Outlier detection before the claim arrives.
The same substrate the health system uses to retain a complex patient surfaces, at the cohort level, the patterns the payer needs to avoid one.
The same data, opposed sides, one chain.
Lucy reads the same patient records on both sides of the provider-payer line. The provider uses her to retain. The payer uses her to identify high-cost trajectories before they become unavoidable claims. The substrate is one substrate; the audit chain proves both sides got the same answer.
Cohort-level pattern recognition under glass.
Lucy ingests the population the payer covers and surfaces the cohorts where small interventions today change large outcomes twelve months from now: medication non-adherence patterns, post-discharge cliff risk, missed follow-ups that predict ED re-admissions, polypharmacy with adverse-interaction signatures.
Every signal is sourced. Every flag is signed into the chain at the moment of production. The actuarial team sees an evidence ledger, not a black-box score.
A record, not a recommendation.
- De-identified cohort access mediated by IRB or equivalent.
- Every outlier flag carries its provenance chain back to the source data and the literature.
- HIPAA business-associate agreement and signed BAAs across the covered-life data plane.
- Pricing on covered lives, not on flags raised. The incentive is to be accurate, not to flag.
A specific offer
We will run a sample outlier report against a de-identified test cohort you choose, returned in five business days, no contract.