LOMAinsights Healthcare operating intelligence
Talk to the team See the Demo

The Layer Underneath

The unified data layer your stack doesn't have.

Most medical groups, plans, and hybrid networks run on five to ten disconnected systems — EHR, practice management, HRIS, payer portals, GL. LOMAinsights builds the governed, HIPAA-eligible data layer that reconciles them all, so your analytics, dashboards, and AI agent reason on a single source of truth.

Google Cloud Healthcare API BigQuery system of record HIPAA BAA FHIR R4 native CMEK encryption

What this layer delivers

  • One canonical model. Division × month × measure, conformed across every source.
  • Scheduled pipelines. Daily, weekly, and monthly ingest with validation.
  • Governed semantics. Every KPI defined once, used everywhere.
  • PHI minimized. Only the operational fields needed for analytics.
  • Auditable. Lineage from any number back to its source row.

Architecture

Source systems land in Google Cloud, get conformed in BigQuery, and surface through the LOMAinsights semantic layer to your dashboards and AI agent.

Source systems

EHR (Athena, Epic, eCW)
PM / Billing
HRIS
Payer portals
GL / NetSuite
Surveys
→

Google Cloud · system of record

Cloud Healthcare APIFHIR R4 · HL7v2 · DICOM
Cloud Storage (Bronze)Immutable raw landings, versioned
BigQuery (Silver)Canonical, partitioned, CMEK-encrypted
Dataform / dbtConformance + tests
Cloud Scheduler + Cloud RunOrchestrated pipelines
→

LOMAinsights layer

Semantic catalog (Gold)Governed metrics, formulas, thresholds
Diagnostic AgentAnomaly + cross-domain reasoning
Predictive EngineTime-series forecasting
Analytics Demo + APIsDashboards · BI · write-back

Why Google Cloud as the system of record

Built for healthcare compliance

Google Cloud signs a HIPAA Business Associate Agreement covering Cloud Healthcare API, BigQuery, Cloud Storage, Cloud Run, and Cloud Scheduler — the exact services we use. Every customer ingest path is HIPAA-eligible by design.

FHIR R4 native

Cloud Healthcare API speaks FHIR R4, HL7v2, and DICOM natively. EHR data lands in a managed FHIR store with built-in de-identification, consent management, and audit logging — no bespoke integration to maintain.

BigQuery as the warehouse

BigQuery is the analytic system of record: serverless, columnar, partitioned, with row- and column-level security and customer-managed encryption keys (CMEK). It scales from a single FQHC to a multi-state IPA without re-architecting.

Tenant isolation

Every customer gets a dedicated GCP project (or dataset) with its own service identity, KMS keys, and audit trail. Cross-tenant data movement is structurally impossible.

Integrated ML and AI

BigQuery ML and Vertex AI live next to the warehouse, so forecasts and embeddings run without exporting data. The LLM only sees pre-aggregated, governed context — never raw rows.

Audit + governance

Cloud Audit Logs capture every data access, query, and pipeline run. Combined with Data Catalog and Dataplex, lineage and access governance are first-class — not bolt-ons.