Papers

How does progress towards the child mortality millennium development goal affect inequalities between the poorest and least poor? Analysis of Demographic and Health Survey data

BMJ 2005; 331 doi: https://doi.org/10.1136/bmj.38659.588125.79 (Published 17 November 2005) Cite this as: BMJ 2005;331:1180
  1. Kath A Moser, lecturer (kath.moser{at}ons.gov.uk)1,
  2. David A Leon, professor1,
  3. Davidson R Gwatkin, consultant on health and poverty2
  1. Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London WC1E 7HT
  2. 6483 Wishbone Terrace, Cabin John, Maryland 20818, USA
  1. Correspondence to: K A Moser
  • Accepted 13 October 2005

Introduction

The millennium development goals (MDGs) have been widely accepted as a framework for improving health and welfare worldwide. Child mortality is one of the most crucial and avoidable global health concerns. In many low income countries, 10-20% of children die before reaching 5 years (compared with, for example, 0.7% in England and Wales). The child mortality MDG (to reduce the under 5 mortality rate by two thirds between 1990 and 2015) is formulated as a national average. The World Health Report 2003 posed an important question: how does progress towards the MDGs affect equity? We investigated this by examining, across a range of settings, how inequality in the under 5 mortality of the poorest and least poor changes as progress is made towards the MDG.

Participants, methods, and results

Using published data1 we examined changes in inequalities in under 5 mortality within 22 low and lower middle income countries (11 in Africa, five in Latin America or the Caribbean, and six in Asia) each with two Demographic and Health Surveys between 1991 and 2001 (http://www.measuredhs.com/). These countries encompass high and lower mortality situations, varied sociodemographic conditions, and in 2000 they accounted for 27% of the world's population. Under 5 mortality was estimated, using standard methods, from information on births in the 10 years preceding the survey derived from birth histories collected from women of reproductive ages. Socioeconomic position was described using an index of household wealth calculated from information on ownership of household assets (for example, a radio), housing characteristics (for example, floor materials), drinking water source, toilet facilities, and availability of electricity. The method is described elsewhere.2 Households, ranked by wealth index, were split into five groups each containing 20% of individuals and representing the poorest up to the least poor quintiles of the population. Under 5 mortality rates (deaths under age 5 per 1000 live births) were calculated for each quintile and the rate ratio (ratio of mortality in poorest and least poor quintiles) used to describe relative inequality. Inequality was considered to have increased or decreased over time if the rate ratio changed by at least ±10%.

National under 5 mortality rates vary between 30 and 250 deaths per 1000 live births (table). In all surveys mortality is higher in the poorest as compared with the least poor quintile. Most rate ratios lie within the range 1.5 to 3.0 and almost all the 95% confidence intervals exclude 1.0. Thirteen countries had statistically significant declines in overall under 5 mortality between surveys. Despite this, in only four of the 13 countries did the rate ratios decrease; five saw increasing rate ratios. None of the changes over time in the rate ratios were statistically significant.

Trend in level and inequalities in under 5 mortality in 22 low and lower middle income countries

View this table:

Comment

We confirm that there are large and persistent inequalities in under 5 mortality within many low and lower middle income countries and show that improvements in national under 5 mortality, in line with the MDG, do not necessarily bring about decreasing inequalities in mortality between the poorest and least poor in society. Indeed, such society-wide improvements seem as likely to be accompanied by increasing as decreasing inequalities. This finding indicates the importance of monitoring under 5 mortality among different socioeconomic groups. It also argues for reformulating the child mortality MDG to incorporate an equity dimension and thus provide an impetus to adopt policies that tackle health inequalities.

For most countries considered here the Demographic and Health Surveys give the only nationally representative data on child mortality and are thus widely used for that purpose. However, as shown by the width of the confidence intervals in our analysis, thesedata allow us to give only indicative results rather than make statistically robust assessments of trends in inequality in under 5 mortality. This points to a need to strengthen health information systems for equity purposes.3

What is already known on this topic

Progress is being made in some low and lower middle income countries towards achieving the millennium development goal on under 5 mortality

What this study adds

National improvements in under 5 mortality, in line with the millennium development goal, are as likely to be accompanied by increasing as decreasing inequalities in child mortality within countries; adding an equity dimension to this goal would give an impetus to adopting policies that tackle health inequalities

This article was posted on bmj.com on 11 November 2005: http://bmj.com/cgi/doi/10.1136/bmj.38659.588125.79

A version of this paper was presented at the Global Forum for Health Research, Mexico City, November 2004. We thank Eldaw Abdalla Suliman and Agbessi Amouzou for supplying us with the standard error data used in calculating the confidence intervals.

Footnotes

  • Contributors The idea was developed by KAM in discussion with DRG and DAL. KAM analysed the data and drafted the paper. All authors contributed to the interpretation of the data and the development and writing of the final manuscript. KAM is guarantor.

  • Funding KAM was supported by the Dreyfus Health Foundation.

  • Competing interests None declared.

  • Ethical approval Not needed.

References

View Abstract