The BJP Paper on Abortion

I’ve now finished reading the paper on abortion and mental health that appeared in the latest issue of the British Journal of Psychiatry. The paper is dynamite and I’ll summarise it below.

  1. This is not an individual study but rather a “quantitative synthesis and analysis” of all the research on the subject published between 1995 and 2009. What that means in real language is that the author, Priscilla K Coleman, has looked at the 22 studies in this area during the past 15 years, tried to minimise any bias in the individual studies and presented the “overall” picture that the amalgamated data from 877181 women (163831 of whom had had an abortion) tells us. In the authors own words, this exercise produces “the largest quantitative estimate of mental health risks associated with abortion available in the world literature“. The broad picture reveals that women who have an abortion have an 81% increased risk of mental health problems when compared with women who don’t. This constitutes “a moderate to highly increased risk of mental health problems after abortion“.
  2. Specifically, the pooled odds ratio for mental health issues for women who had had an abortion (compared to those who hadn’t) was 1.81. The 95% Confidence Interval was 1.57–2.09 at P <0.0001. In laymans terms that means that it is very unlikely that the real figure (81% is, coming from a sample, simple the “most likely figure”) is less than 57% (i.e. that women who had an abortion were 57% more likely to have mental health issues). The key point is this – there is PRACTICALLY NO DOUBT that women who have abortions are more likely to have mental health issues, even if the figure is not in reality exactly 81% more likely.
  3. Within this general figure of 81% more likely, there were variations in individual mental health issues. The specific figures were:
    Cannabis Use 230% Increase
    Suicide 155% Increase
    Alcohol Misuse 110% Increase
    Depression 37% Increase
    Anxiety 34% Increase
  4. Comparisons were also made between different subsets of women who had or hadn’t had pregnancies and who did or didn’t take those pregnancies full term. The figures were as follows.
    Abortion v No Abortion 59% Increase
    Abortion v Carried to Term 138% Increase
    Abortion v Carried to Term (unplanned pregnancy) 55% Increase

The paper summarises the findings with the following comment.

Based on data extracted from 22 studies, the results of this meta-analytic review of the abortion and mental health literature indicate quite consistently that abortion is associated with moderate to highly increased risks of psychological problems subsequent to the procedure. The magnitude of effects derived varied based on the comparison group (no abortion, pregnancy delivered, unintended pregnancy delivered) and the type of problem examined (alcohol use/misuse, marijuana use, anxiety, depression, suicidal behaviours). Overall, the results revealed that 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 was shown to be directly attributable to abortion. The strongest effects were observed when women who had had an abortion were compared with women who had carried to term and when the outcomes measured related to substance use and suicidal behaviour.

Why is this all significant? Here is the wording of the 1967 Abortion Act (Amended):

(1) Subject to the provisions of this section, a person shall not be guilty of an offence under the law relating to abortion when a pregnancy is terminated by a registered medical practitioner if two registered medical practitioners are of the opinion, formed in good faith—

(a)that the pregnancy has not exceeded its twenty-fourth week and that the continuance of the pregnancy would involve risk, greater than if the pregnancy were terminated, of injury to the physical or mental health of the pregnant woman or any existing children of her family; or

(b)that the termination is necessary to prevent grave permanent injury to the physical or mental health of the pregnant woman; or

(c)that the continuance of the pregnancy would involve risk to the life of the pregnant woman, greater than if the pregnancy were terminated; or

(d)that there is a substantial risk that if the child were born it would suffer from such physical or mental abnormalities as to be seriously handicapped.]

(2) In determining whether the continuance of a pregnancy would involve such risk of injury to health as is mentioned in paragraph (a) or (b)] of subsection (1) of this section, account may be taken of the pregnant woman’s actual or reasonably foreseeable environment.

The number one reason given for permitting an abortion is injury to the mental health of the pregnant woman (97.7% in 2010). If the evidence proves definitively that, on average, a woman is more likely to suffer injury to her mental health by undertaking an abortion rather than carrying on with the pregnancy, how can it be justified for a doctor to sign the necessary paperwork arguing the opposite?

