MCE publications present detailed findings for each site as well as the results of cross-site analyses. The full set of findings includes both positive and negative results, and methodological as well as programmatic conclusions. The brief list below highlights only some of the key findings that have particular relevance to child health policies and programmes.

• Child survival efforts must begin with local epidemiology, targeting the major causes of death within each region, country, and even district.

• IMCI guidelines for the case management of ill children in health facilities, supported by high-quality training and supportive supervision, are the gold standard and should continue to be implemented widely. New MCE findings that IMCI is efficient and costs less than routine care in some settings are encouraging.

• To be successful in reducing child mortality, programmes must move beyond health facilities and develop new and more effective ways of reaching children with proven interventions to prevent mortality. In most high-mortality settings, this means providing case management services at community level, as well as focusing on prevention and on reducing rates of undernutrition.

• Coverage should be the driving force behind district, national, regional and global child survival programmes. Only by paying close attention to whether mothers and children receive interventions can we decipher whether the delivery methods are effective and equitable, and whether mortality reductions are likely to occur. Public accountability at all levels can bring delivery bottlenecks to the attention of all, and encourage rapid action to address them.

• Countries and districts should be encouraged to prioritise, and to put first those interventions known to be cost-effective in reducing under-five mortality. Better tools to support policy and decision makers in estimating the costs and impact of their choices are needed urgently.

• Ensuring that resources are available, not only for time-limited projects but in the longer term, is essential to allow sufficient time for planning and implementation to mature and yield impact.

• Much more policy-relevant research is needed and should be conducted at country level, especially studies that focus on the effectiveness of policies and strategies implemented by Ministries of Health and their partners.