed by regulatory agencies, it requires prespecification of candidate models and is highly sensitive to model misspecification. To address these challenges, we introduce MAP-curvature, a general model-free framework for dose-response modelling that penalises the total curvature of the dose-response curve through a prior. Within this framework, LiMAP-curvature arises as the linear special case, whereas SEMAP-curvature, the focus of this work, employs the sigmoid Emax model, providing greater flexibility to capture nonlinear pharmacological patterns. Through extensive simulations, we show that SEMAP-curvature generally outperforms LiMAP-curvature and MCP-Mod in detecting the dose-response signal, estimating the dose-response curve and identifying the minimum effective dose, with particularly significant improvements under concave downward shapes resembling the sigmoid Emax model. Although SEMAP-curvature exhibits slightly greater variability, it remains robust in accuracy and reliability. We further extend MAP-curvature by integrating it with the Bayesian hierarchical model to enable flexible borrowing of historical data, which improves power and precision, particularly when dose levels overlap across studies. These results highlight MAP-curvature, and in particular SEMAP-curvature with historical borrowing, as a robust and efficient framework for dose selection in early-phase clinical trials.

stat.ME↗

arXiv2025

Optimal scheduling of interim analyses in group sequential trials

Group sequential designs (GSDs) are well established and the most commonly used adaptive design in confirmatory clinical trials with interim analyses. However, they remain underutilised, and their implementation involves unique theoretical and practical decisions that demand careful consideration to optimise efficiency. A common practice is to schedule interim analyses at equal intervals based on calendar time or accumulated data. While straightforward, this approach does not completely exploit the potential sample size savings achievable with GSDs. To address this challenge, we develop OptimInterim, an R-based tool that can determine the optimal scheduling of interim analyses to minimise the expected sample size under the alternative hypothesis while controlling overall type I and type II errors. Our method accommodates trials with continuous or binary endpoints, allows multiple interim analyses and supports a range of stopping boundaries. Through extensive simulations, we demonstrate that optimally spaced interim analyses can yield substantial savings in expected sample size compared to equally spaced interim analyses, without compromising the maximum sample size, across various endpoint types, effect sizes, error rates and stopping rules. We illustrate its practical utility with two landmark trials evaluating steroid use in septic shock. Notably, for given type I and type II error rates, the optimal scheduling is independent of endpoint types and effect sizes, ensuring broad applicability across a wide range of trial contexts. To facilitate implementation, we offer a ready-to-use reference table of optimal schedules for up to eight interim analyses under commonly used error rates and stopping rules. Access OptimInterim at https://github.com/zhangyi-he/GSD_OptimInterim.

stat.ME↗