48 Comments on “The BJP Paper on Abortion

  1. Useful summary, thanks. Does raise the usual correlation vs causation questions, and I wonder if the full paper unpacks the following sentence:

    Overall, the results revealed that 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 was shown to be directly attributable to abortion.

    Does this mean:

    (a) nearly 10% of the "additional" mental health problems in those women who'd had abortions were "shown to be directly attributable to abortion" (leaving 90% not being so shown); or

    (b) nearly 10% of mental health problems in the full cohort of women covered by the research was "directly attributable to abortion"?

    Of course, it still leaves imponderables such as what the effect would have been of denying abortion to those women who did have abortions.

    Overall, I doubt that this is the knock-down argument against the legal status of current abortion practice that your final paragraph suggests, but I'd be interested to see what other reactions there are to this research from all sides of the debate.

    • It's (b).

      The particular breakdown is 9.9% overall, 26.5% of all cannabis related mental health issues, 8.3% for anxiety, 8.5% depression, 10.7% alcohol use and a whopping 34.9% of all suicides (including 20.9% of all people exhibiting suicidal behaviours).

      • Thanks. Still don't see how that addresses the correlation vs causation query. I would expect (for example) suicide among young women to be associated with mental illness and/or a generally "chaotic" lifestyle (including drug and alcohol misuse), which in turn I would expect to be associated with unwanted pregnancy and consequent termination.

        • You need to test correlation vs causation by asking some specific questions about the things you want to rule out / include. You can also isolate particular factors (i.e. qualify and quantify "chaotic lifestyle") and then see if that has any effect on the figures. For example, if you think chaotic lifestyles lead to both abortion and drug misuse you can see if the odds ratio varies between chaotic and non-chaotic groups.

          The meta-analysis includes a number of papers that do these kind of checks and I suspect if anything significant had been reported in these that gave an alternative correlation that would have been reported.

          • Hmmm. I can see that the rate of suicides among women who've had abortions would be higher than among women who haven't, but I'd need a lot of persuading that more than a minority of the "additional" suicides were a consequence of those women having had an abortion, rather than both having their origin in a common cause. Which is why I'm interested in seeing other responses to this research (and even tweeted Ben Goldacre to see if I could pique his interest in it, without success so far…).

            • What will be interesting is whether pro-abortion peeps engage with the actual content of the paper or simply reel out "woman's choice" mantras.

              I suspect that some of the research you're asking to be done has already been provided in the papers covered by this meta-analysis. The common causes could easily have been controlled for and reported if they were identified. Indeed, the meta-analysis itself seems to indicate that such controlling has been done. For example,

              Contemporary research on abortion and mental health has addressed a number of shortcomings of the earlier work by employing comparison groups with controls for third variables. However, there has also been increased emphasis on incorporating nationally representative samples, prospective designs, controls for prior psychiatric history and comprehensive assessments of mental health outcome measures which in some cases included actual medical records. A secondary objective of this review was to calculate population-attributable risk (PAR) percentages using pooled odds ratios derived from the meta-analysis subdivided by outcome measures. These statistics reflect the incidence of a disorder in the exposed sample (e.g. women who have undergone abortion) that is directly due to the exposure (the abortion procedure).

              So yes, some more work could be done, but I think the quick answer is, "This is the increase solely due to having an abortion".

              Which just makes it all the more damning.

  2. Peter – even if one was wholly pro-life surely, logically speaking, a like-for-like comparison would be "women who are pregnant, don't want the baby, but have it anyway" vs "women have abortion"? (an "unplanned" pregnancy is not necessarily an unwanted one) As (not-only) pro-choicers could point out that abortion is an invasive medical procedure and, as such, is bound to have psychological consequences?

    And that's aside from the fact that doctors,psychiatrists and psychologists by definition have to tailor advice to the individual patient. For example (to pick a hypothetical example at random): if MOST people who divorced reported unhappiness minor depression etc etc that would in no way mean that a particular individual patient ought to be advised to stay in a (for example) loveless or abusive marriage. Surely you can concede that there are plenty of individual women who might well be more depressed if they are denied an abortion, and that it is the particular patient in front of them that doctors et all are supposed to be treating? That they may underestimate the negative psychological effects of going ahead with abortion in no way means that their clinical opinion on a *particular* woman is flawed, so I'm not sure how the above is much of a victory for the pro-life side. Statisticians dictating to doctors is silly at best, and psychiatry is not a science in the lab-test sense of the word.

