Advancing Sexual and Reproductive Health and Rights
 
International Family Planning Perspectives
Volume 34, Number 1, March 2008
DIGEST

Sub-Saharan Africa Is Sole Exception to the Global Decline in Maternal Deaths

Maternal mortality decreased worldwide during a recent 15-year period, but some regions and countries did not experience any meaningful decline, according to a recent analysis.1 For countries with relevant data, the maternal mortality ratio fell by roughly 3% per year between 1990 and 2005; however, the ratio remained stable in Sub-Saharan Africa, in low-income countries and in countries where the ratio was higher to begin with. In 2005, the estimated worldwide maternal mortality ratio was 402 deaths per 100,000 live births, with 95% of these deaths occurring in Sub-Saharan Africa and Asia.

To address shortcomings of previous estimates of maternal mortality and to develop new estimates that allow comparisons across countries and over time, researchers assessed maternal mortality in 171 countries during 2005 using only nationally representative sources of data. The maternal mortality ratio (the number of women who died from pregnancy-related causes during pregnancy or the postpartum period, divided by the number of live births) was estimated using methods tailored to the type of data source. Trends in the ratio between 1990 and 2005 were assessed by two methods: a time-series regression analysis that used data from only the 125 countries with empirical data and used countries as the unit of analysis; and a reestimation analysis that used data from all 171 countries and weighted the changes by the number of births during this time period. In addition to calculating ratios by country, the researchers analyzed the data by region, national income (using World Bank classifications), type of data source and initial maternal mortality ratio.

The type and quality of data sources on maternal mortality varied widely. Fifty-nine countries (accounting for 13% of births worldwide in 2005) had satisfactory civil registration data; six (1% of births) had complete registration of deaths but an excessive proportion of deaths attributed to ill-defined causes; 28 (16% of births) had data from sisterhood surveys, which collect information regarding siblings and tend to underestimate female mortality; four (5% of births) had data from surveys on mortality of reproductive-age women; two (32% of births) had data from civil registration systems implemented in a representative sample of geographic areas; five (2% of births) had data from census–based estimates; six (5% of births) had other empirical data, generally of unknown quality; and 61 (25% of births) lacked appropriate empirical data, necessitating the use of models.

The researchers estimated that 535,900 maternal deaths occurred worldwide during 2005, which translates to a maternal mortality ratio of 402 maternal deaths per 100,000 live births. The vast majority of the maternal deaths took place in Sub-Saharan Africa (50%) and Asia (45%). Fully 48% of the deaths occurred in only five countries: India, Nigeria, the Democratic Republic of Congo, Afghanistan and Ethiopia.

Across regions, the maternal mortality ratio differed sharply; it was lowest in developed countries (nine maternal deaths per 100,000 live births) and highest in Sub-Saharan Africa (905 per 100,000). Across countries, the range was even greater; Ireland had the lowest ratio (one per 100,000), while Sierra Leone had the highest (2,100 per 100,000).

In the time-series analysis of temporal trends, the worldwide maternal mortality ratio in countries with empirical data fell by about 3% per year between 1990 and 2005. Lower-middle-income countries experienced a reduction of 4% per year and upper-middle-income countries a reduction of 5% per year; in contrast, the ratio remained essentially stable in low-income and high-income countries. Findings also differed by type of data source: Countries with sisterhood surveys did not experience a decrease in maternal mortality ratio, whereas the ratio fell by 2% per year among countries with vital registration systems, by 7% per year among countries with reproductive-age mortality surveys and by 4% per year among those with other data sources. Finally, the researchers found that the maternal mortality ratio remained essentially unchanged among countries where the initial ratio was at least 200 maternal deaths per 100,000 live births, but declined among countries with lower initial ratios.

In the reestimation analysis of temporal trends, the worldwide maternal mortality ratio fell by only 5% between 1990 and 2005, corresponding to an average annual reduction of about 0.4%. (The annual change is smaller in this analysis than in the time-series analysis because of differences in the countries included and in the weighting of data.) The 15-year decline was largest in northern Africa (36%) and smallest in Sub-Saharan Africa (2%). In addition, the absolute number of maternal deaths fell by 7% worldwide during this time period and by 15–35% in every region but Sub-Saharan Africa, where the number actually increased by 28%.

The researchers assert that progress in reducing maternal mortality worldwide has been limited, and note that the "huge" differences among countries in the maternal mortality ratio—differences far exceeding those seen with other frequently used health indicators—suggest that disparities exist in the implementation of effective interventions for reducing maternal mortality. Noting that accurate measurement will be essential for assessing progress in reducing maternal deaths, they recommend international investment in efforts to improve the reporting and collection of data on maternal outcomes. To achieve Millennium Development Goal 5—reducing the global maternal mortality ratio to 25% of its 1990 level by 2015—will require "a huge and urgent emphasis on improved pregnancy and delivery care throughout the developing world," they conclude.—S. London

REFERENCE

1. Hill K et al., Estimates of maternal mortality worldwide between 1990 and 2005: an assessment of available data, Lancet, 2007, 370(9595):1311–1319.