Research

Research Areas

Commercial Determinants of Health

I study how corporations market tobacco, alcohol, sugary beverages, and ultra-processed foods through digital channels, and how this marketing undermines public health. My most recent work in this area included leading the CANARY program at Vital Strategies — large-scale digital marketing surveillance using AI-assisted content classification, social media analysis, and real-time monitoring systems. I continue to develop independent tools in this space through projects such as Digital Sentinel AI.

Public Health Campaign Evaluation

I design and execute evaluations of large-scale public health media campaigns, using multi-stream data integration to measure reach, awareness, behavioral change, and policy impact. My most recent work focused on overdose prevention campaigns targeting Black communities across U.S. cities.

AI and Machine Learning in Public Health Research

I develop and validate AI-assisted tools for public health research, including LLM-based content classifiers, automated QA systems with explainable outputs, and Python-based data pipelines. I am particularly focused on making AI tools reproducible, auditable, and appropriate for use in policy-relevant research contexts.

Behavioral and Policy Research

I apply advanced statistical methods — including survey-weighted logistic regression, factor analysis, structural equation modeling, and behavioral prediction modeling — to understand determinants of health behavior and policy support across diverse populations.

Methodological Expertise

Quantitative methods: Logistic regression, survey-weighted analysis, multivariate regression, factor analysis, structural equation modeling, principal component analysis, behavioral prediction modeling

Qualitative methods: Message testing, focus groups, content analysis, systematic literature review

Evaluation frameworks: Monitoring, Evaluation, and Learning (MEL), process and outcome evaluation, multi-stream data integration

AI/computational methods: Large language model (LLM) classification, prompt engineering, automated validation, inter-rater reliability (Cohen’s kappa), explainable AI

Geographic Scope

My research has been conducted across or in collaboration with local partners and institutions in:

  • United States - overdose prevention campaign evaluation across seven cities; digital food marketing surveillance
  • India - tobacco-free film and television policy evaluation; tuberculosis mass media message testing
  • Brazil - sugary drinks taxation campaign evaluation; ultra-processed food perception research
  • Colombia - front-of-package labeling campaign evaluation; school nutrition message testing
  • Jamaica - sugary drinks communication campaign evaluation; ultra-processed food perception research
  • Barbados - school nutrition message testing
  • South Africa - front-of-package labeling evaluation; sugary drinks tax campaign research