Release Roadmap

Last updated: March 2026 · Monthly harvey ball updates

Data freshness: Epic data refreshes quarterly. CDW last refreshed March 2025 (no longer receiving updates).

SHIPPED Production ready ACTIVE In development PLANNED On roadmap

Complete Substantial Partial Early Planned

H High Priority   M Medium   L Low

---

v1.0

Shipped — v1.0.0

March 2026 · 271 commits · 3 contributors · CDM v5.4 · Vocabulary v5.0

SHIPPED

Patient Identity Stabilization

Foundational cross-source identity management across legacy Cerner and modern Epic EHR systems. Stable person_id across data loads via Medallion architecture (Bronze/Silver/Clustering/Gold) with graph-based clustering. Near-100% Epic PAT_ID coverage (742K/742K matched).

80+ commits

SHIPPED

HIPAA De-identification

Production de-identification pipeline with cohort_person_filter and deident_driver for safe data exports.

35 commits

SHIPPED

Drug & Condition Eras

Longitudinal exposure and disease duration analysis via DRUG_ERA and CONDITION_ERA tables.

Core Clinical

SHIPPED

Cross-Project Subsampling

40-patient reproducible sample with alias_prefix isolation. Supports multi-project dbt builds.

20 commits

SHIPPED

CI/CD & Test Framework

Automated dbt test pipeline, data quality checks, and deployment automation.

Infrastructure

SHIPPED

OHDSI Agent MCP Servers

Three MCP servers in production: vocabulary search (7.4M+ concepts), CIRCE cohort compiler, and lineage sidecar. 82.1% tool-routing accuracy across 820 evaluation tasks.

AI Tooling

SHIPPED

Custom Vocabulary Builder v0.3.0

Flowsheet mapping pipeline — 17,745 items mapped from unmappable pool. Clinical reasoning pass completed on review batches.

Vocabulary

SHIPPED

Documentation Site & DQD

Public documentation site, Data Quality Dashboard with completeness, conformance, and plausibility metrics. ARES integration live.

Community

CQ1

CQ1 2026 (Calendar Quarter 1)

January – March · Brain Health, FALCON-Bladder, CIRCLE, SON Prototype

H

ACTIVE

Brain Health v1.0

First project-specific OMOP deployment for the Personalized Brain Health Initiative (Goizueta ADRC). BrainHealthEnterprise dbt models and subsample infrastructure. Serves 34 researchers across 12 departments (SOM, SON, RSPH).

Brain Health ~Mar 20–21

H

ACTIVE

FALCON-Bladder Data Readiness

Data readiness assessment for FALCON-Bladder restart. Execute 3 SQL scripts (general concepts, genomic concepts, episode concepts) against Emory OMOP, generate CSV outputs, submit to study coordinators. First monthly meeting March 24.

OHDSI Oncology Due Apr 15

H

ACTIVE

ARPA-H CIRCLE AP1 Solution Summary

AP1 = single-award data platform role for ARPA-H critical illness digital twin program. Solution summary due March 30; full proposal May 28 if encouraged. Consumes the OMOP-on-FHIR Streaming MVP. ~$2M/3yr.

Grants Due Mar 30

M

ACTIVE

School of Nursing — Prototype

1M patient subsample created for SON experimentation against MVP — milestone achieved. Nursing cohort definitions in development.

Research Teams Milestone ✓

CQ2

CQ2 2026 (Calendar Quarter 2)

April – June · v1.1 Notes & NLP, CASSIDY, SON Export, Winship R01 Decision

H

PLANNED

v1.1 — Notes, NLP & Governance Infrastructure

Clinical notes access with governance, NLP extraction pipeline, and shared API platform enabling standard LLM/NLP pipelines across the institution. Brain Health is the first consumer. Includes CDW notes archival from Cerner sunsetting (BMI intermediate storage). BMI HPC cluster POC informs production requirements.

