Establishing a centralized data mart from the Rakai community cohort study to improve HIV research in Rakai, Uganda. Issue 2 (18th May 2022)

• Record Type: Journal Article
• Authors: Ndyanabo, Anthony; Footer, Kevin; Ahmed, Tanvir; Glogowski, Alex; Whalen, Christopher; Ssekasanvu, Joseph; Ssentongo, Lloyd; Lutalo, Tom; Nalugoda, Fred; Ha, Grace K; Rosenthal, Alex
• Is Part Of: JAMIA open. Volume 5:Issue 2(2022)
• Publication Date: 2022-05-18
• Subjects: data management -- data mart -- data reproducibility -- infectious diseases; Medical informatics -- Periodicals; 610.285
• DOI: 10.1093/jamiaopen/ooac032
• Languages: English
• ISSNs: 2574-2531

Abstract

Abstract: To improve timely access to quality HIV research data, the Rakai Health Sciences Program (RHSP) Data Mart was developed to store cohort study data from a legacy database platform in a modernized system using standard data management processes. The RHSP Data Mart was developed on a Microsoft SQL Server platform using Microsoft SQL Server Integration Services with custom data mappings and queries. The data mart stores 20+ years of longitudinal HIV research data and includes standard processes for managing data, data dictionary, training materials, and a library of queries to fulfill data requests and load new data from completed survey rounds. The RHSP Data Mart enables efficient querying and analysis of multidimensional research data by simplifying data integration and processing. A sustainable database platform with well-defined data management processes promotes data accessibility and reproducibility, enabling researchers to advance their understanding and management of infectious diseases. Lay Summary: This article highlights the approach that was taken to develop a unified data mart for Rakai Health Sciences Program (RHSP) to migrate 20+ years of HIV research data from legacy systems to a modern, sustainable, database platform to optimize data management process. Data from multiple domains such as Survey, Census, and Lab were extracted from 100s of individual files, with 1000s of columns, and integrated into a custom data model designed to resemble prior file structures enabling users to easily access and analyze longitudinal research data. The original implementation of the data mart included data for first 14 rounds of study, and it was developed to accommodate future study rounds with minimal code changes. A data freeze process and change tracking method were introduced to improve data quality, reproducibility, and integrity. Last, a comprehensive data dictionary, user guides, tutorial videos, and query libraries were created to expedite user adoption. The RHSP Data Mart significantly reduces processing time for data managers to fulfill collaborator requests with accurate and reproducible HIV research data while providing a foundation for the organization to highlight their work to the global research community.