CARE Lab

Research → Innovative study design methodology

Innovative study design methodology

Supporting regulatory decision-making

Real-world evidence is only as trustworthy as the design that produced it. Regulators and payers act on these studies, so the standard is not whether an analysis can be run but whether its result can be defended. We develop and evaluate designs for the situations where conventional approaches do not reach — and we work on them because specific decisions depend on them, not as methodology for its own sake.

External control arms

When a therapy is approved on the basis of a single-arm trial, there is no comparator, and the size of the benefit is difficult to judge. We build external control arms from registry and claims data to supply that missing comparison, work that began with CAR-T therapies in lymphoma and now spans several hematologic indications. Most of the difficulty lies in questions that sound simple: when follow-up should start, how to handle patients with more than one plausible eligibility date, and which comparator population makes a comparison fair rather than merely available. These choices decide whether a result holds up under regulatory review, and they are where our methodological work concentrates.

RCT generalization and transportability

A trial answers its question for the patients it enrolled. Clinicians and regulators need to know what happens to everyone else. We are developing methods that use trial data to predict treatment effects in two groups the trial never observed: patients who met the eligibility criteria but were not enrolled (generalization), and patients who would not have been eligible at all — typically older, sicker, or taking more medications (transportability). The aim is to anticipate what a drug will do once it leaves the trial and enters practice, and to make that prediction checkable against what actually happens.

Selected work

See all publications →