Answers are grounded in this paper and cite their sources.
Background Routine health information systems (RHIS) such as the District Health Information Software (DHIS2) are essential for evidence-based health system planning and monitoring, yet few studies examine trends in data quality over time. This study presents findings from a longitudinal data quality assessment (DQA) conducted across six provinces between 2019 and 2024, using both DHIS2 data and facility-level verification. Methods We conducted a longitudinal DQA across 121 health zones at three timepoints. Trends in reporting completeness and timeliness were assessed using DHIS2 data for eight indicators. To assess accuracy, we compared data from a sample of health facilities with corresponding DHIS2 values. Verification factors, reporting rates, and Bland-Altman analysis were used to assess concordance across rounds. We assessed changes in average bias using repeated measures ANOVA design. Results DHIS2 reporting completeness improved to over 90% in all provinces studied. Timeliness also improved over time. Accuracy of DHIS2 reporting, measured by verification factors, showed good concordance with health facility registers. Indicators for under-five uncomplicated malaria, diarrhea, and pneumonia cases did not show significant differences between data sources. However, underreporting of ITNs distributed during ANC1 and overreporting of BCG vaccinations and ANC1 visits remained consistent across rounds. Conclusion The quality of RHIS data in DHIS2 improved over the study period, particularly in completeness and consistency. Persistent indicator-specific biases highlight the need for improvements in data recording and reporting practices. With regular data quality audits and appropriate interpretation, DHIS2 data can support routine monitoring and program evaluation in DRC and similar contexts.
No takes yet. Share an insight, caveat, or question.
Niles et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: