# Migrating longitudinal African mental health data from staging to the OMOP common data model within the INSPIRE network datahub

## TL;DR

The paper describes a framework for harmonizing African mental health data using the OMOP model, enabling global analysis and policy development.

## Contribution

A metadata-driven pipeline was developed to standardize and integrate African mental health datasets using OMOP and OHDSI tools.

## Key findings

- 202,013 person records and 7 million observations were successfully migrated to the OMOP model.
- Mapping completeness exceeded 99.9% with high data quality across all OMOP domains.
- Custom vocabularies enabled robust analysis of context-specific mental health outcomes.

## Abstract

The standardization and integration of longitudinal mental health data from African cohort studies are critical in advancing research and informing policy. There are several challenges posed by diverse sources, instruments adapted for locals, and the absence of an interoperable framework to allow for meaningful analysis and cross-study comparisons.

We designed and executed a metadata-driven pipeline using the OMOP Common Data Model within the INSPIRE Network Datahub to harmonise multi-country African mental health datasets. Data extracted previously from longitudinal studies, standardised via a snowflake schema staging database, is now mapped to OMOP vocabularies with local extensions, and validated through quality assurance protocols using OHDSI tools.

A total of 202,013 person records and over 7 million observations across fourteen cohort studies were successfully migrated. Mapping completeness exceeded 99.9%, with high conformance, completeness, and plausibility across all OMOP domains. Custom vocabularies ensured the coverage of context-specific exposures and outcomes, thereby supporting robust cohort construction, event characterization, and longitudinal analyses.

This framework demonstrates scalable harmonisation and integration of African mental health data, bridging the gap between local datasets with global standards. This then enables the performance of federated analysis and reproducible research, increasing the utility and impact of mental health data in informing evidence-based policies and future collaborative studies across Africa.