    (none of which negates the moral arguments against abortion, of course.

    • Your second paragraph shows that you refuse to accept the research. What the paper tells us is that if a doctor continues to think that most of the women who come to him asking for an abortion would have better mental health outcomes if they had abortions, that doctor is wrong. Yes, you can always find individual anecdotal cases which are opposite to the trend, but the undeniable fact that this research demonstrates is that a woman having an abortion is almost twice as likely to have mental health issues because of the abortion then a woman who didn't have an abortion.

      If you think that is incorrect then perhaps you could point out to us where to meta-analysis fails?

      • Most women don't have abortions, no? As such, a like-for-like comparison would be, as I said, "women who have abortions" vs "women who don't want to keep the child, but don't have an abortion". None of the above would preclude a doctor believing that a particular woman wanting an abortion and not having it would be more likely to suffer poor mental health (and indeed he could believe this and of course believe that the woman who has the abortion will still suffer psychological consequences not true of a woman who never has to make such decisions). Someone who regarded abortion (as I suspect that many of the pro-choice crowd, albeit perhaps just instinctively, do) as a "necessary evil" would hardly be swayed by the data and analysis above.

        Would you concede that whether x, y or z ON AVERAGE leads to worse mental health problems that in no way means doctors should not recommend it for *particular cases*? Tardive Dyskensia is a probability (or at least high-risk) for anyone who spends a long-enough time on antipsychotics, and indeed the rates of recovery from schizophrenia are better in non-developed countries than the drugged-up west. But none of those facts (i.e. : generalisations) 'prove' that psychiatrists are wrong or incompetent to prescribe drugs for particular, individual patients.

        • As such, a like-for-like comparison would be, as I said, “women who have abortions” vs “women who don’t want to keep the child, but don’t have an abortion”.

          That would be a subgroup of the third comparison above in bullet 4. These women would be 55% more likely to have mental health issues if they had an abortion then if they didn't.

          Someone who regarded abortion (as I suspect that many of the pro-choice crowd, albeit perhaps just instinctively, do) as a “necessary evil” would hardly be swayed by the data and analysis above.

          So you're saying they would accept almost a doubling of mental health issues for those who had abortions as an acceptable risk?

          Your last paragraph is irrelevant. You cannot argue against statistical research of this quality with anecdotes. This meta-analysis "trumps" anecdotes any day of the month.

          • >>>>>That would be a subgroup of the third comparison above in bullet 4.>>>>

            Yes, and how big is this subgroup? Why is the fairest analogy – abortion v unplanned and UNWANTED pregnancies – not explicitly drawn? Given the difference between "unplanned pregnancies" (devout Catholics have lots of these) and "unplanned and UNWANTED pregnancies" surely the study is in error by conflating the two groups?

            >>>>So you’re saying they would accept almost a doubling of mental health issues for those who had abortions as an acceptable risk?

            Of course, as 'doubling' a low number doesn't necessarily make a high number. I have had, due to my caffeine intake, periods of having double the risk of a heart attack; this (unlike the caffeine ;-)) doesn't exactly keep me up nights. People who think that women shouldn't be forced to have babies they don't want are not liable to alter that opinion just because the alternative has risks. I'd say that a number of women would say that "Have a baby you don't want" is not a better experience to choose than "have a medical procedure that doubles your risk of mental health problems" in the manner you seem to think it is. Most people know that heavy drinking increases the risk of depression (etc); this knowledge has hardly turned us into a nation of sensible drinkers.

            And my last paragraph is not 'anectodal' in the pejorative sense; it's relevant information on how psychiatry works. You seem to be saying that the number of abortions represents a failure in the aims and practice of psychiatry (correct me if I'm wrong). I'm saying that the logic you use above could be used to object to all sorts of elements of psychiatry, so psychiatrists could rightly question why abortion-related mental health issues should be singled out in that way.

