MD, PhD
Behavioral Science · Data Science · Health Care Delivery
My work sits at the intersection of behavioral science and data science — designing, testing, and scaling interventions that improve how patients take their medications, how clinicians prescribe, and how technology can be harnessed to optimize care. The challenge is not just discovering what works in theory, but translating it into practice — understanding for whom, and how to deliver it at scale.

“The more we understand what drives behavior, the more effectively we can close the gap between what we know and what we do. And that’s the key to turning scientific discovery into real-world impact.”
Niteesh K. Choudhry, MD, PhD · Harvard Medical School Insights, 2026Designing and testing scalable strategies — financial incentives, behavioral nudges, data-driven personalized messaging — to help patients take medications as prescribed.
Applying behavioral science to improve what clinicians prescribe — and to safely stop what they shouldn't. Techniques include EHR-based decision support, pre-commitment tools, audit and feedback, and patient-directed outreach.
Applying reinforcement learning and machine learning to predict non-adherence before it happens and deliver personalized interventions — the right message, to the right person, at the right moment.
A national randomized trial with Aetna (n=5,855) showing that eliminating medication copayments after heart attack improved adherence, reduced major vascular events, and lowered patient spending — without increasing insurer costs. Aetna subsequently changed its national benefits policy.
A cluster-randomized trial at 14 primary care sites (n=4,078) testing a pharmacist-led, behaviorally tailored intervention combining telephone consultations, text messaging, and automated report cards for patients with poorly controlled chronic conditions.
A five-arm trial (n=228,000) conducted with the White House Social and Behavioral Science Team showing that a single behaviorally designed mailed letter increased influenza vaccination in Medicare beneficiaries — a meaningful effect at national scale.
A randomized trial testing a reinforcement learning algorithm that identified each patient's individual responsiveness to text message content and personalized messaging accordingly. Improved diabetes medication adherence by 13.6 percentage points versus control.
A randomized trial showing that EHR tools designed using behavioral science principles — including pre-commitment strategies and booster reminders — increased safe deprescribing of high-risk medications in adults aged 65 and older by 40% compared to usual care.