Event Detection
Transformer encoders process longitudinal events — claims, pharmacy, and labs — weighting time gaps, care settings, and medication patterns to forecast the next event.
See the next health event
before it costs everyone.
Ruya Health pairs transformer-based risk prediction with orchestrated care workflows — so payors and employers can act on the right member, at the right moment, with measurable ROI.
Built on transformer architectures for event-sequence modeling
SAMPLE COHORT (ILLUSTRATIVE)
POP-7831 · East Region
POP-7831 · East Region
RISK INDEX
62.4/100
62.4/100
EVENTS · 24H
1,284 processed
1,284 processed
NEXT-EVENT WINDOW
14 days
14 days
HbA1c rising+0.6 pts
CHF cohort312 mbrs
Outreach scheduled+47
ED risk · 14d8.4%
The Ruya Platform
Risk detection, care orchestration, and proof — in one loop.
An event-sequence AI stack pairs transformer-based risk detection with orchestrated interventions, enabling prevention-first operating models.
Care Orchestration
Care plans translate risk into outreach sequencing, eligibility checks, and concierge workflows — delivered inside contact centers, EHRs, and member experiences.
Closed-loop ROI
Closed-loop analytics quantify savings, quality lifts, and experience metrics across populations — with attribution back to specific interventions and cohorts.
How it works
A single event sequence, end to end.
From the first claim to the next intervention — Ruya stitches risk signals into a single timeline that operations, clinicians, and contact centers can act on.
01
Ingest longitudinal events
Streaming connectors land claims, pharmacy, eligibility, and lab data — normalized into a per-member event sequence that preserves time gaps and care settings.
02
Forecast the next event
Transformer encoders weight medication patterns, utilization, and chronic-disease trajectories to forecast events 7–90 days ahead, with factor-level explanations for clinician review.
03
Trigger the right workflow
Care plans translate scores into outreach sequencing, eligibility checks, and concierge handoffs — delivered as serverless APIs inside your existing stack.
04
Measure & close the loop
Every intervention writes back to the cohort. Closed-loop analytics attribute savings and quality lifts to specific actions — not just dashboards.
// member.event_sequence
mbr_8f31a · 64F · CHF + T2D
Annual wellness visit−182d
CPT 99397 · BP 138/86 · A1c 7.1
claim · primary careMetformin refill gap−74d
PDC dropped 0.91 → 0.62 over 8 weeks
pharmacy · adherenceA1c trend rising−41d
A1c 7.1 → 8.3 · loinc 4548-4
lab · trajectoryEdema-related PCP visit−9d
ICD-10 R60.0 · referral to cardiology pending
claim · symptomRuya forecast · 14d window (Illustrative)
CHF exacerbation · ED visitP = 0.78
Inside the platform
Risk scores you can defend. Actions you can ship.
Factor-level insights highlight the drivers influencing risk — and the recommended actions for clinician review.
MH
Maria H. (Sample)
mbr_8f31a · 64 · F · plan PPO-N3
HIGH RISK
78/100
Next-event risk · 14 days
CHF exacerbation · ED visit likely without intervention
Top drivers · contribution to risk score
A1c trajectory · 7.1 → 8.3 (+1.2)
+14 pts
Metformin PDC drop · 0.91 → 0.62
+11 pts
Edema symptom claim (R60.0)
+9 pts
Recent PCP touchpoint
−4 pts
BASELINE 48 · NET +30SCORE 78 / 100
Recommended next actions
Cardiology referral — concierge booked
Surface 3 in-network cardiologists with <7d availability. Handoff to nurse navigator on accept.
SLA 24hChannel SMS + concierge
Adherence outreach — Metformin
Pharmacy partner refill nudge + benefits eligibility check for 90-day supply.
SLA 48hChannel App + pharmacy
Daily weight monitoring
Ship connected scale; alert care team if >2 lb / 24h or >5 lb / 7d gain.
SLA 7dChannel Device + EHR
Built to deliver
Prevention-first economics, proven population by population.
Prevention-first economics — what we're optimizing our platform to achieve across deployed cohorts. Numbers below are design targets, not historical customer results.
Projected spend
$47T
Projected global healthcare spend by 2030.
Chronic care savings
30%
Estimated reduction opportunity in chronic care costs.
PMPM Impact
Measurable
PMPM impact across rising-risk cohorts.
Time-to-action
Faster
Time from signal to action for care navigators.
Built for the operators
For payors and self-insured employers.
For payors
Bend the trend on your highest-cost cohorts.
Replace fragmented stratification with a single event-sequence model. Wire risk signals into your existing care management stack — without ripping out workflows.
- EHR + claims integration via FHIR / X12
- Stars, HEDIS, and risk-adjustment uplift
- Concierge and contact-center workflows
- Closed-loop attribution for VBC contracts
For employers
Get to the right benefit at the right moment.
Surface members heading toward avoidable cost — and route them to the benefits you already pay for: navigation, COE, behavioral, GLP-1 stewardship, and more.
- Point-solution orchestration, not duplication
- GLP-1 eligibility & step-care guardrails
- Quarterly cohort-level ROI reporting
- Member-facing experience SDK
Integrations roadmap — designed to drop into your stack.
Serverless APIs and EHR integrations deliver guidance inside the workflows, contact centers, and member experiences you already operate.
Note: Active integrations are listed in our docs. The items below represent planned and supported targets.
FHIR R4
X12 837/835
Epic · App Orchard
Cerner
Salesforce HC
Twilio Flex
Snowflake
Databricks
Datavant
Surescripts
Okta SSO
SCIM 2.0
Frequently asked
Honest answers.
What data does Ruya Health require?
We need historical claims (medical + pharmacy) to train the baseline, plus a daily or weekly delta feed. Lab feeds (HL7) and eligibility files are highly recommended but not strictly required for V1 deployment.
How do you handle compliance and data residency?
We follow standard security practices: TLS 1.2+ in transit, AES-256 at rest, scrypt password hashing, session revocation on sign-out, rate-limited authentication. We are not yet independently audited under SOC 2, HITRUST, or HIPAA. If you are evaluating Ruya Health for a use case that requires those, please contact us — we will be transparent about timeline and scope. Details: ruyahealth.com/data-protection.
Can we bring our own models?
Yes. If your data science team has already trained risk models, you can deploy them into our orchestration engine via our custom model API.
What does implementation look like?
Implementations average 6 to 12 weeks, depending on data cleanliness and the number of integrations required. A dedicated technical account manager will guide your team through data mapping, historical backtesting, and workflow integration.
Ready to look ahead?
Join the early access program and see what our transformer models can find in your historical claims data.