            • Yes, and how big is this subgroup?

              I suspect given the current abortion regime, very small and possibly too small for significant analysis.

              There is no attempt to hide this group in the larger group – it is simply that no-one has asked the question to the data isn't available. We therefore have to go on the wider group which these very few cases are part of, and that wider group has a pooled odds ratio c. 1.55

              And my last paragraph is not ‘anectodal’ in the pejorative sense; it’s relevant information on how psychiatry works. You seem to be saying that the number of abortions represents a failure in the aims and practice of psychiatry (correct me if I’m wrong).

              I'm saying that the reason cited for permitting abortions in almost 98% of all cases (preventing mental health injury) has now been empirically demonstrated to produce the exact opposite effect then that desired. Having an abortion increases rather than decreases your odds of having mental health issues EVEN when you claim that you're having it to reduce mental health issues.

          • I don't see that Ryan's last paragraph is "anecdotal". The law requires the doctors to assess the balance between the risk to physical or mental health (a) if the pregnancy were terminated, and (b) if the pregnancy were to continue. If, in their view, the latter exceeds the former then a termination is lawful.

            It's perfectly possible for a doctor to say: "One of the risks of abortion is an increase in the risk of mental health problems. However, I still consider the risks, to this patient, of a continuance with the pregnancy against her wishes exceed the risks from a termination."

            This is the kind of risk-assessment that doctors make all the time. For example, at the moment I am taking a nasal spray whose possible side-effects include severe allergic reaction, glaucoma, perforated septum and nose-bleeds. In other words, on average I am at a higher risk of suffering those things as a result of taking this medicine than if I didn't. But my doctor has evidently concluded that that risk is exceeded by the benefits to my health from taking the medicine – or, to put it another way, the risks from not doing so.

            To that extent, all that the law is doing is stating that abortion is to be a medical procedure rather than simply a matter of personal preference. To say that a medical procedure should only go ahead if the risks of doing so are exceeded by the risks of not doing so is almost to state a triviality. What makes abortion unusual is not that doctors are required to assess the risks to their patient, but that two doctors are required to make the assessment.

            But as I've said before: I'm currently suspending judgment on what this research says about causation, because I'm not qualified to assess the evidence myself and am not convinced by your own analysis of it (not saying that out of any disrespect for you – just I'd like, well, a second opinion, doctor… ;-) ).

            • Jon,

              I understand what you're saying in the body of your comment, but given the overwhelming evidence now in place that, on average, having an abortion significantly increases your risk of mental health issues, the example you give should be an isolated occurrence not a regular happening.

              I'm happy to send you the paper if you want to read it yourself. The author make the very clear statement that the increase she is observing is independent of and had been isolated from any other underlying factors. Remember that the BJP is a very highly respected journal and the peer-review prior to publication would have addressed this issue amongst others.

              • Thanks! Taking a look at it now.

                Let's say as rough figures (based on figures quoted in the text) that the annual suicide rate among women who have not had abortions is 6 per 100,000, while the suicide rate among women who have had abortions is 11 per 100,000. (The precise figures don't make much difference to the argument.)

                Around 180,000 women each year have abortions in England & Wales. Those figures would seem to suggest that abortions would be associated with around nine additional suicides each year. That's nine suicides out of 180,000 patients, and compared with 1,500 suicides a year among women in the UK and Ireland (don't have E&W figure to hand).

                Obviously an awful thing to happen – but it does put another perspective on the risks that the doctors are assessing.

                If I've made some glaring error in my understanding then, well, I wouldn't be in the least surprised. But it wouldn't be the first time that one of these "81% increased risk in [x]" statements begins to look less daunting when you look at the absolute numbers involved.

                • In 2008 there were c. 1430 female suicides in the UK (16 yo and older). The 2001 census female (16y+) was 25,194,751, so the suicide rate is 0.0000568%, which comes out at about 10 of the 180k women. The Paired Odds Ratio for suicide is 2.55, so that means that you would reduce suicide by 6 deaths a year if women didn't have abortions.

