Effective coverage of institutional deliveries under the Janani Suraksha Yojana programme in high maternal mortality provinces of India: analysis of data from an annual health survey

Abstract

Background

Three-quarters of India's maternal deaths come from nine (of 35) provinces. In 2005, the government launched a conditional cash transfer programme, Janani Suraksha Yojana (JSY), to reduce maternal mortality through promotion of institutional deliveries (ID). Despite steep rises in ID, recent reports have been unable to detect a significant fall in the maternal mortality ratio (MMR). This suggests that merely increasing coverage of ID may not dent maternal mortality and indicates the need for measuring effective coverage of ID rather than just ID. Crudely, effective ID coverage is a measure of ID proportions that actually have access to emergency obstetric care (EmOC). Given the lack of data on EmOC functions, caesarean section proportion in an area could be used as a crude proxy indicator of availability/utilisation of EmOC. We estimate the effective coverage of ID in nine provinces.

Methods

Population-level estimates of ID and caesarean section proportions for all 284 districts in nine provinces were sourced from the Annual Health Survey (2010–11). WHO recommended a minimum 5% caesarean section proportion to be considered as 100% need met. District caesarean section rates, expressed as the proportion of need met, were calculated. This fraction was multiplied by the ID proportion to give an effective coverage. District-level predictors of effective coverage were estimated in a regression model.

Findings

In 139 districts, median ID proportion and effective coverage of ID were equal at 62% each, while in 145 districts these were 51%, and 35%, respectively. Bivariate association between MMR and effective coverage in the districts was stronger than with simple ID coverage. Literacy, urbanisation, poverty, and fertility had a significant association with effective coverage of ID.

Interpretation

We present a proxy method for estimating effective coverage of ID. A gap exists between access to just ID and access to EmOC, possibly indicating that mothers delivering in institutions are not necessarily receiving appropriate or adequate care. This study suggests effective coverage of ID is a better indicator for monitoring safe motherhood interventions than mere ID coverage. Ideally, effective ID coverage should be measured using a proportion of all obstetric complications treated; however, this study used caesarean section rate as a proxy indicator of EmOC utilisation in view of limited data availability.