Clinical Notes MVP NLP Extraction Pipeline Community Shared Governance Infrastructure API Platform for LLM/NLP Pipelines

NOTE / NOTE_NLP

H

PLANNED

CASSIDY Phenotype

Diabetes surveillance computable phenotype for the CASSIDY Network (CHOA + Emory). Multi-stage pipeline: identification, classification, complication staging (§4.1–4.4), multi-year index 2018–2025.

Pediatrics Beginning CQ2

H

PLANNED

School of Nursing — Export

Deliver data export to School of Nursing for independent research use. Prototype and 1M patient subsample already created (CQ1 milestone).

Research Teams Early CQ2

M

PLANNED

OMOP-on-FHIR Streaming MVP

Prototype to experimentally validate OMOP as a real-time streaming storage target. FHIR-to-OMOP architecture (Kafka + DuckDB vocabulary resolution + hot/cold OMOP stores). Feeds ARPA-H CIRCLE AP1 if funded.

Streaming

M

PLANNED

CVB Flowsheet Expansion

~38K remaining unmappable flowsheet items routing through CVB pipeline for expanded measurement coverage.

Vocabulary

M

PLANNED

Winship Oncology R01

Clinical trial matching LLM incorporating OMOP. R01 submitted — decision expected May 2026.

Grants Decision May '26

CH2

CH2 2026 (Calendar Half 2)

July – December · NLP Production, SDOH, OHDSI Agent GA

H

PLANNED

NLP Pipeline — Production

Production entity extraction with quality metrics. Researcher-queryable NLP outputs in NOTE_NLP table.

NOTE_NLP

M

PLANNED

OHDSI Agent GA

General availability release of the MCP-native OHDSI agent with production-grade tool routing and evaluation framework.

AI Tooling

M

PLANNED

SDOH & Geocoding Enhancement

Expanded social determinants of health data. GaCTSA geocoding of patient and care site addresses planned for contribution back to Enterprise OMOP.

OBSERVATION

L

PLANNED

Provider Specialty Mapping

Improved provider specialty concept mapping for more accurate attribution and network analysis.

PROVIDER

2027

2027 Horizon

Long-range initiatives

H

PLANNED

OMOP on UDP / Azure Cloud Transition

Migrate OMOP into the Unified Data Platform on Azure. Enables ATLAS-based cohort discovery at speed, de-identified research data for cross-institutional collaboration.

Infrastructure CQ1 2027

M

PLANNED

Imaging & DICOM Linkage

Radiology report integration and DICOM metadata linkage to OMOP clinical events. Gates Foundation grant submitted.

Extension

L

PLANNED

Waveform Data

ECG waveforms and bedside monitor data integration for high-frequency clinical signal research.

Extension

L

PLANNED

Advanced NLP Products

LLM-assisted clinical note summarization, phenotyping from unstructured text, and multi-modal data integration.

AI Tooling

---

Data Product Maturity

Domain	Table	CDW	Epic	Combined	Notes
Demographics	PERSON				Stable person_id shipped
Encounters	VISIT_OCCURRENCE				~12M unmapped visits/yr
Visit Detail	VISIT_DETAIL				ICU/OR segments
Conditions	CONDITION_OCCURRENCE
Condition Eras	CONDITION_ERA				Shipped v1.0.0
Medications	DRUG_EXPOSURE
Drug Eras	DRUG_ERA				Shipped v1.0.0
Labs	MEASUREMENT				LOINC mapped
Vitals/Flowsheets	MEASUREMENT				CVB expanding (17K mapped, ~38K remaining)
Procedures	PROCEDURE_OCCURRENCE
Providers	PROVIDER				Specialty mapping improving
Social Hx	OBSERVATION				Smoking, alcohol; SDOH expansion planned
Notes	NOTE				v1.1 target
NLP	NOTE_NLP				CQ2 2026
Device	DEVICE_EXPOSURE
Imaging	—				CQ4 2026
Waveforms	—				CQ4 2026

---

• Back to Roadmap Overview

  ---

  Lucidchart interactive diagram and GitHub project board links.

  Roadmap Overview

• GitHub Project Board

  ---

  Real-time status on individual work items (core team access).

  Open Project Board