                  6.7 million abortions in the UK between 1967 and 2007. Add another 600k to bring us up to date so we arrive at 7.3 million. This means we can account c. 250 suicides since 1967 to the mental health effects of having an abortion.

                  Is that good or bad? You tell me…

                  • It's bad, but I don't think it's enough to significantly affect a doctor's assessment under s.1. Let's say the average GP refers 5 women a year for an abortion (40,000 GPs, 200,000 abortions). So over a career of say 35 years that's say 200 women (rounding up for ease of calculation).

                    So if one in every 30,000 abortions results in a suicide that would not otherwise have occurred, then that's one suicide per 150 GP-careers. I suspect there are plenty of other areas of medical practice where GPs would be very grateful for those odds.

                    I don't know what the figures would be for the doctors who make the actual s.1 assessments. Clearly they would be involved in far more decisions over their lifetimes.

  3. Peter

    Thanks for posting this summary, especially your added commentary with a statisticians expertise! I don't have time currently to go into the paper itself but I agree with you that, based on the stats that you quote, 'the paper is dynamite'. Meta-analysis can be quite a powerful tool as long as the original studies have been rigorously done. It seems that this is such a powerful meta-analysis based on the accumulated sample size from the studies, which is huge! Your quoting of confidence intervals is also very useful as it shows that women who have an abortion are at least 57% more likely to suffer from mental health issues at the level that is popularly termed 'proven'. This is powerful evidence indeed.

    As John H rightly points out, this data doesn't immediately demonstrate causality rather than correlation. However, the power of a rigorous meta-analysis is that it enables possible alternative causal factors to be investigated, especially with the lareg sample size of this particular study. And at least some of the potential alternative causal factors have been analysed, at least by proxy.

    The overall higher abortion correlation of risk for anxiety and depression of broadly a third higher is a powerful finding as there is quite a high general lifelong risk for women generally – so a third higher is quite significant irrespective of any other possible alternative causal factors.

    The 55% higher risk for mental health issues on abortion versus unplanned pregnancy carried to term is very powerful as the latter group is that which would be expected to seriously consider an abortion as an alternative. Wearing my economist's hat as a social scientist I would have liked to have seen some differentiation based on class and race (gender is obviously a given in this case!) It would be useful to look at differences in the prevelance of abortion (would the relevant stat be annual numbers per 100,000 women of reproductive age?) by class and race.

    However, these are minor quibbles. Overall, the findings are substantive and provide strong support to the findings of the churches that many women who come for counsel are dealing with unhealed trauma from abortion and the deep, if even subconcious, feeling of a women that an abortion has taken a life.

    Excellent and needed post!

  4. This is not the first study to have shown there to be an increase in psychiatric illness following an abortion . As a GP the mental health clause has become the reason given for termination of pregnancy ,medical causes are extremely rare less then 0.1% . I am fully behind the current attempt to separate abortion counselling from provision since the providers have a vested interest in promoting abortion since it is a revenue source !!. In my area all requests for terminations are channeled through one of the so called charities who do not have to pick up the pieces possibly years later .

  5. Ummm, correct me if I'm wrong here, but there are some fundamental problems with your interpretation of the research.

    i) This paper is looking at the relationship between abortion and possible mental health problems / substance abuse. However, I don't see any mention in the paper of trying to control for previous mental health problems substance abuse, which is a pretty big problem if you are trying to make a causal link between the 2. It is more than conceivable that being prone to these problems could make people engage in more risky behaviour (in fact I can point you to some articles on this), including substance abuse, promiscuity and not using sexual protection, and therefore the link here could just be that people with mental health problems and substance abuse are more prone to unwanted pregnancies (which is just as plausable in my mind.

    ii) There is no ground whatsoever from the meta-analysis here to suggest that having abortions is more likely to lead to unwanted pregnancies. This is because the crucial control group (women who WANT to have an abortion, but are PREVENTED from doing so for the experiment) can't be used for obvious ethical reasons. If in some imaginary world this could be done, and you could show that the mental health problems and substance abuse is reduced compared to the experimental group, then this would go some way to suggest the causal link you are trying to make. However, without it you are stuck in the situation that it is merely a correlation, with way too many factors involved to make a sensible conclusion.

