Selected work

Representative Papers

Lazy Evaluation: A Comparative Analysis of SAS MACROs and R Functions

Jan 14, 2026

This study addresses the lack of systematic comparison between SAS and R regarding their lazy evaluation mechanisms, a gap that hinders cross-language code optimization. It presents the first comprehensive analysis contrasting SAS macros—implemented via symbol-table-based call-by-name—with R functions, which rely on promise-based call-by-need semantics. The work elucidates fundamental differences in evaluation strategies, memory consumption, and execution efficiency, demonstrating their tangible impacts on programming productivity through representative case studies. By clarifying these mechanistic distinctions, the paper provides both theoretical grounding and practical guidance for optimizing code during the transition from SAS to R, particularly in the pharmaceutical industry where such migration is increasingly prevalent.

1 citationsRead paper

A Unified Causal Inference Framework for the Desirability of Outcome Ranking Paradigm in Benefit-Risk Evaluation

Aug 05, 2026

This study addresses the absence of a unified causal inference framework for estimating the probability of Desirability of Ordered Random Outcomes (DOOR) in benefit–risk assessment. The authors propose the first covariate-adjusted, unified causal inference framework that expresses DOOR probabilities as bilinear functionals of marginal ordered outcome distributions under two treatment strategies. Conditional outcome distributions are estimated via sequential risk-set hazard models, and the corresponding efficient influence function is derived. The framework integrates G-computation, normalized inverse probability weighting (IPW), augmented IPW (AIPW), and targeted maximum likelihood estimation (TMLE), employing generalized linear models or Super Learner to estimate nuisance functions and incorporating cross-fitting to enhance robustness. The proposed CVTMLE-SL estimator demonstrates superior performance in terms of DOOR estimation bias, recovery of ordered outcome distributions, standard error accuracy, and confidence interval coverage, with empirical validation on real-world antimicrobial resistance data.

0 citationsRead paper

A Practical Framework for Sensitivity Analysis in Externally Controlled Trials: An Illustration with a Bayesian Hybrid Evidence Synthesis Case Study

Jun 07, 2026

This study addresses the lack of a systematic framework for evaluating the validity, benefit, and robustness of borrowing historical data in external control arm trials. The authors propose a “three-pillar, eight-module” sensitivity analysis framework that integrates eight analytical techniques—heterogeneity diagnostics, source influence analysis, no-borrowing benchmarks, effective sample size quantification, prior sensitivity assessment, tipping point analysis, alternative borrowing methods, and model structure sensitivity—to comprehensively address these core concerns. This work presents the first structured, methodology-agnostic template compatible with both Bayesian and frequentist paradigms and adaptable to diverse data types. Through simulated evidence synthesis mimicking real-world regulatory review scenarios, the framework demonstrates practical feasibility and reproducibility, offering a standardized paradigm for sensitivity analysis in external control trials and helping to fill a critical gap in regulatory guidance.

0 citationsRead paper

Synergy Area with FDR-controlled Evaluation (SAFE) to robustly assess safety profile in clinical trials

May 04, 2026

This study addresses the lack of robust quantitative methods in clinical trial safety assessment that integrate clinical knowledge, control error rates, and rely on sufficient evidence. To this end, the authors propose SAFE, a two-tier framework that first leverages clinical prior knowledge to identify clinically meaningful signal areas (SAs), and then applies cross-area false discovery rate (FDR) control for multiplicity adjustment. SAFE is the first approach to jointly model signal areas and enforce global FDR control, thereby enhancing clinical interpretability without compromising statistical rigor. Simulation studies and real-data analyses demonstrate that SAFE effectively controls both within-area and cross-area error rates while filtering out extreme values, leading to more reliable safety conclusions.

0 citationsRead paper

On the Calibration of Bayesian Success Criteria and Operating Characteristics for Clinical Trials

Mar 20, 2026

This study addresses the need for more reliable regulatory decision-making in Bayesian clinical trials by systematically calibrating Bayesian success criteria to control decision errors. It establishes the first theoretical correspondence between Bayesian decision error metrics and frequentist operating characteristics—specifically Type I and Type II error rates—and proposes a practical calibration strategy grounded in this relationship. The approach is illustrated through a case study on a revascularization trial in cardiogenic shock. To facilitate adoption under the FDA’s emerging Bayesian framework, the authors also developed an interactive Shiny web application that enables sponsors and regulators to efficiently and reliably formulate decisions while maintaining rigorous error control.

0 citationsRead paper

Recent publications

Latest Papers

A Unified Causal Inference Framework for the Desirability of Outcome Ranking Paradigm in Benefit-Risk Evaluation

Aug 05, 2026

This study addresses the absence of a unified causal inference framework for estimating the probability of Desirability of Ordered Random Outcomes (DOOR) in benefit–risk assessment. The authors propose the first covariate-adjusted, unified causal inference framework that expresses DOOR probabilities as bilinear functionals of marginal ordered outcome distributions under two treatment strategies. Conditional outcome distributions are estimated via sequential risk-set hazard models, and the corresponding efficient influence function is derived. The framework integrates G-computation, normalized inverse probability weighting (IPW), augmented IPW (AIPW), and targeted maximum likelihood estimation (TMLE), employing generalized linear models or Super Learner to estimate nuisance functions and incorporating cross-fitting to enhance robustness. The proposed CVTMLE-SL estimator demonstrates superior performance in terms of DOOR estimation bias, recovery of ordered outcome distributions, standard error accuracy, and confidence interval coverage, with empirical validation on real-world antimicrobial resistance data.

0 citationsRead paper

A Practical Framework for Sensitivity Analysis in Externally Controlled Trials: An Illustration with a Bayesian Hybrid Evidence Synthesis Case Study

Jun 07, 2026

This study addresses the lack of a systematic framework for evaluating the validity, benefit, and robustness of borrowing historical data in external control arm trials. The authors propose a “three-pillar, eight-module” sensitivity analysis framework that integrates eight analytical techniques—heterogeneity diagnostics, source influence analysis, no-borrowing benchmarks, effective sample size quantification, prior sensitivity assessment, tipping point analysis, alternative borrowing methods, and model structure sensitivity—to comprehensively address these core concerns. This work presents the first structured, methodology-agnostic template compatible with both Bayesian and frequentist paradigms and adaptable to diverse data types. Through simulated evidence synthesis mimicking real-world regulatory review scenarios, the framework demonstrates practical feasibility and reproducibility, offering a standardized paradigm for sensitivity analysis in external control trials and helping to fill a critical gap in regulatory guidance.

0 citationsRead paper

Synergy Area with FDR-controlled Evaluation (SAFE) to robustly assess safety profile in clinical trials

May 04, 2026

This study addresses the lack of robust quantitative methods in clinical trial safety assessment that integrate clinical knowledge, control error rates, and rely on sufficient evidence. To this end, the authors propose SAFE, a two-tier framework that first leverages clinical prior knowledge to identify clinically meaningful signal areas (SAs), and then applies cross-area false discovery rate (FDR) control for multiplicity adjustment. SAFE is the first approach to jointly model signal areas and enforce global FDR control, thereby enhancing clinical interpretability without compromising statistical rigor. Simulation studies and real-data analyses demonstrate that SAFE effectively controls both within-area and cross-area error rates while filtering out extreme values, leading to more reliable safety conclusions.

0 citationsRead paper

On the Calibration of Bayesian Success Criteria and Operating Characteristics for Clinical Trials

Mar 20, 2026

This study addresses the need for more reliable regulatory decision-making in Bayesian clinical trials by systematically calibrating Bayesian success criteria to control decision errors. It establishes the first theoretical correspondence between Bayesian decision error metrics and frequentist operating characteristics—specifically Type I and Type II error rates—and proposes a practical calibration strategy grounded in this relationship. The approach is illustrated through a case study on a revascularization trial in cardiogenic shock. To facilitate adoption under the FDA’s emerging Bayesian framework, the authors also developed an interactive Shiny web application that enables sponsors and regulators to efficiently and reliably formulate decisions while maintaining rigorous error control.

0 citationsRead paper

Lazy Evaluation: A Comparative Analysis of SAS MACROs and R Functions

Jan 14, 2026

This study addresses the lack of systematic comparison between SAS and R regarding their lazy evaluation mechanisms, a gap that hinders cross-language code optimization. It presents the first comprehensive analysis contrasting SAS macros—implemented via symbol-table-based call-by-name—with R functions, which rely on promise-based call-by-need semantics. The work elucidates fundamental differences in evaluation strategies, memory consumption, and execution efficiency, demonstrating their tangible impacts on programming productivity through representative case studies. By clarifying these mechanistic distinctions, the paper provides both theoretical grounding and practical guidance for optimizing code during the transition from SAS to R, particularly in the pharmaceutical industry where such migration is increasingly prevalent.

1 citationsRead paper