Illuminate Health

In-Warehouse Product

EMPI Engine + Workbench

EMPI Engine + Workbench gives healthcare data teams transparent patient identity resolution and analyst-grade stewardship without moving PHI outside your cloud.

It is the same model as the rest of Illuminate Health delivery: forward-deployed engineers help you ship and operationalize an in-warehouse product your team can inspect, tune, and own.

  • Warehouse-native matching
  • Human-in-the-loop review
  • Auditable golden records
Book a demo See implementation plan

Problem

Patient identity breaks quietly, then breaks trust.

Most teams can list the symptoms: duplicate records, split histories, and brittle patient-level metrics. The hard part is fixing identity without creating a black box.

EMPI Engine + Workbench is designed for teams that need confident matching logic plus human review workflows that stand up to governance scrutiny.

What goes wrong when identity is underpowered

  • Silent fragmentation: one person appears as multiple patients across claims, EHR, and operational sources.
  • Opaque logic: teams cannot explain why records matched or why they did not.
  • Stewardship bottlenecks: uncertain candidates pile up with no clear queueing or adjudication path.

Approach

Transparent probabilistic matching with practical governance.

EMPI Engine + Workbench combines a configurable matching engine and analyst workflow into one operating model: tune rules in your warehouse, route uncertainty for review, and keep every decision traceable.

  1. 01

    Configure openly

    Define blocking logic, thresholds, and survivorship behavior in a guided configuration layer with readable logic.

  2. 02

    Run in your warehouse

    Execute dbt-powered matching (batch or incremental) where your source data already lives, with no external PHI movement.

  3. 03

    Resolve with context

    Route uncertainty to stewards for split, merge, ignore, and golden record decisions with an audit trail.

Capabilities

What EMPI Engine + Workbench gives your team on day one.

Matching and configuration capabilities

  • Guided configuration for blocking and threshold strategy
  • dbt-native execution patterns for scheduled and incremental processing
  • Explainable score behavior for match candidate review and tuning
  • Golden record inputs aligned to survivorship rules

Stewardship and governance capabilities

  • Queue management with practical prioritization controls
  • Split, merge, and ignore workflows for ambiguous cases
  • Golden record review with survivorship oversight
  • Full audit trail for actions and outcome tracking
  • Admin controls that support repeatable governance

Implementation

A practical rollout path from first run to production stewardship.

1. Baseline and configure

Connect source domains, define a baseline matching policy, and align survivorship logic to downstream reporting expectations.

2. Run and calibrate in your warehouse

Execute EMPI Engine + Workbench in dbt, review candidate quality, and tune thresholds with your analysts and data engineering leads.

3. Operationalize stewardship workflows

Promote queue operations and audit reporting into day-to-day practice so identity quality improves continuously, not quarterly.

Currently supports Snowflake and Databricks Compatible with the tuva project No external PHI movement required

See EMPI Engine + Workbench on your data model.

We will walk through architecture, matching logic, and stewardship workflows with your team and map a practical launch plan.

Book a demo