    I'm prefectly willing to be corrected on this, but as I see it you are massively misrepresenting this piece of research.

    • i) And I quote (again)

      Contemporary research on abortion and mental health has addressed a number of shortcomings of the earlier work by employing comparison groups with controls for third variables. However, there has also been increased emphasis on incorporating nationally representative samples, prospective designs, controls for prior psychiatric history and comprehensive assessments of mental health outcome measures which in some cases included actual medical records. A secondary objective of this review was to calculate population-attributable risk (PAR) percentages using pooled odds ratios derived from the meta-analysis subdivided by outcome measures. These statistics reflect the incidence of a disorder in the exposed sample (e.g. women who have undergone abortion) that is directly due to the exposure (the abortion procedure).

      Your assertion is therefore incorrect.

      ii) You cannot argue from an absence of evidence what might or might not happen. As I have quoted above, the meta-analysis involved systematic controls for the very kind of objections you raise. Without it it would never have passed peer review and been published in the BJP as, since you and John have pointed out, these are the first valid criticisms that would have been raised.

      • Thanks Peter – I see that they have attempted to control for prior use, although if you read the paper in more detail I'm sure you'll see a few problems with this. Not all the studies they looked at would have had information about prior mental health problems, and those that do have very very different types of outcome measurements. They also provide very sparse information on how they "controlled for prior mental health problems", so I have to say I'm still not utterly convinced about this.

        Secondly, my main point was that the conclusion you are making about the abortion causing mental health problems above and beyond getting pregnant against your wishes is completely groundless. Yes, I know they have a control group of women that had unintended pregnancies that ended up going through with it. This is an entirely appropriate control for the conclusions they are trying to make, but not for the insinuation that you are trying to push forward. Having a control group of women that had an unwanted pregnancy that they ended up choosing to keep is VERY different from women who want an abortion and are MADE to go through with the pregnancy. I think it would be pretty silly to suggest that forcing women against their will to have a child would somehow reduce mental health problems.

        So to summarise, yes it's vaguely interesting that there are these associations. However, what you are suggesting is that it is going against current legislation. For that to be true you would have to prove the following is false:

        "and that the continuance of the pregnancy would involve risk, greater than if the pregnancy were terminated, of injury to the physical or mental health of the pregnant woman"

        And that my friend you cannot get from this paper. Because to conclude that you would need the control group discussed above, which you do not have.

        • So to summarise, yes it’s vaguely interesting that there are these associations. However, what you are suggesting is that it is going against current legislation. For that to be true you would have to prove the following is false:

          “and that the continuance of the pregnancy would involve risk, greater than if the pregnancy were terminated, of injury to the physical or mental health of the pregnant woman”

          And that my friend you cannot get from this paper.

          Nonsense. It is the very thing that this paper highlights and I suggest that your rejection of that finding has less to do with good science and more to do with supporting abortion.

      • Also, hope this doesn't come across rude, but I am aware of the peer review process, and it would be rather naive to think that just because something is published in a good journal, it is above reproach (even more so for journals like the BJP – impact factor of 5 is nothing to write home about!). I'm shocked on a daily basis at how many papers with blinding flaws are published in very high impact papers (look at the famous autism article published in the lancet for example).

        Also, about the following quotation from yourself:

        "You cannot argue from an absence of evidence what might or might not happen"

        I think you'll find that my point is completely valid – the onus is not on me but on you, who is making the ill-founded hypothesis. However, I can throw your criticism back at you. The rules in the world of science are if you don't have the data to back up your theories (which as pointed out above, you do not), you assume the null hypothesis.

        • I am not making any hypothesis. The paper is presenting meta-analysis which demonstrates conclusively that, on average, women who have an abortion experience higher levels of mental health injury then those who do not. Arguing about sub-populations which have not been analysed is a demand for further research, not a criticism of the populations which have been already analysed.

