'Lies, damned lies and statistics' from the Alan Guttmacher Institute

One of the principal techniques used by the pro-abortion lobby to advance their agenda of legalising abortion in developing world countries is to argue that ‘safe, legal abortion’ will decrease overall maternal mortality whilst not appreciably increasing the overall number of abortions.

In order to make this case they obviously have to establish first that there are already lots of illegal abortions happening and that many women are dying from them.

To achieve this end lobbyists need statistics about levels of illegal abortions and this where the Alan Guttmacher Institute (AGI) comes in.

For many years AGI’s astronomical figures of illegal abortions from developing countries have gone virtually unchallenged by both prolife and prochoice campaigners alike.

But this is about to change now that more rigorous research is being published.

Jacqueline Harvey this week writes about a new study published in the International Journal of Women’s Health showing that AGI’s figures for illegal abortions in Mexico in 2006 and 2009 were grossly overestimated.

The study titled, ‘Fundamental Discrepancies In Abortion Estimates And Abortion-Related Mortality: A Reevaluation Of Recent Studies In Mexico With Special Reference To The International Classification Of Diseases’ was conducted by a panel of six epidemiologists at four universities in the U.S., Mexico and Chile and examines the actual figures produced by the Federal District of Mexico and confirmed by an independent, non-governmental agency that supports legal abortion.

AGI’s estimate for illegal abortions in Mexico in 2006 was 725,070-1,024,424. But the actual number of abortions in 2007 after abortion was legalised (which typically increases rather than decreases the numbers), was only 10,137! So AGI’s estimate was 70-100 times the actual figure.

After legalisation the AGI estimate for legal abortions in Mexico in 2009 was 122,455. But the actual number was 12,221. This is a 10 fold overestimation.

These gross disparities discredit not only AGI figures for illegal abortions and abortion-related mortality in Mexico, but in all countries where they apply their flawed methodologies to create these bogus estimates.

The researchers also discovered that AGI purposefully includes women who died from ectopic pregnancies, miscarriage and assault in their calculations of illegal abortion-related mortality, a case of intentional deception. This leads them to over-estimate abortion-related mortality rate by almost 35%. I have previously blogged about gross overestimates of maternal deaths in the US and UK abortions before legalisation here.

Harvey concludes:

‘Nonetheless, AGI uses these false calculations and deceptive figures about illegal abortion deaths to push for decriminalization of abortion around the world. This new study authoritatively discredits the Alan Guttmacher Institute and its findings.’

This new study adds hard evidence to suspicions I have had for some time. Last July I was speaking at an ICMDA (International Christian Medical and Dental Association uniting over 70 national bodies of which CMF is one) conference in Nigeria where there were 1,700 Christian doctors and medical students from all over Africa.

A leading obstetrician in Kenya told me then that she thought the AGI stats for death from abortion for Kenya were grossly inflated and based on small urban samples along which included miscarriages and other gynaecological diagnoses.

And another doctor who was working in the main teaching hospital in Lagos, Nigeria had done a research project on abortion deaths and said that the actual number of illegal abortions was very small relative to AGI estimates.

This strategy used in Kenya and other developing countries is similar to that used by US abortion supporters in their efforts to legalize abortion in the late 1960s and early 1970s. Dr Bernard Nathanson, a leading supporter of abortion rights and an abortionist himself, later admitted to deception:

‘We aroused enough sympathy to sell our program of permissive abortion by fabricating the number of illegal abortions done annually in the U.S. The actual figure was approaching 100,000 but the figure we gave to the media repeatedly was 1,000,000. Repeating the big lie often enough convinces the public. The number of women dying from illegal abortions was around 200-250 annually. The figure we constantly fed to the media was 10,000.’

The most impressive catalogue of known abortion statistics on line is that of William Johnston whose latest totals of abortions worldwide (last updated in August 2012) are listed here.

What struck me about these numbers was how much lower they were than AGI figures for all developing countries.

When I raised this discrepancy with Johnston he answered as follows (reprinted with permission):

'(My) figures for worldwide abortions differ because AGI includes estimates of unreported illegal abortions, estimates which are inflated by bad methodology (in my opinion).

My figures cover only reported abortions (with limited use of estimates, eg. interpolation for missing years) thus, while they are incomplete they are well documented. They are also limited to countries with legal abortion and where statistics are compiled. Some of the higher AGI/WHO figures involve estimated underreporting from countries with legal abortions, but most of the difference is from their estimates for developing countries where abortion is illegal or legal under very limited circumstances.

These latter estimates are generally based on hospitalisation samples, household surveys, and a variety of assumptions. This process yields illegal abortion rates that are as high as legal abortion rates in the developed world, coincidentally supporting the AGI thesis that abortion should be unrestricted everywhere because laws have no effect on occurrence rates.

The key here is of course the set of assumptions that turn small sample sizes into multi-national estimates of abortion rates. Some obvious issues I see include: surveys of urban populations on abortion, and treating results as applicable to the general population; bias by basing results on surveys of people willing to talk to these survey takers; the validity of the assumptions used for underreporting, for deciding what fraction of hospital miscarriage cases are illegal abortions, or for turning such “detected" abortions into figures including "undetected" abortions.

I do not dispute that many illegal abortions take place in developing countries but I suspect that the actual numbers are significantly below the AGI/WHO estimates, because the methodology of their estimates involves assumptions biased by their policy position. I have little evidence to produce an estimate of total worldwide abortions, but I'm inclined to suspect that the AGI/WHO figures (of 42 million per year) are high by about a factor of two.

A few years ago Laura Antkowiak and Randall O'Bannon analyzed the AGI methodology in an article series in the National Right to Life newsletter. They indicate, for example, that some of the sample sizes involved are only a few dozen. Here are links to their articles.

1. WHO Claims of Unsafe Abortions and Deaths
2. World Abortion Estimates: An Audit (Part 1)
3. World Abortion Estimates: An Audit (Part 2)
4. World Abortion Estimates: An Audit (Part 3)
5. World Abortion Estimates: An Audit (Part 4)

Here are figures for comparison: AGI/WHO estimate worldwide abortions at 45.6 million in 1995, 41.6 million in 2003, and 43.8 million in 2008. (This is from the jointly AGI-WHO-authored article Sedgh et al., 2012, The Lancet, 379(9816):625+) For those three years what I can document are 18.1, 15.1, and 16.0 million. The drop to my current figure of ~12 million/year is mostly due to fluctuations in reported figures from China.

Some perspective on using AGI as a source: for current abortions in the US, their data is better than official data because the abortion providers provide statistics to AGI that they withhold from state health departments. In contrast, AGI survey-based statistics (based on estimates) tend to be biased.

Another point: the above Lancet article claims: ‘The abortion rate was lower in sub-regions where more women live under liberal abortion laws’ - a counter-intuitive claim to anyone but an abortion proponent, one that rests entirely on methodological assumptions, and one that is refuted by regional-level data in the US and Europe.'

Harvey and Johnston’s work needs much wider circulation to counter the 'lies, damned lies and statistics' that pro-abortion campaigners and population control advocates are using to advance their case. More research is also needed.

Please spread the word.

The fact that abortions in developing countries have been overestimated does not in any way of course alter the fact that abortion remains the number one cause of human death worldwide.

Even when one takes Johnston’s ‘revised-down’ figures the total number of abortions is utterly staggering. Johnston has documented almost 1 billion abortions worldwide from figures gleaned for the 90 years between 1922 and 2012, a figure equivalent to one seventh of the world’s current population.

Given the timespan the vast majority of these babies, had they not been aborted, would still be alive today.