Sunday, 4 September 2011

Abortion, Mental Health Problems and the Politicisation of Science

Just the other day, several UK newspapers (here, here) reported about a new, ambitious study on the relationship between abortion and mental health problems. Undertaken by Priscilla Coleman, professor at the Bowling Green State University, and published in the well-renowned journal, British Journal of Psychiatry, the study is a so-called meta-analysis where data and results from many different studies in this area are brought together and analysed as a whole. The result is that evidence speaks in favor of a significant statistical correlation between abortion and elevated risks of mental health problems after abortion. Not surprisingly, this news have been ceased on by political campaigners who want to push for more restrictive abortion legislation. Coleman herself is markedly more modest, concluding only that...

Consistent with the tenets of evidence-based medicine, this information should inform the delivery of abortion services.
In the news reporting, this statement has generally been read as a call for telling abortion seeking women about the risks. Now, if you believe in individual liberty, women's right to decide over their own body or is generally oriented towards pro-choice positions on abortion, this may not appear to be very shocking. After all, abortion is a serious medical or (when undertaken in later stages of pregnancy) surgical procedure that is often undertaken in the context of a personal crisis. If there indeed is a link between undergoing this procedure and suffering mental health problems, this should be explained to women contemplating abortion, shouldn't it? However, there are several question marks surrounding Coleman's result and what conclusions can be inferred from it, both scientifically and with regard to policy.

First, neither the news reporting, nor the abstract to the actual article reveals if and to what extent Coleman's analysis has controlled for relevant so-called confounders – i.e. other factors contributing to the onset of mental health problems. In particular, it is unclear to what extent the analysis has factored in the presence of mental health problems or risk factors for such problems before abortion and, indeed, pregnancy. When I had reason, several years back, to look at research in this area undertaken in Sweden, a repeated phenomenon was that there indeed were correlations between abortion and mental health problems, but that the occurrence of this problems almost always could be linked to the presence of such problems (or risk factors for them) before abortion or even the actual pregnancy. Now, Coleman's results indeed indicate that women who undergo abortion have elevated risks compared to a substantial control group. However, in lack of controlling for the presence of problems before abortion, this may just as well support the notion that women already suffering from mental health problems or who are victims of risk factors for this are more likely to establish pregnancies that they eventually would prefer to have terminated. Indeed, such a link was suggested by a large Danish study published last year in the highly ranked journal New England Journal of Medicine. This, in turn, may depend on a lot of different things, such as being placed in a dysfunctional psycho-social context (e.g. lack of a reliable partner, poverty, disorganised housing situation, et cetera). Hopefully, time will tell what quality Coleman's study actually possess in this respect and, in the case of deficiencies, that further studies are undertaken to investigate the issue.

However, even if it turns out that Coleman's results hold up to closer scrutiny, the sort of factors just mentioned make the interpretation of Coleman's practical recommendation made in the media quite odd, if not biased – just as the one reported about the reaction of many campaigners against liberal abortion legislation. For suppose that Coleman indeed is right in the strongest sense, is the most obvious conclusion then that we should move to restrict access to legal abortion? Actually not, since there is strong scientific support for the claim that such actions lead to no good (in particular, they do not prevent abortions). How about informing about the risk then? Well this looks more sensible, although, if the underlying explanation of abortion being a risk factor is that a certain portion of abortion seeking women are already burdened by mental health problems that threaten to become more serious if they are exposed to trauma of some kind, one may doubt the efficacy of such actions.

Instead, the practical conclusion that would seem to be gaining the most support would be this: Researchers like Coleman should rapidly proceed to develop instruments to identify those at risk, and abortion services should offer these women special post-abortion care and counseling, or even preventive actions before the procedure is undertaken that may serve to decrease the risk. If such an instrument proves difficult to develop, such care and counseling should become a standard ingredient of good clinical abortion practice all across the board. I must say that I find it a bit odd and worrying that Coleman's own practical suggestion does not focus on this. After all, if women's mental health is what you care about, practical implications should focus on actions directed at promoting that aim.


  1. In my opinion an abortion services should offer these women special post-abortion care and counseling before the procedure is undertaken that may serve to decrease the risk.

  2. I agree in general terms that advising about the variety of reactions that may follow an abortion (or any other procedure, including a delivery) is good practice, but note that pre-counseling that mention risks is problematic in the abortion case in view of the controversy on the source of the risk. If abortion itself turns out to be a cause of mental health problems it is, of course, appropriate. However, if it is not, it is much more problematic, since it is then unclear what one is to say in counseling. The statistical association is there, of course, but it is not at all clear that it is good practice to use it as basis for saying that there is a risk in having an abortion. This since if, for instance, the overrepresentation of mental health problems among women that have had abortions is explained by the factor that women with mental health problems are more likely to establish unwanted pregnancies, then the abortion may very well be a mitigating factor – i.e., the risk to mental health would have been even worse had it not been for the abortion.

    Having good mental health caring resources, however, is a good thing, whatever the truth about the possible causal link between abortion and mental health problems may be.