How many mothers die from ‘unsafe abortion’?

Reproductive health lesson. Photo: WHO/Abir AbdullahMaternal deaths during pregnancy or childbirth, the vast majority in developing countries, constitute one of the world’s greatest hidden epidemics. The death toll is estimated at more than half a million mothers a year.

World leaders meeting at the UN agreed that reducing maternal mortality is essential to fulfilling their “collective responsibility to uphold the principles of human dignity, equality and equity at the global level.” UN Millennium Development Goal 5 aims at reducing maternal mortality by 75 per cent between 1990 and 2015, but it is a target most unlikely to be reached.

One important reason for that, says American doctor and researcher Donna J Harrison, is unreliable data on the causes of maternal death -- in particular abortion. In this interview with MercatorNet Dr Harrison explains what’s wrong with the abortion data and why it matters.


MercatorNet: There are several direct causes of maternal death in the developing world. A New York Times article, for example, listed the five leading ones in this order: bleeding, infection, high blood pressure, prolonged labour and botched abortions. Why single out abortion for analysis?

Donna Harrison: Each of the other causes of maternal death has clear definitions, and is not in itself connected to a political agenda. However, the international politics pushing worldwide legalization of elective induced abortion does not foster clear thinking about the maternal deaths and injuries which accompany that legalization.

In order to understand the policy implications of a decision, one must be able to assess the effects of that decision. In common with many of the other causes, however, the incidence of induced abortion in developing countries, along with the number of maternal deaths and complications from them is not accurately known, despite estimates that are published from time to time.

Mifepristone and misoprostol abortions in the United States have been linked with severe adverse events, and women could easily die from infections and haemorrhage in areas where they do not have immediate access to transfusion and surgical facilities. Thus, introducing chemical abortions, whether with mifepristone and misoprostol, or with misoprostol alone, in medically underserved areas will mean that these severe adverse events will become maternal deaths.

MercatorNet: What are the main obstacles to getting good abortion data?

Donna Harrison: There are three main problems. First, the use of the terms “safe” and “unsafe”, which are not scientific terms but more legal and political, because they are directed to changing the law. In a 2007 article co-sponsored by WHO, for example, unsafe abortion is defined as “abortions in countries with restrictive abortion laws”. That means that any abortion in such a country, no matter how medically superior the conditions, would be counted as “unsafe”.

This can produce some unintended and even amusing consequences. In 2007 I attended the UN-sponsored Women Deliver conference in London, which was dedicated to advancing maternal health. During the presentation of a paper estimating the worldwide number of “unsafe” abortions, a Marie Stopes International representative from a clinic performing abortions in a country where it is illegal rose in indignation and said, “By your definitions, are you saying that all the abortions performed in my clinic are unsafe?” The presenters did not answer her question.

The second major problem comes from the way WHO collects data on hospital admissions due to abortion. In a 2003 document drawn up in conjunction with the UN Population Fund it defines these admissions as “due to abortion (spontaneous and induced, but excluding planned termination of pregnancy)”. This could mean in practice, “Don’t count complications in women admitted precisely to have an abortion,” or, “Don’t count any women who have had complications from planned termination of pregnancy.” It’s not at all clear what statistics to count. And, since planned termination of pregnancy in a hospital implies legal abortion, this method would give a distorted picture of maternal deaths and morbidity, or ill health, under a legal regime.

The third major problem comes from the statistical manipulation of such data as it has. For example, in estimating the number of “unsafe” abortions and related morbidity, WHO combines spontaneous abortions and induced abortions, and then “corrects” for spontaneous abortion according to what it believes the proportion should be.

I should add that this guesswork extends beyond abortion to the whole field of maternal mortality in developing countries. At the Women Deliver conference a WHO researcher, Dr Cindy Stanton, admitted: “In some areas we make huge adjustments to make the numbers turn out right. More than fifty percent of some numbers are ‘adjusted.’” A physician from Benin asked her why WHO doubled the number of maternal deaths that country had reported. Dr Stanton replied that although they could not validate their methods of adjusting, “we did have twelve to fourteen people who worked on the criteria”. In other words, it was a matter of expert opinion.

MercatorNet: Does it matter very much if the numbers are not accurate? Governments do have to address all aspects of maternal health, including abortion, don’t they?

Donna Harrison: Actually, it matters enormously because of the implications for planning national policy. If a country finds that a large percentage of maternal deaths happen from a lack of skilled birth attendants at delivery, and a very small number come from induced abortions, then it becomes clear that funding should be directed to skilled birth attendants and not to abortion agencies.

But if a country where abortion is illegal uses methodology which allows spontaneous abortions to be counted as induced abortions, a falsely high number of maternal deaths may be attributed to “unsafe abortions”. The country will then be under international pressure to legalize abortion, on the assumption that it would result in the immediate lowering of the hospital admissions owing to abortion, and the associated costs.

As we have seen, however, this reduction would be more apparent than real, since adverse outcomes of planned and legal abortions might not be counted.

MercatorNet: What are your recommendations for improving abortion data in the interests of maternal health?

Donna Harrison: There is a clear and immediate need to define terms used by the UN and its member states. To make maternal mortality data understandable for policy purposes, a country must be able to clearly see the event which preceded the death of a mother.

Induced abortion should be a separate category from spontaneous abortion, and at least three sub-categories of induced abortion should be specified: (a) abortions medically necessary to save the life of the mother, which is rare; (b) voluntary induced abortion performed in a hospital setting; (c) voluntary induced abortion in the outpatient setting.

The last subcategory is increasingly important with the growing prevalence of misoprostol abortions and mifepristone/misoprostol abortions, especially in light of the known increased risk of haemorrhage and fatal infections resulting from them. Given that WHO is putting significant emphasis on its programmes promoting medical abortions in developing countries, this category is urgent and important.

MercatorNet: Are there any signs that things are moving in this direction, or is the goal of providing good care to all the world’s mothers going to continue to be undermined by a bias towards “safe abortion”?

Donna Harrison: I have concerns about the direction that the U.N. may pursue under the leadership of Secretary of State Hillary Clinton. In her recent statement before congress, Mrs Clinton stated clearly that the term “reproductive health” included abortion. This inclusion of abortion in the term reproductive health means an increasingly greater pressure to legalize abortion will be exerted on countries whose own people and legitimate governments have decided to ban it. The issue is significant because it becomes a kind of reproductive cultural imperialism from the United States and Europe.

That said, Millennium Goal 5 does not include a “reproductive health” target and the addition of such language was again rejected at a meeting of the UN Human Rights Council in Geneva this month. The HRC resolution retains the agreed-upon language where it explicitly references MDG 5, which it limits to “improving maternal health.”

Donna J Harrison, MD, is a diplomat for the American Board of Obstetrics and Gynecology and president of the American Association of Pro Life Obstetricians and Gynecologists. Her paper, "Removing the Roadblocks from Achieving MDG5 by Improving the Data on Maternal Mortality," can be found here.


Join Mercator today for free and get our latest news and analysis

Buck internet censorship and get the news you may not get anywhere else, delivered right to your inbox. It's free and your info is safe with us, we will never share or sell your personal data.

Be the first to comment

Please check your e-mail for a link to activate your account.