Assessing the impact of a facility-based digital intervention on district-level data quality in rural Burkina Faso

Digital clinical decision support tools are increasingly used in primary-level health facilities to improve maternal and child healthcare. These tools guide health workers to input patient records digitally, providing a critical opportunity to strengthen routine health information systems with real-time data and improve quality of care. However, often, these service-level data are not always effectively integrated into existing data systems, reducing their downstream impact on data quality and data-driven decision-making. Our study evaluated the impact of a digital job aid tool (Registre Electronique de Consultations-Maternite, REC-Maternity tool), used by healthcare workers in rural primary healthcare facilities in Burkina Faso, on routine data quality at the district level.

First, we conducted an analysis comparing REC-Maternity data (n=79 895 visits) with District Health Information System 2 (DHIS2) data from 34 facilities to assess agreement in service delivery records. Next, in a quasi-experimental study design, we used DHIS2 data in Toma intervention and Gourcy comparison districts to conduct a difference-in-difference analysis with 13 months preintervention and postintervention time horizon. We assessed the impact of the intervention on three DHIS2 data quality outcomes: completeness, timeliness and internal consistency.

The validation analysis revealed limited data agreement between REC-Maternity and DHIS2, particularly for postnatal care (ratio: 1.56) and family planning (ratio: 3.05). These discrepancies suggest parallel data flows from paper-based and electronic forms, indicating limited integration of the digital tool into routine reporting. The results also suggested the potential for digital records to help distinguish true zero-reporting from missing values in DHIS2. The difference-in-difference analysis indicated a significant decrease in timely reporting in the intervention district compared with the comparison district (-66.5% of facilities (95% CI -73.0% to -60.1%)).

While digital tools offer a potential data revolution in resource-limited settings, they should not be presumed to be more efficient and replace paper-based data collection without continuous monitoring and quality improvement strategies.