Effect of community-based newborn care on cause-specific neonatal mortality in Sylhet district, Bangladesh: findings of a cluster-randomized controlled trial

Abstract

Objectives: Community-based maternal and newborn intervention packages have been shown to reduce neonatal mortality in resource-constrained settings. This analysis uses data from a large community-based cluster-randomized trial to assess the impact of a community-based package on cause-specific neonatal mortality and draws programmatic and policy implications. In addition, the study shows that cause-specific mortality estimates vary substantially based on the hierarchy used in assigning cause of death, which also has important implications for program planning. Therefore, understanding the methods of assigning causes of deaths is important, as is the development of new methodologies that account for multiple causes of death. The objective of this study was to estimate the effect of two service delivery strategies (home care and community care) for a community-based package of maternal and neonatal health interventions on cause-specific neonatal mortality rates in a rural district of Bangladesh.

Study design: Within the general community of the Sylhet district in rural northeast Bangladesh. Pregnancy histories were collected from a sample of women in the study area during the year preceding the study (2002) and from all women who reported a pregnancy outcome during the intervention in years 2004 to 2005. All families that reported a neonatal death during these time periods were asked to complete a verbal autopsy interview. Expert algorithms with two different hierarchies were used to assign causes of neonatal death, varying in placement of the preterm/low birth weight category within the hierarchy (either third or last). The main outcome measure was cause-specific neonatal mortality.

Result: Deaths because of serious infections in the home-care arm declined from 13.6 deaths per 1000 live births during the baseline period to 7.2 during the intervention period according to the first hierarchy (preterm placed third) and from 23.6 to 10.6 according to the second hierarchy (preterm placed last).

Conclusion: This study confirms the high burden of neonatal deaths because of infection in low resource rural settings like Bangladesh, where most births occur at home in the absence of skilled birth attendance and care seeking for newborn illnesses is low. The study demonstrates that a package of community-based neonatal health interventions, focusing primarily on infection prevention and management, can substantially reduce infection-related neonatal mortality.