Payer
Health plans and employers
Population value, intervention prioritisation, HEOR and benefit-contract economics.
Company
RISX Science
RISX Science builds governed analytic infrastructure for health plans, provider and dental organisations, and life-science teams — infrastructure in which the provenance of a result matters as much as the result itself.
Mission
Make the reasoning behind a clinical decision inspectable
Health care generates more predictive output than it can act on, and far more than it can defend. Our mission is narrower and harder: identify the specific people for whom an available clinical decision would change the expected outcome, and show the complete reasoning behind that judgement.
An answer that cannot be reconstructed is not an answer. Every result the platform produces carries its inputs, its method, its model versions, its validity checks and a replay manifest — or it is refused with an explicit reason.
Problem thesis
Why we built it this way
Three failures recur across health analytics. The platform is designed around avoiding them rather than apologising for them.
Risk is mistaken for opportunity
The highest-risk people are frequently the people a decision can no longer help. Ranking by risk misdirects finite clinical capacity. We rank by modeled incremental benefit of a specific, available decision.
Absence is mistaken for evidence
A missing result is not a negative result. Facts in the platform carry explicit states — observed, inferred, contradicted, stale, unknown, refused — and each state changes what an engine is permitted to conclude.
Results outlive their justification
A number that survives its own audit trail becomes folklore. Bitemporal facts, versioned artifacts and replay manifests keep every figure attached to the evidence and code that produced it.
Operating approach
How we work with an organisation
Engagements move in one direction: from a decision worth improving, to the evidence that can carry it, to a governed production capability.
- Define the decision, not the datasetWe start from a clinical or benefit decision with real capacity behind it, and specify the analytic question as a target trial.
- Establish the governed evidence layerSource systems are mapped into bitemporal facts with states, dates, provenance and identity separation before any engine runs.
- Develop and validate enginesModels are developed on retrospective data with temporal splits, then evaluated for discrimination, calibration, balance and overlap.
- Operate with inspection and monitoringOutputs run through domain workflows with registry-controlled versions, monitoring, refusal gates and a documented kill switch.
Capabilities
What the company builds
Four analytic products and the governance layer that keeps them distinct.
Analytic products
| Product | Answers |
|---|---|
| Phenotype detection | Is an unrecognised condition likely present now? |
| Event prognosis | What is the risk of a defined severe event over a fixed horizon? |
| Conditional intervention benefit | What would a specific intervention be expected to change? |
| External clinical / control arm | Is a fit-for-purpose external comparator available at all? |
Platform capabilities
| Capability | What it provides |
|---|---|
| Bitemporal evidence layer | Facts with state, valid time and transaction time |
| Individualised value policy | Ranking by expected net benefit rather than risk |
| Evidence inspection | Data, method, rules, validity and replay on every result |
| Governance registry | Versioned models, rules and artifacts with environment gates |
| Health-economic modelling | Clinical impact, plan cost and contract value on separate ledgers |
Markets served
Where the work applies
The same evidence layer supports four distinct buyers with different obligations.
Provider
Clinical organisations
Pathway intelligence and clinician-led review of the cases where a decision still has room to work.
Dental
Dental organisations
Cross-domain detection at the health contact many people keep most reliably.
Life sciences
Research organisations
Feasibility, target-trial emulation, external comparators and replayable evidence.
Company
Organisation and governance
RISX Science is a private company building the platform described on this site.
Leadership and advisers
Leadership and scientific advisory profiles are published as each appointment is confirmed. Enquiries about the team, references or diligence materials are answered directly.
Scientific and data governance
Model versions, rule sets, environment gates and inspection requirements are documented in the platform itself, so a reviewer can examine the controls rather than take them on description.