I use routine health data and causal inference to study the benefits and harms of medicines, cancer and surgical outcomes, and access to care.
Postdoctoral Research Fellow, Institute for Cancer Care, Mercy Medical Center, Baltimore
Each one started at the bedside, and each one is answered with the data health systems already collect.
How quickly regulators act on safety signals, and how coverage decisions change what patients receive.
Who benefits from major cancer surgery, and how cost and insurance design shape when people get operated on.
Whether the metrics used to compare intensive care units reward the right hospitals.
Analysis code is public wherever the data use agreement allows it.
A 12-year cohort of all 652 safety signals the FDA disclosed from its Adverse Event Reporting System, across 579 products, with G. Caleb Alexander. The median lag from first case report to FDA disclosure was 5.7 years, and 71% of resolved signals led to regulatory action, mostly labeling changes.
An active-comparator cohort of 8,240 ICU admissions. Allergy labels were associated with lower odds of β-lactam therapy (OR 0.26) and higher odds of carbapenem use (OR 1.74), with weighting balanced on 46 covariates and three null negative controls.
Five states that withdrew coverage compared with 12 that continued it, using difference-in-differences and event-study models on State Drug Utilization Data, 2021 to 2026. Withdrawal was associated with an 85% decline in obesity-indicated dispensing within one quarter.
Since 2021 I have mentored students through Open Dreams, a Cameroonian nonprofit that guides high-achieving students from low-income families to university admission and full scholarships. Each year I take on about ten students, mark every draft, and run mock interviews. Three of last year's cohort now hold full university scholarships, and they mentor the students behind them.
I am looking for research groups working on pharmacoepidemiology, cancer outcomes, and causal methods for routine health data.