Goal & Strategy

• Goal / Problem we want to solve: Make OpenMRS-collected data easier to use for Researchers and Program Decision Makers (e.g. Researchers like (e.g. Epidemiologists, Health System Strengthening, Public Health, Whole-Health/OneHealth, etc; and Program Decision Makers like Population/Public Health leads).

• Strategy: (1) Translate OpenMRS to the OMOP Common Data Model and (2) set up a re-useable tooling pipeline for data to be extracted for research question use.

Outcomes

1. Standardized Data: A fully converted OpenMRS dataset from Kisumu County into OMOP CDM, ready for analysis using OHDSI tools.

2. Improved Research Capabilities: Enhanced data analysis access leading to better health outcomes and stronger research capacities.

3. eLearning Resources for Local Capacity Strengthening: Increased local expertise in data management and analysis.

4. Documented Knowledge: Comprehensive documentation of the data conversion process for future replication in other regions.

Plan

Milestone 1: Build Community Understanding of Researchers, their Questions, and their Tools

Q4 CY24 -

Q2 CY25

Goal: We need to know, “What kinds of questions do people want to answer?” with OpenMRS data, and “What tools do researchers use?”

1. I.d. people to interview:

   1. OpenHIE: WhatsApp group and conference

      1. Grace Potma

   2. People with known interest in using OMOP for OpenMRS Data Research:

      1. Kisumu County & OHDSI contacts: Dr. Gregory Ganda, Dr Andy Kanter - Jan Flowers ASAP

      2. UW Malawi project (what is their target use-case? current tools?) Jan Flowers Nov

      3. Rwanda Implementer (Ian & Jan to remember)

   3. Post to community forum

      1. … Grace Potma ASAP

   4. Search Talk for past OMOP interest

      1. … Grace Potma ASAP

   5. Google it:

      1. Looks like a masters' student? https://github.com/M-Gwaza/OpenMRS-to-OHDSI

   6. Researchers we find through searching for terms “OpenMRS OMOP”

      1. Reach out to Barry Levine about this OpenMRS-OMOP Mapping & ETL paper and their findings: https://www.sciencedirect.com/science/article/pii/S2666990023000277

         1. https://github.com/sikder-sab/OPENMRS_OMOP_PROJECT

      2. https://github.com/sikder-sab/OPENMRS_OMOP_PROJECT

   7. Follow Up with Julia connection

      1. Grace Potma set up call

2. Interview/survey researchers across public health departments and implementers, academic colleagues in global health

Milestone 2: Assessment of OpenMRS database landscape

Q4 CY24 -

Q2 CY25

1. Goal: Check for unexpected data model structures/workarounds happening in the real-world which we should know about going-in.

   1. Data Inventory:

      1. e.g. Conduct a full assessment of the OpenMRS schema used by the healthcare facilities in Kisumu County.

      2. Assessment questions:

Is the data already patient-level and extracted?

Milestone 2.2: Decide if we need to review de-identified patient data (and then obtain necessary approvals for data extraction and documentation).

Milestone 3: Use Interview Findings to Guide PRACTICAL Use Cases

Q1 CY25 - Q2 CY25

• Goal: I.d. Practical Use Cases and Validate with community

Milestone 4: Data Mapping

Q1 CY25 - Q3 CY25

• Goal: Map OpenMRS data model elements to OMOP CDM for standardized data use, with a focus on the practical use cases identified above.

• Terminology Mapping:

• Leverage CIEL (Columbia International eHealth Lab) to map the OpenMRS concepts to the OMOP

• CDM standard concepts.

• Identify gaps in the current terminology and work closely with Andrew Kanter (CIEL) to address

• these through custom mappings or proposing additions to standards bodies.

• Consider how local adaptations of the data model should be accounted for in the mapping.

Milestone 5: Data Extraction Pipeline

Q2 CY25 - Q1 CY26

• ETL (Extract, Transform, Load) Pipeline Development and validation:

• Design and implement a robust ETL pipeline that can extract the required data from OpenMRS,
  transform it according to OMOP standards, and load it into the OMOP database.

• Ensure the pipeline is scalable for use beyond Kisumu County and can be adapted for other LMIC
  regions.

• Address specific challenges related to patient-level data extraction if needed.

• Testing & Validation

Milestone 6: Testing & QA

Q1 CY26 - Q2 CY26

1. With OHDSI Tools:

• Run tests using OHDSI tools to ensure the converted data set is functional for advanced health
  data analytics.

2. Check for Data Changes:

   1. Cross-check that mapped terminologies are correct and that no critical data is lost during

      transformation.

Milestone 7: Develop Documentation and eLearning Resources for Local Capacity Building

Q1 CY26 - Q2 CY26 (END of GRANT)

4. Capacity Strengthening
   Goal: Build local expertise for future data standardization and analysis.

• Academy Video on understanding of OpenMRS tooling and pointers to the OMOP CDM, ETL processes, and OHDSI tools.

• Identify and work with community partner to develop documentation and training materials for Kisumu County teams to maintain the ETL pipeline and conduct future data conversions without external assistance.

Process Documentation: Document what we did, and how we did it.

Scope Boundaries:

• We will not be implementing this in production ourselves, that is outside the scope of this project. For example, in the Kisumu case, we would recommend that a in-country vendor partner to support such implementation.

Resources

https://github.com/M-Gwaza/OpenMRS-to-OHDSI

https://github.com/sikder-sab/OPENMRS_OMOP_PROJECT