Horizon Population Health API
AI-powered analytics — natural language querying, risk stratification, predictive models
Live API Status — Horizon Analytics
Probing production endpoints…
Rate Limits & SLA
Rate limit
500 queries/min
Burst
50 req/sec
Uptime SLA
99.9%
Support
Business hours · 4h
P95 latency
< 2s semantic
Retention
Query logs 90d
Natural Language Query
Translate questions into SQL against the FHIR data lake. "How many diabetic patients have A1C above 9%?"
Risk Stratification
Admission risk, readmission risk, ED utilization, and mortality prediction scoring.
Quality Measures
HEDIS, CMS Star Ratings, and custom quality measure calculation.
Natural Language Query Example
Request
POST /horizon/v1/query/nl
Authorization: Bearer {token}
{
"query": "How many diabetic patients have
an A1C above 9% in the last 6 months?",
"max_rows": 100
}Response (200 OK)
{
"sql": "SELECT COUNT(DISTINCT p.gpid) ...",
"result": {
"columns": ["count"],
"rows": [[ 847 ]],
"total_rows": 1
},
"explanation": "Found 847 patients with
Type 2 Diabetes (E11.x) and most recent
HbA1c (LOINC 4548-4) > 9.0%",
"execution_time_ms": 1240
}Query playground
Two-step NL → execute flow against the sandbox gateway. Requires a bearer token with horizon.query scope.
Horizon Portal (11 Pages)
Horizon API Endpoints
/horizon/v1/query/nlBearerNatural Language Query
/horizon/v1/query/executeBearerExecute SQL Query
/horizon/v1/scoreBearerRisk Stratification Score
Integration toolkit
Official SDKs
Apex Horizon certification track
12 required items · platform + product-specific checks
- First authenticated API call in sandbox
- OAuth or API key with least-privilege scopes
- FHIR R4 read against sandbox lake
- GPID boundary headers on write paths
- Webhook HMAC-SHA256 verification implemented
- Webhook idempotency by delivery ID
+9 more on the certification page
Complete all required items before production go-live