About My Work
I am a biostatistician and infectious disease researcher focused on developing statistical methods and modeling approaches motivated by real-world data with the goal of producing tools and insights that inform public health and clinical decision-making. I work closely with clinicians, epidemiologists, and other scientists to ensure my methods are clinically relevant, while also developing open-source software to make these approaches broadly accessible to the research community.
Training & Background
I earned my PhD in Biostatistics from Boston University, where I worked with Dr. Laura White to develop statistical methods for questions in tuberculosis (TB) epidemiology. My doctoral research focused on integrating complex data sources and addressing methodological challenges in observational data to better understand TB dynamics in high-risk populations. I also collaborated with Dr. Karen Jacobson at Boston Medical Center on applied TB studies in South Africa.
Current Work
As a postdoctoral research fellow, I study infectious disease transmission using phylodynamics, mathematical modeling, and machine learning. These approaches help investigate how pathogens spread and how interventions influence epidemic outcomes. I continue TB transmission collaborations with Dr. White and Dr. Jacobson. I have also expanded to work with Dr. Ashlee Earl at the Broad Institute on a project studying methicillin-resistant Staphylococcus aureus (MRSA) transmission in nursing homes in which we are leveraging phylodynamic models that integrate genomic and epidemiological data.
Funding
My postdoctoral training is supported by the National Institute of Allergy and Infectious Diseases (T32AI052074).
Statistical Methods
Data-Driven TB Disease Phenotypes
Developed TB-STATIS, a model that integrates clinical information (symptoms, smear, chest X-ray) to generate TB disease phenotypes and stratify patients.
Software: tbSTATIS
Data Application: TRUST study (a TB clinical cohort) and the REMox trial
Inferring Associations with Respondent-Driven Sampling Data
Developed a semi-parametric randomization test that accounts for the correlated nature of respondent-driven sampling data to accurately infer associations.
Software: RDSAssociation
Data Application: TOTAL study
Multiple Imputation for Time-to-Event Data
Explored methods for handling missing data in repeated measures for time-to-event outcomes, motivated by the need to assess time to culture conversion in TB clinical trials.
Software: imputeTBculture
Data Application: TRUST study (a TB clinical cohort)
Modeling Infectious Disease Transmission
Role of Reinfection in Sustaining TB Transmission
Developing a population-level mathematical model to quantify the role of reinfection in sustaining TB epidemics. Results inform best intervention strategies such as active case finding, vaccination, and preventive therapy in high- and low-burden settings.
TB Transmission and Substance Use
Applying mlTransEpi, a machine learning method that integrates whole genome sequencing and metadata, to explore TB transmission dynamics among people who smoke drugs in a rural community in South Africa.
MRSA Transmission in Nursing Homes
Studying the transmission of multidrug-resistant organisms (primarily MRSA) in nursing homes in Orange County, California. Applying TransPhylo2 to identify transmission risk factors, evaluate decolonization strategies, and explore the impact of sampling in endemic settings.
Other Projects
TB Prevalence Among People Who Smoke Drugs
Estimated TB and HIV prevalence among people who smoke drugs in a rural community in South Africa using a respondent-driven sample of 750 participants. TB prevalence was ~10% which was three times higher than community rates.