My research focuses on evaluating interventions that aim to prevent infectious diseases using methods in causal inference.

My published and ongoing research includes work related to

  1. Child Growth Failure
  2. Influence of Weather on Enteric Pathogens and WASH Interventions
  3. School-Based Influenza Vaccinations
  4. Epidemiological Methods for Infectious Diseases
  5. Reproducibility

Child Growth Failure

Current estimates of global child growth faltering rely on cross-sectional data, but little is known about the timing and recurrence of stunting and wasting. I contributed to a large meta-analysis that pooled individual-level data from 35 longitudinal cohort studies to study child growth failure in low- and middle-income countries. In these studies, I produced manuscript figures, assisted with revisions, and refactored analysis code to improve study reproducibility. Our team found that the highest incidence of child wasting and stunting happened within the first three months of life and that early growth faltering was associated with persistent growth faltering. This work resulted in three papers that are currently under revision at Nature and will provide evidence to guide the development of targeted prenatal and early-postnatal interventions.

Influence of Weather on Enteric Pathogens and WASH Interventions

Our group previously found that low-cost, point-of care water, sanitation, and hygiene (WASH) interventions did not have substantial effects on enteric disease. We hypothesized that environmental conditions might influence the prevalence of enteropathogens and efficacy of WASH interventions. I led a large data curation process to pull high-resolution spatiotemporal data on these risk factor variables, allowing us to merge data from a randomized clinical trial with accurate estimates of environmental conditions at the time of infection. I supported one paper using this data to assess how climate and environment impacted enteropathogen carriage and am leading two first author papers to assess how these environmental factors impacted the effectiveness of WASH interventions on childhood diarrhea, STH infection, and giardia infection. We found that temperature and precipitation had heterogenous impacts on carriage by pathogen type and that WASH interventions were more effective at preventing diarrheal disease following heavy precipitation. We also found that intervention effectiveness is expected to increase in high-emission climate change scenarios that would lead to more frequent and extreme rainfall. We hope that this work will help target the implementation of future WASH interventions to improve population resilience again enteric diseases under climate change.

School-Based Influenza Vaccinations

Children are super-spreaders of influenza, largely due to high contact rates with their peers in school settings. Increasing vaccination rates among children might lower the burden of influenza by directly preventing infections in the age group and indirectly reducing transmission in the broader community. Our team evaluated the impact of a school-based influenza vaccination program using survey data, school absentee data, hospitalization surveillance data, and medical records from a large health system. Using difference-in-differences analyses, we found that the vaccination program was associated with higher school-aged vaccination, lower school absenteeism, and all-age hospitalizations in the community. In my master’s thesis, I led a first author paper to investigate differences in program impact by race/ethnicity. We found that the program was associated with the largest increases in vaccinations among White and Hispanic/Latino students and lower hospitalizations among Black older adults and White/Asian Pacific-Islanders of all ages. We saw evidence that differences in program effectiveness were linked to underlying reasons for vaccine non-receipt; school-based vaccinations helped overcome logistical barriers but did not sufficiently address hesitations about vaccine effectiveness or safety. Several co-authors for these papers are based in the California Department of Public Health, and our work has helped plan school-based vaccination campaigns for influenza and COVID-19 in the state.

Epidemiological Methods for Infectious Diseases

Traditional statistical methods often make parametric assumptions that cannot be assessed when transmission dynamics are not well defined, such as in emerging disease and disease elimination settings. To better understand the epidemiology of infectious diseases, we can apply novel techniques that leverage advances in statistics and computing. I’ve contributed to two papers that examined how different sources of bias impact statistical estimates in epidemiologic studies: (1) we estimated the prevalence of COVID-19 after correcting for imperfect testing using a semi-Bayesian probabilistic bias analysis and (2) comparing intervention effect estimates in an observational study using causal inference methods versus estimates from a mathematical transmission model. In these projects, I helped review statistical methods, replicate statistical analyses, create a web page to display interactive visualizations, and review manuscript drafts.

Reproducibility

In my research, I maintain reproducible computational workflows and regularly publish analysis code on my GitHub page. I have also contributed to the Benjamin-Chung Lab Manual which outlines best practices for transparency and reproducibility in computational epidemiology, and have written an invited commentary on rigor and reproducibility for Paediatric and Perinatal Epidemiology.