Community-based intervention packages facilitated by NGOs demonstrate plausible evidence for child mortality impact.

INTRODUCTION Evidence exists that community-based intervention packages can have substantial child and newborn mortality impact, and may help more countries meet Millennium Development Goal 4 (MDG 4) targets. A non-governmental organization (NGO) project using such programming in Mozambique documented an annual decline in under-five mortality rate (U5MR) of 9.3% in a province in which Demographic and Health Survey (DHS) data showed a 4.2% U5MR decline during the same period. To test the generalizability of this finding, the same analysis was applied to a group of projects funded by the US Agency for International Development. Projects supported implementation of community-based intervention packages aimed at increasing use of health services while improving preventive and home-care practices for children under five. METHODS All projects collect baseline and endline population coverage data for key child health interventions. Twelve projects fitted the inclusion criteria. U5MR decline was estimated by modelling these coverage changes in the Lives Saved Tool (LiST) and comparing with concurrent measured DHS mortality data. RESULTS Average coverage changes for all interventions exceeded average concurrent trends. When population coverage changes were modelled in LiST, they were estimated to give a child mortality improvement in the project area that exceeded concurrent secular trend in the subnational DHS region in 11 of 12 cases. The average improvement in modelled U5MR (5.8%) was more than twice the concurrent directly measured average decline (2.5%). CONCLUSIONS NGO projects implementing community-based intervention packages appear to be effective in reducing child mortality in diverse settings. There is plausible evidence that they raised coverage for a variety of high-impact interventions and improved U5MR by more than twice the concurrent secular trend. All projects used community-based strategies that achieved frequent interpersonal contact for health behaviour change. Further study of the effectiveness and scalability of similar packages should be part of the effort to accelerate progress towards MDG 4.