World Abortion Estimates: An Audit

Part I--What are the Numbers, and Where Do They Come From?

By Laura Antkowiak and Randall K. O'Bannon, Ph.D.

Editor's note. This important but widely unreported aspect of the abortion debate requires extensive examination and much documentation. This series will run over the next several months and eventually be placed on NRLC's web page at www.nrlc.org.

Whenever there are special U.N. conferences, whenever there is a debate in Congress about aid to organizations that promote abortion overseas, at some point you're likely to hear someone assert that there are 30, 40, 50 million abortions a year (or more) in the world with tens of thousands of "abortion-related" or maternal deaths due to "unsafe abortions." Abortion's promoters argue that figures like these prove not only the need of more money for "family planning" (which includes abortion) but also the medical necessity to legalize abortion worldwide.

What is the source of such claims? And is there any solid statistical basis for these assertions? National Right to Life decided to try to track down the source of these claims and find out how and where the advocates of abortion got their numbers. It was not easy.

What we did discover is that there is an abundance of oft-times questionable assumptions and very often skimpy hard data to document the assertions. Not surprisingly the areas of the world with the most incomplete methods of tracking abortions and abortion complications are listed as having the most "unsafe" abortions and abortion deaths.

Estimates of "Unsafe" Abortions and Abortion Deaths

Check on different sites on the Internet and you'll see estimates of annual world abortions ranging from 20 million (Maryland Sierra Club) to 88 million (One World Supersite) abortions. Among the recent publications most commonly cited in academic studies or governmental literature, however, are the Alan Guttmacher Institute's Sharing Responsibility: Women Society & Abortion Worldwide (1999); Unsafe Abortion, a 1998 report from the World Health Organization (WHO); and two articles by Christopher Murray of Harvard and Alan Lopez of WHO that estimate worldwide mortality by cause, including abortion mortality.1

These sources put the number of annual world abortions at 46 million,2 with some 20 million of these officially deemed "unsafe." Both WHO and AGI estimate the number of maternal deaths associated with abortion at 78,000.3

The precision implied in such numbers is highly misleading. The numbers cited for developed nations, such as those of North America and Western Europe, are relatively low. These figures tend to come from government health records or comprehensive national data collection, which are thought to be fairly accurate. WHO charts do not even list "unsafe abortion" or "unsafe abortion death" rates for North America or for Western Europe because it says the "incidence is negligible."

Figures given for developing countries and regions, however, where researchers report the vast majority of "unsafe" abortions and abortion-related deaths (see chart), are based on much more meager data and a lot of assumption-laden extrapolations. Many of these countries do not maintain detailed birth or mortality records, much less abortion statistics, making even the roughest of estimates problematic.

Little Official Data in the Developing World

WHO's highest estimates of "unsafe abortions" and abortion-related deaths come from the developing world. Again, these are the regions that generally do not keep official statistics on abortion. Estimates for abortions in these regions are constructed by WHO researchers from material in its "unsafe abortion database" (discussed in Part II of our series).

Data for most countries in Sub-Saharan Africa (every African region except middle-income Northern Africa4), South America, and Southcentral and Southeastern Asia, the regions to which WHO assigns the highest numbers of unsafe abortions and abortion-related deaths, come almost entirely from "hospital studies," "community surveys," and "estimates." Each of these comes with certain limitations.

Another type of data that also shows up in the listings for developing countries is what WHO calls "public source" data. Typically, a "public source" is a journal article, report, or unpublished document, often from a pro-abortion organization, raising questions about its objectivity. Part II of this series will discuss these studies and their methods in more detail.

Not only is the quality of unsafe abortion data suspect, but its quantity is sometimes thin as well. In Uruguay, for example, WHO's data listing consists of "public source" data with sample sizes of 5, 14, and 23 individuals, as well as five other pieces of "public source" data with unidentified sample sizes. These all come from the 1970s and 1980s. While, anecdotally, these cases may be informative, any projection based on such limited data is highly speculative and simply unreliable. The box [at location] lists a sampling of other egregious examples.

Despite the implied exactness of WHO's numbers, many countries in the developing world do not even have basic registration systems of births and deaths. Further, according to Murray and Lopez, even among those that do, these national vital registration systems "are often faulty." In a number of countries, the system is thought to be incomplete, and sometimes not all deaths are medically certified.

Not only are the countries with minimal or nonexistent records generally the least developed countries, but--probably not coincidentally--WHO tends to find the most unsafe abortions and abortion deaths in these countries. Murray and Lopez show that the only country in Sub-Saharan Africa with a vital registration system, for example, is South Africa.

Though WHO does not publish abortion estimates by country, the region it calls Southern Africa--which contains South Africa and four other middle-income countries--contributes the lowest number of Africa's estimated unsafe abortions and abortion deaths. These countries, however, constitute only a small subset of the otherwise very poor area known as Sub-Saharan Africa.

There are no national vital registration systems in place in the rest of those Sub-Saharan countries. The numbers WHO reports for the rest of the region are a great deal higher, to the point that Sub-Saharan Africa accounts for nearly over 40% of WHO's worldwide abortion-related deaths.

That same theme--that the largest numbers of unsafe abortions and abortion-related deaths are found in regions with the sketchiest data--resurfaces when one examines the main data and methods used by WHO to estimate "unsafe" abortions and abortion-related deaths. We shall do this in Part II, in next month's NRL News.

Laura Antkowiak is NRL-ETF special research assistant.

Randall K. O'Bannon, Ph.D., is NRL-ETF director of education and research

Notes:

1. WHO estimated the current most commonly cited numbers of illegal or "unsafe" abortions and abortion-related deaths. AGI uses WHO's numbers for illegal or "unsafe" abortions. The AGI report also presents a figure for total worldwide abortions; it estimates the number of worldwide legal abortions from national statistics that are adjusted for incompleteness and presumed underreporting, based on "experts' judgment." Guttmacher also provides more individual estimates on selected countries, drawn from "official data," surveys, and adjusted estimates from hospital studies. The Murray and Lopez articles provide estimates of abortion-related deaths. Prior to WHO's report, this work had been a commonly cited source for abortion-related death figures. Because Murray and Lopez's descriptions of data and methodology are not specific to abortion, this report will focus on WHO's derivation of figures and complement that with comments from Murray and Lopez on estimating mortality generally.

2. AGI's report contains the 46 million figure. Apparently the United Nations does not mind tacking on a few million abortions here and there: its latest publication on abortion policies (2001) says that "approximately 50 million abortions" take place worldwide. That source is a 1994 WHO report.

3. Murray and Lopez estimate abortion mortality at a lower level, 61,000, in their 1994 and 1996 articles. It is unclear what degree of overlap exists between the data and methods used by Murray and Lopez, and those used by WHO. Murray and Lopez, however, revise their figures downward for many causes of deaths in much of the developing world, since their sum exceeded the estimated sum of total deaths of given age-sex groups. They say this is because experts' estimates--which they rely upon to calculate mortality from abortion and many other causes in the developing world where official statistics are incomplete or nonexistent--tend to be higher than official records indicate. Abortion deaths are included among the adjusted figures.

4. WHO includes low-income Sudan in its definition of Northern Africa. The World Bank, however, from which our definition of Sub-Saharan Africa comes, includes Sudan in Sub-Saharan Africa.