Operational study of the quality of health data aggregated by community health workers in the Upper East Region of Ghana

Objectives: Assess the quality of health data compiled by community health workers (CHWs) by comparing district data summarized from data aggregated by CHWs during routine operation to data aggregated by independent researchers. Methods: Compared summary reports aggregated from patient data by CHWs (“site report”) to summary reports compiled by data audit experts using the same raw data (“validated report”). Completeness and accuracy of data from site reports, including number of antenatal care visits, postnatal follow-up visits, immunizations, and malaria cases were compared to that of independently validated reports. Predictors of data quality such as community health worker, community, and system characteristics were included in the analysis using binary logistic regression. Results: All sites provided site reports to the district health management teams, but there were significant errors in the data aggregated by CHWs. Some variables were over-reported in site reports whereas others were under-reported. Staff shortages during community outreach were associated with declines in CHWs recording information about clinical encounters in patient registers or in patient health books. Conclusion: Quality of data compiled by CHWs is important to monitoring site performance, informing district level decision-making about budget allocations, and guiding national policies. This highlights the importance of streamlining data aggregation processes as a pathway to improve evidence-based decision making; at the community health center as well as higher levels of the health system.