INTRODUCTION

Prognostic modeling utilizing real-world data

Predicting health outcomes in high-risk individuals is essential. While machine learning (ML) offers powerful risk stratification capabilities, robust prognostic modeling depends fundamentally on rigorously planned study design and statistical analysis. To move beyond adopting ML as merely a "fabulous" tool, it is critical to revisit epidemiological and biostatistical principles and ensure that the model outputs are clinically valid. By integrating advanced statistical methods grounded in epidemiological theory, I aim to develop prognostic models that are both clinically plausible and practically useful.

My research interests include:

• Explainable AI for prognostic modeling:
  to develop locally transparent model for clinical interpretations

• Simulation-based population insights:
  to inform population-level decision making from individual-level prognostic models

• Mixed-effects machine learning:
  to handle longitudinal data and capture both inter- and intra-individual variability

Causal inference in observational settings

Evaluating the effectiveness and safety of pharmacological intervention is critical for both clinical and regulatory decision-making. Robust causal estimates derived from real-world data can yield high-quality evidence and clinically valid insights. However, observational studies are susceptible to bias from non-experimental settings and unmeasured latent variables, which can undermine causal inference. To overcome these challenges, I aim to implement rigorously designed studies with state-of-the-art methodologies to generate valid clinical evidence for pharmacological interventions.

My research interests include:

• Drug-target Mendelian randomization:
  to infer the causal effect of drug-target perturbation on clinical outcomes by leveraging genetic instruments

• Biobank-scale epidemiological studies:
  to evaluate pharmacological effects with high-dimensional covariate adjustment, accounting for both genetic and environmental factors

Outcome-based pharmacy education

The College of Pharmacy oversees the training of pharmacists through its undergraduate and graduate program. Pharmacy practice primarily revolves around "actions" such as dispensing medications, providing medication counseling, and collaborating with other professionals. Therefore, it is crucial for students to internalize what instructors and preceptors teach and translate it into practice. In particular, education in clinical pharmacy must serve as a catalyst at each transitional stage, which includes: (1) building foundational pharmaceutical knowledge to enable pharmacists to involve in decision-making; (2) applying knowledge to practice during pharmacy rotations; and (3) accumulating advanced knowledge post-graduation to perform sophisticated pharmacist roles and apply it in real-world settings. By focusing on achieving competencies, I aim to design and evaluate the curriculum and successfully cultivate pharmacists.