Founder • 2025 to present
Matchway - AI Patient Identification
I dove deep into this problem space after learning a few shocking statistics about patient identification in clinical trials. Research coordinators spend 30 hours a week manually scanning records to find eligible candidates, yet 80% of trials still miss their enrollment targets. Even more striking: only 5% of eligible patients are ever offered a trial, despite 75% saying they would join one if recommended. This gap feels solvable with software.
I built a working prototype, simulation model, website, brought on physician advisors from Jefferson, Penn Medicine, and Mount Sinai, and designed two structured pilots. Got it to a real place, excited to take matchway.org to the next step.
I built a working application that connects to EHR data, parses inclusion and exclusion criteria from active trials, and surfaces a ranked shortlist of matched patients with clear eligibility evidence. Built a financial simulation model that lets a research sites input their real numbers and see projected time savings, enrollment increases, and ROI. Brought on an advisory board of physicians from Jefferson Health, Penn Medicine, and Mount Sinai through direct networking. Designed two structured pilot frameworks for validation.
If there’s one thing I’ve learned working on this… this problem is real and the solution itself is validated. There’s a ton of large hospital networks tackling this with a similar product. I got it to concept validation with a working prototype and physician advisors. Time to execute!
What I’d do is run 3 or 4 in-person pilots at Jefferson Health and Penn Medicine, compare AI-matched enrollment against manual baseline, and use the results to apply to a health tech accelerator. The prototype is built. The pilots are designed. See matchway.org
