iptw

Nonparametric estimators of the average treatment effect with doubly-robust confidence intervals and hypothesis tests

Horvitz-Thompson estimator for RCTs, with Joel Middleton

Comparison of vedolizumab and tofacitinib via inverse propensity of treatment weighting

Analysis of Parkinson’s patients to explore how dopaminergic medication dose (LEDD) affects gait and fall risk. Combining causal inference (IPTW) with machine learning (XGBoost, logistic models), we reveal that high LEDD improves some gait metrics but increases fall risk, highlighting a clinical trade-off.

Code for Harnois-Leblanc, Rifas-Shiman, Switkowski et al. (2025) Estimating sex-specific population-level effects AJE

Reproducible clinical/RWE causal-inference workflow estimating the effect of early right heart catheterization on 30-day mortality using IPTW, doubly robust AIPW, overlap diagnostics, sensitivity analysis, and Python.

Config-driven Quarto/Rmd workspace for survey-weighted observational papers: survey design, IPTW, LCA, network analysis and reporting skeleton. Clone once per paper.

An end-to-end causal-inference workflow using synthetic observational longitudinal data: propensity-score weighting, balance diagnostics, GEE, and weighted survival analysis.

Official R implementation of the preoperative dual-score framework (O-score and S-score) developed to evaluate oncological and surgical risk and estimate treatment benefit from curative-intent resection in hepatocellular carcinoma.