The worldwide abortion rate was essentially stable between 2003 and 2008, after having declined during the previous eight years, a new analysis indicates.1 About 44 million abortions were performed in 2008, which translates to 28 per 1,000 women of reproductive age, a rate only slightly lower than that five years earlier (29 per 1,000). More than four-fifths of these abortions took place in developing countries, where more than half (56%) of abortions were unsafe.

The new study was a sequel to prior analyses that estimated worldwide abortion statistics for 1995 and 2003. The researchers estimated numbers and rates of abortion by country and region, and compared the findings to those for the earlier years. They obtained data from a variety of sources, assessed its quality and made adjustments as needed to compensate for shortcomings.

As in the prior analyses, the researchers generally considered abortions unsafe if they were performed in a country with highly restrictive abortion laws, or if they were done in a country with less restrictive laws but did not meet the country’s legal requirements. Safe abortions were those that met legal requirements in a country where abortion was allowed on request or on socioeconomic grounds, or where, if laws were interpreted liberally, abortion was legal to preserve a woman’s physical or mental health. However, the researchers deviated from these classifications when called for by empirical evidence; for example, although abortion is broadly legal in India, many pregnancies are terminated by unauthorized providers and categorized as unsafe.

In two-thirds of the countries with liberal abortion laws, formal abortion statistics were available. These data were used unaltered if experts familiar with the country considered the statistics at least 95% complete; if the data did not meet this threshold, the number of abortions was increased by 5–154% (mean, 26%) to yield a more accurate estimate. Similarly, adjustments for underreporting were made to data from countries whose only abortion estimates were from national surveys. If no official statistics, survey data or estimates were available, the investigators assigned estimated rates on the basis of the country’s fertility and contraceptive rates and other factors.

For countries with restrictive abortion laws, the investigators used data from official sources, nongovernmental organizations, databases and experts, giving preference to estimates published in peer-reviewed journals. In some cases, numbers and rates were estimated using data on the number of women who received medical treatment for abortion complications and the estimated proportion of abortion recipients who required treatment for complications. In other cases, survey-based data were used and adjusted for underreporting. For some countries (most of them small), the investigators generated national estimates by extrapolating from subnational data, or by assuming that rates mirrored those of countries that were nearby or had similar abortion laws, fertility rates and levels of contraceptive use.

Overall, an estimated 44 million abortions were performed in 2008, slightly more than in 2003 (42 million) but fewer than in 1995 (46 million). About 38 million (86%) were in developing countries, including 27 million in Asia, six million in Africa and four million in Latin America; the number of abortions increased by 2.8 million in developing countries, but declined by 0.6 million in developed regions. The worldwide proportion of pregnancies ending in abortion changed little between 1995 (22%) and 2008 (21%).

The global rise in the number of abortions was entirely due to increased population size, as the number of abortions per 1,000 women aged 15–44 declined slightly between 2003 and 2008, from 29 to 28 per 1,000, and was essentially unchanged in developing countries (29 per 1,000). Likewise, regional rates for 2008 in Africa (29 per 1,000), Asia (28 per 1,000) and Latin America (32 per 1,000) were similar to the corresponding 2003 rates. The stability in abortion rates suggests a stalling of the decline seen between 1995 and 2003, when the global abortion rate fell from 35 to 29 per 1,000. A linear regression analysis revealed that abortion rates tended to be lower in subregions where the proportion of reproductive-age women living under liberal abortion laws was higher.

Within regions, abortion rates varied widely. In Africa, for example, the rate ranged from 15 per 1,000 in Southern Africa to 38 per 1,000 in Eastern Africa; similar variation occurred in Asia (26–36 per 1,000) and Latin America (29–39). Subregional rates, like regional ones, generally changed little between 2003 and 2008; the most notable exceptions—a decrease from 24 to 15 per 1,000 in Southern Africa, and an increase from 26 to 36 per 1,000 in Middle Africa—were probably attributable at least in part to fluctuations in data quality, though a true decline in abortion incidence likely occurred in Southern Africa.

The proportion of abortions that were unsafe was slightly higher in 2008 (49%) than in 2003 (47%) and 1995 (44%). This trend was most marked in Western Asia, where the proportion rose from 42% to 60% between 1995 and 2008. Fifty-six percent of abortions in developing countries were unsafe in 2008, compared with 6% of those in developed countries.

Limitations of the analysis include the underreporting, misreporting and bias typically seen in abortion statistics. The investigators believe that their adjustments largely corrected for these problems; the intervals of certainty that they calculated suggest that the actual rates of abortion are roughly within 10% of their estimates, both globally and regionally. They further note that the stability in the global abortion rate between 2003 and 2008, following a period of decline, mirrors trends in contraceptive use, levels of which rose during the 1990s but stabilized in the ensuring decade. "Measures to reduce the incidence of unintended pregnancy and unsafe abortion," including "improving access to family planning services and safe abortion care, are crucial steps" toward reducing maternal mortality, the researchers conclude.

–P. Doskoch

REFERENCE

1. Sedgh G et al., Induced abortion: incidence and trends worldwide from 1995 to 2008, Lancet, 2012, 379(9816):625–632.