Examining coverage, content, and impact of maternal nutrition interventions: the case for quality-adjusted coverage measurement

BACKGROUND: Reductions in neonatal mortality remain stagnant, despite gains in health care access and utilization. Nutrition interventions during antenatal care (ANC) and in the immediate postpartum period are associated with improved neonatal outcomes. Adjusting coverage estimates for the quality of care provided yields greater insight into health system performance and potential population health benefits of accessing care. In this cross-sectional study, we adjust maternity care coverage measures for quality of nutrition interventions to determine the impact on infant birth weight and breastfeeding. METHODS: We used household data from the Malawi 2013-2014 Multiple Indicator Cluster Survey to assess use of maternal health services and direct observations of ANC and delivery from the 2013 Service Provision Assessment to measure nutrition interventions provided. We adjusted coverage measures combining self-reported utilization of care with the likelihood of receipt of nutrition interventions. Using adjusted log-linear regression, we estimated the associations of these nutrition quality-adjusted metrics with infant birthweight and immediate breastfeeding. RESULTS: Health facility data provided over 2500 directly observed clinical encounters and household data provided 7385 individual reports of health care utilization and outcomes. Utilization of ANC and facility-delivery was high. Women received nutrition-related interventions considerably less often than they sought care: over the course of ANC women received a median of 1.6 interventions on iron, 1 instance of nutrition counseling, and 0.06 instances of breastfeeding counseling. Nutrition quality-adjusted ANC coverage was associated with a reduced risk of low birthweight (adjusted relative risk [ARR] 0.87, 95% confidence interval (CI) = 0.79, 0.96) and increased likelihood of immediate breastfeeding (ARR = 1.04, 95% CI = 1.02, 1.07); nutrition quality-adjusted post-delivery care was also associated with greater uptake of immediate breastfeeding (ARR = 1.08, 95% CI = 1.02, 1.14). Based on these models, delivering nutrition interventions consistently within the existing level of coverage would decrease population prevalence of low birthweight from 13.7% to 10.8% and increase population prevalence of immediate breastfeeding from 75.9% to 86.0%. CONCLUSIONS: Linking household survey data to health service provision assessments demonstrates that despite high utilization of maternal health services in Malawi, low provision of nutrition interventions is undermining infant health. Substantial gains in newborn health are possible in Malawi if quality of existing services is strengthened.