Linking Household and Service Provisioning Assessments to Estimate a Metric of Effective Health Coverage: A Metric for Monitoring Universal Health Coverage

TL;DR

This study develops a new metric to measure effective health coverage by combining household and service data, revealing significant gaps in care quality in rural India.

Contribution

A novel composite metric of effective coverage is proposed by integrating multiple data sources and adjusting for service quality.

Findings

Effective coverage estimates dropped significantly compared to crude coverage across all linking approaches.

Postnatal care had the lowest effective coverage (10.1%) while immunization had the highest (78.7%).

Structural and process quality adjustments revealed large disparities between exact and ecological linking methods.

Abstract

Background: The framework of measuring effective coverage is conceptually straightforward, yet translation into a single metric is quite intractable. An estimation of a metric linking need, access, utilization, and service quality is imperative for measuring the progress towards Universal Health Coverage. A coverage metric obtained from a household survey alone is not succinct as it only captures the service contact which cannot be considered as actual service delivery as it ignores the comprehensive assessment of provider–client interaction. The study was thus conducted to estimate a one-composite metric of effective coverage by linking varied datasets. Methods: The study was conducted in a rural, remote, and fragile setting in India. Tools encompassing a household survey, health facility assessment, and patient exit survey were administered to ascertain measures of contact coverage and quality. A gamut of techniques linking the varied surveys were employed such as (a) exact match linking and (b) ecological linking using GIS approaches via administrative boundaries, Euclidean buffers, travel time grid, and Kernel density estimates. A composite metric of effective coverage was estimated using linked datasets, adjusting for structural and process quality estimates. Further, the horizontal inequities in effective coverage were computed using Erreygers’ concentration index. The concordance between linkage approaches were examined using Wald tests and Lin’s concordance correlation. Results: A significantly steep decline in measurement estimates was found from crude coverage to effective coverage for an entire slew of linking approaches. The drop was more exacerbated for structural-quality-adjusted measures vis-à-vis process-quality-adjusted measures. Overall, the estimates for effective coverage and inequity-adjusted effective coverage were 36.4% and 33.3%, respectively. The composite metric of effective coverage was lowest for postnatal care (10.1%) and highest for immunization care (78.7%). A significant absolute deflection ranging from −2.1 to −5.5 for structural quality and −1.9 to −8.9 for process quality was exhibited between exact match linking and ecological linking. Conclusions: Poor quality of care was divulged as a major factor of decline in coverage. Policy recommendations such as bolstering the quality via the effective implementation of government flagship programs along with initiatives such as integrated incentive schemes to attract and retain workforce and community-based monitoring are suggested.

Figures