research themes

My lab’s research program aims to explore and expand how advances in causal inference, statistical machine learning, and computational statistics catalyze discovery in the biomedical and public health sciences. Our methodological research emphasizes an assumption-lean, model-agnostic philosophy for statistical inference, applying a translational perspective that embraces the rich interplay between the applied sciences and the development of tailored statistical methods. This approach relies on causal inference principles to translate scientific questions into precise, interpretable statistical estimands, which can be learned from data generated by properly designed studies through the application of analytic methods that

1. avoid imposing restrictions not justified by available domain knowledge;
2. incorporate flexible, adaptive modeling strategies (e.g., machine learning); and
3. apply semi-parametric efficiency theory for best-in-class uncertainty quantification.

Thus, two key themes of our research program are the integration of causal inference ideas with tools and techniques from non-parametric estimation and statistical machine learning and the application of semi-parametric theory for asymptotically efficient estimation. These approaches have yielded novel insights applicable for de-biased or targeted causal machine learning (e.g., targeted minimum loss estimation, sieve estimation); non-parametric causal mediation analysis to study questions of mechanism; causal effect heterogeneity or effect modification analysis to support stratified medicine; large-scale, moderated semi-parametric inference for biomarker discovery; corrections needed to reliably draw accurate inferences when data are collected using two-phase, auxiliary- or outcome-dependent sampling designs; and quantifying the causal effects of flexible intervention regimes for continuous exposures.

A secondary theme of our research centers on the role of high-performance computing and the development of open-source software tools for statistical science. These distinct areas are unified by the overarching aims of pushing the boundaries of statistical methods development and promoting reproducibility and transparency in the practice of applied statistics. Consistent with our commitment to open science, new methods developed by members of the lab are accompanied by open-source software implementations both to ensure replicability of the reported work and to facilitate widespread use of the proposed techniques.

Here are a few highlights from research projects completed over the last few years:

Hejazi et al. (2023): SARS-CoV-2 pseudovirus neut. antibody as an immune correlate of protection via stochastic interventional vaccine efficacy.

Hejazi et al. (2022): Direct and indirect effects of hypothetically increasing odds of an early-increase dosing schedule on opioid use disorder relapse.

Hejazi et al. (2023): Stability to data erosion of top 30 differentially methylated CpGs identified using moderated semi-parametric estimation of the ATE.

Hejazi et al. (2021): Counterfactual HIV-1 infection risk in vaccinees under hypothetical modifications to CD8⁺ standardized polyfunctionality score.

Browse more about our work on statistical methods innovations, their use in applied sciences, and in developing open-source software for statistical science.