Marketplace · BD / Sponsor partnerships
CuRE Commons
Community-pooled research engagement — site capabilities, sponsor matchmaking, brokered introductions.
What it does
Community-pooled research engagement — site capabilities, sponsor matchmaking, brokered introductions. Data-as-currency reciprocity: sites get free Cue, cleaned-data give-back, and network analytics while sponsors trade engagement rather than raw data.
Key capabilities
- Structured site capability listings
- Sponsor research-interest publishing
- Introduction / match brokering
- Data-as-currency reciprocity for sites
- Live federated feasibility count
- Protocol-AI → computable OMOP eligibility
- Post-match outcome feedback loop
- RWD-verified site performance
What sets it apart
- Data-as-currency reciprocity: sites get free Cue, cleaned-data give-back, and network analytics while sponsors trade engagement rather than raw data.
- Live network patient counts against eligibility criteria — the TriNetX wedge — on CuRE's own federated OMOP.
- Diversity analytics, reusable feasibility questionnaires, live federated counts, shadow profiles, protocol-AI criteria ingestion, and FMV budget benchmarking extend the marketplace.
- Site capability profile is computed from real OMOP data, not a self-reported PDF.
- OHDSI-aligned — designed to work with the partner network sponsors already trust.
Score how well a site fits a sponsor's research interest
Commons is a community-pooled research marketplace: sites publish structured capability listings, sponsors publish research interests, and the platform scores the fit. Pick a sponsor interest and a site, toggle the site's capabilities or drag a patient-count target, and watch the deterministic per-metric → domain → overall match score recompute — including the hard-fail filter that disqualifies a site outright. The ranked panel scores every synthetic site the same way.
Phase II autologous CAR-T · CGT experience is a hard requirement
Academic CGT hub · apheresis + cell-processing lab on-site
A numeric capability scores 1 once it meets the sponsor's minimum-desired target, and degrades linearly below it.
Each shared metric scores 0..1 (booleans/enums exact-match, numerics graded against the sponsor's target), rolls up to a weighted domain mean, then to an overall weighted score. Only metrics present on both sides are scored — Commons matches on the demand side's declared criteria, not on a self-reported PDF.
All sites · ranked against this interest
| Site listing | Match | Disposition |
|---|---|---|
Meridian Cancer Institute Academic CGT hub · apheresis + cell-processing lab on-site | 96% | Strong fit |
Lakeside University Health High-volume academic oncology center · no CGT program yet | 0% | Disqualified |
Riverbend Community Research Community site · broad access, lighter infrastructure | 0% | Disqualified |
Northshore Digital Health OMOP-native, Conduit-connected · deep data, CGT-naive | 0% | Disqualified |
Every score is computed in your browser from the synthetic capability + interest profiles — no backend. Note a site can post a high raw capability count yet be disqualified the moment it fails a hard-fail filter (e.g. a sponsor requiring OMOP-mapped data or CGT experience): a fit is more than a feature checklist.
Why this is more than a toy
The scorer is a faithful, dependency-free port of Commons' real operational match score, from two source files under apps/commons/src/server/matching/: the score.ts per-metric closeness + category/overall weighted rollup (booleans/enums exact-match, numerics graded against the interest's minimum-desired target; each domain a weighted mean, then domains rolled to an overall 0..1), and the service.ts inner-join that scores only the metrics present on both sides — the demand side declaring which capabilities matter (per ADR-CMN-008). The hard-fail filter (ADR-CMN-002) forces the overall score to 0 when a site lacks a capability the sponsor hard-requires. The synthetic site + sponsor profiles are built over the product's real v1 capability catalog domains (prisma/seed.ts, ADR-CMN-004). Per ADR-PLT-044 this deterministic operational score lives in-app (it is the always-available floor); the network-scale trained recommendation model is a separate advisory Calculate artifact and is not what this demo ports. Everything runs client-side on synthetic data — no backend, no real data.
See CuRE Commons in action
Every research ecosystem is unique. Let's discuss how CuRE can be configured for your needs.