Background: Accurate routine surveillance data are essential for malaria control and elimination. However, the multistep reporting process used in most malaria-affected countries can introduce discrepancies between health facility registers and national health management information systems, often based on the District Health Information System 2 (DHIS2). This study assessed the concordance of malaria rapid diagnostic test (RDT) data across facility registers, monthly summary forms (MSFs) and the DHIS2 in four sub-Saharan countries.
Methods: In 2023, we conducted an observational study in Benin, Côte d'Ivoire, Nigeria, and Uganda, using harmonized tools and methods. In each country, 16 public primary health care facilities were selected from two regions. The total number of RDTs and positive results from health facility registers, MSFs and the DHIS2 were compared over three to five months. We assessed concordance using Bland-Altman plots, weighted absolute percentage error (WAPE)-based aggregate data reporting accuracy (WADRA), and verification factors (VFs). System- and facility-level differences were examined by stratifying indicators by region, baseline outpatient volume and test positivity rate.
Results: Across 64 facilities, 104,396 RDTs (58,304 positives, 55.8%) were recorded in registers, compared to 112,435 (62,903 positives, 55.9%) in MSFs and 110,771 (62,761 positives, 56.7%) in DHIS2. Benin showed the highest concordance across data sources, while Nigeria and Uganda had the lowest. Positive RDT results were more likely to be reported than total RDTs, particularly in Nigeria, where VFs indicated consistent overreporting (mean VF 0.74, 95% CI: 0.61-0.89). WADRA analysis showed low reporting accuracy for positive RDTs in 6 (38%) Nigerian and 5 (31%) Ugandan facilities. Regional differences were notable in Nigeria and Uganda. In Nigeria, higher outpatient volume was associated with lower concordance; no trend was seen for baseline test positivity rate.
Conclusions: Substantial variation in RDT data concordance was observed across countries and facilities. Concordance was strongest between MSFs and DHIS2, suggesting data entry was not a significant issue. Strengthening routine data validation and using accuracy and direction-sensitive metrics, such as WADRA and VFs, could improve malaria data reliability. Further research should explore system-level factors influencing data quality and identify scalable solutions.
Keywords: DHIS2; Data quality; Discrepancy; Malaria; RDTs; Surveillance.
© 2025. The Author(s).