Visit my personal 
page here and group IntegrativeStats.
R packages for matching:
- multilevelMatching for implementing a novel matching procedure to compare multiple treatments simultaneously from the observational data. (CRAN)
- dsmatch implements double score matching for the average treatment effect and quantile treatment effect estimation.
- AMW implements augmented matching weighted estimator for the average treatment effects, which selects the number for matches adaptively and achieves double robustness. The double penalties are used to select important covariates for adjustment and encourage the sparsity of the bias function for efficiency gain.
R packags for continuous-time causal inference
- contTimeCausal provides estimation methods for continuous-time structural failure time models (ctSFTM) and continuous-time Cox marginal structural models (ctCoxMSM). (CRAN)
- FPCA implements functional principal component analysis with informative irregularly-spaced observation times.
R packages for generalizing randomized trial findings
- genRCT generalizes the average treatment effect (ATE) from the trial using observational studies (OS) for binary/continuous/survival outcomes.
- genITR conducts transfer learning of the individual treatment regime (ITR) from a source to a target population using summary statistics.
- TansferLearningSurvITR conducts transfer learning of the ITR from a source to a target population for survival outcomes.
R packages for integrative analysis
- IntegrativeCI implements integrative analyses for the average treatment effect combining big main data and smaller validation data.
- IntegrativeHTE implements elastic analyses for the heterogeneous treatment effects combining trials and real-world data.
- IntegrativeHTEcf implements integrative analyses for the heterogeneous treatment effects combining a randomized trial and confounded real-world data using the notion of confounding function.
- intRlearner implements an integrative R-learning analysis of the heterogenous treatment effects by combing a randomized trial and confounded real-world data using the notation of confounding function.
- IntegrativeFPM implements integrative analyses for the finite population mean combining probability and non-probability samples with high-dimensional data.
- DPIE (double penalty integrative estimation) implements an integrative estimator combining randomized clinical trial data and external controls for an augmented ANCOVA + bias function. The double penalty is used to select important covariates for adjustment and encourage sparsity of the bias function for efficiency gain.
- intFRT implements Fisher randomization test of randomized trial data with conformal selection borrowing of external controls.
R packages or functions for missing data, causal inference, and individualized treatment regime learning
- IITR implements a doubly robust estimator of the risk function and Adpative LASSO based algorithm to find interpretable individualized treatment regime.
- miATE implements a unified bootstrap inference of the average treatment effect after multiple imputation based on martingales.
- pace implements various estimators of principal strata average causal effects from observational studies.
- smim implements Survival sensitivity analysis using Multiple Imputation and Martingale.
- spatial_confounding contains R-scripts and other files that were employed to carry out simulation studies and data analysis described in the paper “A Spectral Adjustment for Spatial Confounding.”
- MSNMM solves the doubly robust estimating equation proposed in the paper “Multiplicative structural nested mean model for zero-inflated outcomes.”
- RealWorld-DoublyRobustML implements various doubly robust estimators using machine learned nuisance functions, including DSM, AIPW, TMLE, and PENCOMP