LP LAB

Research Team

Rebecca's research focuses on developing methods to examine and test for complex heterogeneity in the utility of surrogate markers in both randomized and real data settings via meta-learners. She has proposed a nonparametric method for efficient testing using surrogate information (ETSI), which enables treatment effect estimation and hypothesis testing in a setting where the surrogate is valid to substitute for the primary outcome for certain patient subgroups, and not for others.

Emily Hsiao, PhD Student

Emily's research focuses on developing an empirical framework to test assumptions that ensure protection from the surrogate paradox via nonparametric testing methods, and a simulation-based approach to assess resilience to the surrogate paradox in a future study.

Ahantya Sharma, Undergraduate Researcher

Ahantya is an SDS major working with me to develop a website that showcases surrogate marker methods, related software, and educational content. The site will serve as both an informational resource and a platform for implementing surrogate methods through interactive Shiny apps.

Dylan Huynh, Undergraduate Researcher

Dylan's is an SDS major working on improving the computational efficiency of nonparametric testing for monotonicity, significantly optimizing performance of an existing test and making it much faster and more scalable. To enhance accessibility and usability, he developed an interactive Shiny app that allows users to apply the test easily to their own data.

Alum

Yunshan Duan, Graduate Research Assistant

Yunshan's research focused on proposing a Bayesian model averaging approach to estimate the proportion of treatment effect explained by a surrogate marker. This procedure offers a compromise between the model-based approach and the nonparametric approach by introducing model flexibility via averaging over several candidate models and maintains the strength of parametric models with respect to inference.

Thao Nguyen, Undergraduate Researcher

Thao examined risk factors that contribute to the development of mild cognitive impairment and its conversion to dementia and studied the importance of these features in predicting cognitive function among adults at high-risk for diabetes. She implemented and compared various supervised machine learning models to determine which algorithm, as well as which features, were most predictive of cognitive function.

Why not Parast Lab? Because my amazing cousin, Mana Parast, a physician-scientist doing research on placental development and pathology, already has Parast Lab. Learn more about her here.