Lies, damn lies, and statisticsPosted: January 18, 2012 at 11:17 pm | Tags: abortion, body, Feminism, science, study
I’ve been following the MTR debate with some interest. I had planned to write a blog post about how she’s not my kind of feminist, and I may yet do, but a statistic she quoted today in an article with Mamamia caught my eye.
6. How do you resolve the apparent divide between being pro-life and a feminist?
A growing number of feminists are questioning abortion as safe, simple and risk free. Research is also indicating that women have significant negative mental health outcomes after abortions. The UK Royal College of Psychiatrists has published a meta-analysis in the British Journal of Psychiatry finding that women who undergo abortions are 81% more likely to experience subsequent mental health problems. (Substance abuse increased 340%, suicidal behaviour by 155%).
I looked at those statistics and boggled, because when I last looked at Wikipedia regarding mental health and abortion the information suggested that there was no correlation between negative health outcomes and abortion. I went and tracked down what I could find of the British Journal of Psychiatry article. Sadly I found it was behind a paywall, so I went and looked at what other people had said regarding the article, the methods used, and the author of the piece. It was an interesting read. To start off, I’ll quote the Results section of the abstract:
Women who had undergone an abortion experienced an 81% increased risk of mental health problems, and nearly 10% of the incidence of mental health problems* (my own asterisk) was shown to be attributable to abortion. The strongest subgroup estimates of increased risk occurred when abortion was compared with term pregnancy and when the outcomes pertained to substance use and suicidal behaviour.
The first commentary I found was an interview with the author, Dr Priscilla Coleman, and Elizabeth Moncher from the Culture of Life Foundation. The questions asked by the interviewer were quite leading and it was hard to tell what Coleman’s actual opinions regarding the study, whether she had approached the study with one opinion and came out with a different finding, or whether she had always believed that there were negative mental health outcomes for women who had had abortions. I won’t quote any of the interview here because it’s likely to make you go “grrrr”. Though in summary, those who are likely to have negative mental health outcomes are those who are from religious backgrounds, those who felt that they had to have an abortion, and who do not have support in their family or community and as a result suffer in silence.
So I dug a bit further and found a blog called, “Philosopical Comment” written by a Professor of Practical Philosophy at the University of Gothenburg, Sweden (Christian Munthe). He has written several pieces on this study (all detailed in this post) and had written a letter to the British Journal of Psychiatry, the letters are freely available on the interwebs (Yay!). Munthe did not, as some authors have, critique the methodology behind the meta-analysis, but he argued against the conclusions, which are stated in the abstract as:
This review offers the largest quantitative estimate of mental health risks associated with abortion available in the world literature. Calling into question the conclusions from traditional reviews, the results revealed a moderate to highly increased risk of mental health problems after abortion. Consistent with the tenets of evidence-based medicine, this information should inform the delivery of abortion services.
Munth argues that there are mental health issues with many medical procedures and life events, and perhaps the focus should not be on restricting access to said medical procedures but instead focusing on providing mental health services to people:
The debate on Coleman’s study has moved into ethics, as commentators claim it to support conclusions about the organisation of abortion services (AS). In her reply, Coleman sides with these voices. However, closer scrutiny puts these suggestion into question.
Puccetti et.al argue that, regadless of Coleman’s study, AS should include “in-depth analysis of the various factors known to interfere with the psychological outcomes” as a pre-procedure. The basis for this is that abortion has not been shown to benefit mental health. No ethical principle is presented to back up this argument, but one may generalize the suggestion into one: If a health care practice has not been proven to be beneficial to mental health, it should not be allowed unless preceded by an in depth analysis of the various factors known to interfere with the psychological outcomes. This principle has far-reaching implications, since risk factors for mental health problems all across health care, while there is normally no proven mental health benefits. The principle also rules out the original suggestion: the suggested analysis should not be made before another analysis has been made and the same holds with respect to that additional analysis, and so on.
These proposals and Coleman’s own response rest on the notion of the most important task being to discourage women from having abortions. If the ethical baseline is to be care for mental health, however, the proper conclusion should rather be about access to mental health care. If Coleman is wrong, anyone may be burdened by mental health problems independently of whetheher [sic] or not they contemplate abortion. If she is right, there should be a special readiness to attend to mental health needs in AS. If feasible, harmless pre-abortion identification of high-risk cases may be added to this. Neither Coleman nor her supporters have considered anything in this vein and this adds to suspiscions [sic] of an undisclosed ideological conflicts of interest. (citations at link)
The letters page at the British Journal of Psychology also contains strong criticisms from Epidemiologists, Professors of Medicine, and others. I’ll quote Gregory L Kinney and Ben Goldacre, the rest can be viewed at your own pleasure/interest. Kinney details that the meta-analysis is flawed as:
An author should be commended for defending the work that they have done when that work is questioned by peers. Coleman however does not address a specific analytical observation made by authors that have commented (Coyne, Polis, Littel and Thygesen among others). Coleman included multiple analyses performed within multiple studies that represent the same study populations (Table 1, citation 10, 20, 21, 22, 24, 27, 29, 30, 32, 39 where each of these citations represents a study population being counted more than once in the meta-analysis). Doing so inappropriately increases the sample size and reduces the variance in the pooled measure. This causes the pooled estimate to be biased as the analytical technique was not applied in an appropriate manner and represents an inflated N. In her response the author addresses that observation by discussing random verses fixed effects and weighting but this is not the issue at hand
Kinney also pointed out that the Coleman included much of her own research in the study (11 of the 22 articles).
Acre wrote that there were significant omissions in Coleman’s paper:
Professor Coleman’s systematic review and meta-analysis of the literature on termination of pregnancy and mental health featured several significant omissions: an undislosed [sic] conflict of interest; no assessment of publication bias; and no assessment of the quality of studies included. The search strategy was also inadequately reported, and the meta-analytic technique was faulty.
Most of the comments in the letters page request that the article be retracted due to flawed methodology, bias, and undisclosed conflict of interest.
Based on my research tonight, I am not satisfied that this paper genuinely reflects the true picture of mental health outcomes for women who have abortions. I’d want to see another meta-analysis done by another group before accepting that this results is true. And even if it is true, as argued by Munth, surely we should be prioritising mental health care for people and not taking away medical procedures and choice from them.
* If “nearly 10% of the incidence of mental health problems” stem from abortion, then over 90% of mental health problems stem from other causes. I don’t know, and with Wikipedia shut down for the SOPA protest can’t find out, how many mental health problems are diagnosed each year. If approximately 9% of mental health problems are caused by abortion, and 91% are caused by everything else that happens to people, is that really a good argument to call for abortion to be restricted or banned? Surely a better argument would be to ban all the other things that are causing mental health problems.
Further reading: More on Review Claiming Abortion Hurts Women’s Mental Health, James C. Coyne PhD