Child mortality estimation: accelerated progress in reducing global child mortality, 1990-2010

Abstract

Monitoring development indicators has become a central interest of international agencies and countries for tracking progress towards the Millennium Development Goals. In this review, which also provides an introduction to a collection of articles, we describe the methodology used by the United Nations Inter-agency Group for Child Mortality Estimation to track country-specific changes in the key indicator for Millennium Development Goal 4 (MDG 4), the decline of the under-five mortality rate (the probability of dying between birth and age five, also denoted in the literature as U5MR and (5)q(0)). We review how relevant data from civil registration, sample registration, population censuses, and household surveys are compiled and assessed for United Nations member states, and how time series regression models are fitted to all points of acceptable quality to establish the trends in U5MR from which infant and neonatal mortality rates are generally derived. The application of this methodology indicates that, between 1990 and 2010, the global U5MR fell from 88 to 57 deaths per 1,000 live births, and the annual number of under-five deaths fell from 12.0 to 7.6 million. Although the annual rate of reduction in the U5MR accelerated from 1.9% for the period 1990-2000 to 2.5% for the period 2000-2010, it remains well below the 4.4% annual rate of reduction required to achieve the MDG 4 goal of a two-thirds reduction in U5MR from its 1990 value by 2015. Thus, despite progress in reducing child mortality worldwide, and an encouraging increase in the pace of decline over the last two decades, MDG 4 will not be met without greatly increasing efforts to reduce child deaths.

Figure 1. Illustration of the loess fitting procedure.

Senegal is a country that has experienced substantial fluctuations in the rate of change of under-five mortality over the last 40 years. Two loess fitted trend lines are shown, one with α = 1, and the other the UN IGME 2011 trend line where α was defined using the standard α calculation. Generally, the greater the α value used in the loess fitting procedure, the more longer-term trends in the data influence the final trend line.

Figure 2. Global and regional under-five, infant, and neonatal mortality rates and deaths, 1990–2000.

(A) U5MR and under-five deaths; (B) IMR and infant deaths; (C) NMR and neonatal deaths.

Figure 3. Under-five mortality rate by country for 2010.

Data for Sudan refer to the country as it was constituted in 2010, before South Sudan seceded on 9 July 2011.

Figure 4. Annual rate of reduction in the under-five mortality rate compared to annual rate of reduction in the neonatal mortality rate over the periods 1990–2000, 2000–2010, and 1990–2010.