ABSTRACT
Objectives: Estimates from robust cross sectional surveys have brought into question the accuracy of HMIS of VAS in India, including potential errors as HMIS data is rolled up at various levels from health worker- to state-level, limiting the usefulness of these data for program feedback and improvement. In response, we sought to identify errors in monitoring data at different levels, suggest corrective action and establish coverage by a cross sectional survey triangulated with HMIS data. Methods: This exercise was conducted in 16 districts where 48 primary health centers (PHCs) were selected randomly. Data on Vitamin A coverage recorded at districts, PHCs, sub-health centers (SHCs) and health worker/volunteer level were cross-verified with coverage figures recorded at their respective constituent units. Discrepancies between coverage data recorded at health facilities and that aggregated from their constituent units were computed. A 30X7 cluster survey was also undertaken in each of these centers to establish population-based coverage. Results: Overall, there was a 7% discrepancy between figures reported by PHCs and that aggregated from their SHCs. In some PHCs the discrepancy was as large as 40%. The cross-sectional household survey identified 10% higher coverage than that reported by the state government (52% vs 42%). District specific findings were used to identify and facilitate systemic changes in reporting and data compilation. Conclusions: This is a cost-effective method and useful exercise for improving internal data quality and minimizing reporting error. The involvement of government functionaries in the process increased the ownership of study results and facilitated corrective actions.