Applying a Targeted Machine Learning and Causal Inference Approach to Analyzing Long-Term Sequelae of COVID-19 Infection Through the National COVID Cohort Collaborative.

PROJECT SUMMARY / ABSTRACT Candidate: I am an epidemiologist in the Division of Biostatistics at the University of California, Berkeley School of Public Health, and I completed my Ph.D. in Epidemiology in August 2022 at UC Berkeley. Since my graduation, I have worked with the Center for Targeted Machine Learning and Causal Inference (CTML) to

apply cutting-edge biostatistical and causal inference methods to pressing COVID-19 research questions using data from the National COVID Cohort Collaborative (N3C). I led a group of CTML epidemiologists and biostatisticians in the NIH Long COVID Computational Challenge (L3C) competition, where we were honored

with third place for our ensemble machine learning model that accurately predicted the risk of Long COVID diagnosis based on individual electronic health record (EHR) data in N3C. I aim to become a leader in the application of innovative biostatistical, causal inference, and machine learning methods to impactful research

questions related to infectious disease epidemiology. Environment: In order to attain my career goals, my training and mentorship plan will focus on recent advances in biostatistics, causal inference, and data science methods (Targeted Machine Learning) as well as immunology and infectious disease epidemiology. I have assembled an interdisciplinary team of expert

biostatisticians, epidemiologists, and clinicians who will support my training. Alan Hubbard (primary mentor) and Mark van der Laan (co-mentor) will provide expert guidance and mentorship on biostatistics, data science, and causal inference. Rena Patel (co-mentor) and Jack Colford (scientific advisor) will provide mentorship and

guidance in infectious disease epidemiology and immunology. Research: Researchers and clinicians have made enormous progress in understanding, preventing, and treating acute COVID-19 infection, but there is considerable uncertainty regarding the factors associated with long-term sequelae of COVID-19 infection. Although vaccination is a key strategy for COVID-19 epidemic

control, little is known regarding the role of COVID-19 vaccination timing relative to COVID-19 infection (i.e., up-to-date vaccinations and boosters) in preventing long-term sequelae of infection, and the lack of objective Long COVID biomarkers hampers our ability to evaluate, prevent, and treat Long COVID. In Aim 1, I will

evaluate the relationship between vaccination timing and Long COVID diagnosis in order to determine an optimized vaccination schedule to minimize Long COVID. In Aim 2, I will assess the relationship between COVID-19 vaccination timing and individual long-term sequelae of COVID-19 infection. In Aim 3, I will assess

mediation of the relationship between acute COVID-19 infection and Long COVID via interleukin 6 (IL-6) to evaluate a biological mechanism of interest. I will apply Targeted Machine Learning methods to achieve these aims, which will prepare me for an R01-level application to apply these methods to research questions in

infectious disease epidemiology.