  6. Come on Peter, John gave another example (outside of mental health) of the fact that doctors weight up benefits and risks for individual patients. Let's say that the GMC publicly endorsed the above study. What's the next logical step? Suggesting that Doctors – despite their years of medical school and specialised training – are no longer "allowed" to give the clinical opinion that being denied abortion might have negative medical consequences for *individual, particular patients*? A quota system, perhaps, where, before psychiatrists can give such opinions, they have to check that not enough other doctors have recommended abortions on mental health grounds? There would be uproar. The efficacy of antidepressants doesn't seem to accord with the vast number of prescriptions; should there be a quota system to prescribe them too? How liable is it that doctors would go along with such attacks on their authority?

    You might say that doctors should take a general philosophy that abortion on mental health grounds should be less wide-spread, but that wouldn't alter their right to make individual judgments. It's certainly possible to offer an argument that, because of the stakes, abortion *does* justify a change in normal medical practice, but that's not really the same thing as using metastudies to suggest that the individual decisions of particular doctors in regard to particular patients are necessarily fundamentally flawed.

    • The next step would be to provide guidance based on good clinical evidence. This meta-analysis says that the best clinical evidence is that abortion increases, not decreases, mental health injury.

      If you are suggesting that all doctors when asked for an abortion make detailed clinical assessments of the actual likelihood of mental injury in the individual woman then you are having a laugh. What happens in reality is that the doctors just sign on demand. Now the evidence suggests that EVERY TIME they do that, on average that action leads to a significant increase in the woman's likelihood of mental health injury. What's there to argue about? Are you honestly suggesting doctors should be free to act against the best clinical evidence?

      • >>>>The next step would be to provide guidance based on good clinical evidence. This meta-analysis says that the best clinical evidence is that abortion increases, not decreases, mental health injury.>>>>

        Does that mean a judgment along the lines of "this patient would, in my clinical opinion, suffer mental health problems if she is denied an abortion" is necessarily unsound?

        >>>>>If you are suggesting that all doctors when asked for an abortion make detailed clinical assessments of the actual likelihood of mental injury in the individual woman then you are having a laugh. >>>

        Detailed? Possibly not. But then most antidepressant prescriptions are given after brief (>>>>>>>> Now the evidence suggests that EVERY TIME they do that, on average that action leads to a significant increase in the woman’s likelihood of mental health injury. What’s there to argue about? Are you honestly suggesting doctors should be free to act against the best clinical evidence?>>>

        Because it's perfectly logical to believe that a patient would have less mental health problems if she CHOSE to have the baby but that *for a particular patient* being forced to have the baby will lead to greater mental health problems than the act of abortion itself. And that's aside from the issue of patients with pre-existing mental health problems: a woman who was depressed or anxious in regard to an unwanted pregnancy (hardly uncommon) would surely not fit into the "no mental health problems prior to abortion, lots afterwards" group. And how long after (on average) the abortion are those groups who abuse cannabis and alcohol? Is the following an unsound clinical opinion? :

        "the patient, after abortion, may initially experience unease or guilt that may put them at risk for substance abuse. In the long term, however, they will in my opinion be less likely to suffer severe mental illness than if they are denied an abortion"

      • Peter, the problem is that patients don't suffer side-effects of medical treatment "on average". No-one suffers 0.5% of a side-effect.

        The question a doctor has to ask themselves is, "Do the health risks to this woman of continuing with the pregnancy outweigh the risks of a termination?" Obviously one of the factors that has to be weighed in that assessment is any medical evidence as to potential mental health consequences of an abortion.

        In assessing that, absolute figures are more useful than percentages such as "81% more likely to cause mental health problems" (which are the sort of figures beloved of newspaper reports about "doubling the risk of cancer"). To take our earlier discussion of suicide as an example, if only one in every 150 GPs would expect to have a patient commit suicide as a consequence of an abortion across their entire career, then I doubt that risk will weigh very heavily in their assessment of a particular patient.

        The way you talk makes it sound as if a doctor approving a woman's abortion is behaving with reckless irresponsibility akin to suggesting she walk across the M1 with her eyes closed. Thinking only in percentage terms can have that effect. So I'd like to see more absolute figures: out of every 100,000 abortions, how many more women would one expect to experience specific side-effects (depression, substance misuse, etc.) as a consequence?

  7. Well firstly, doctors do take in to consideration mental health issues, but you are completely missing the point here. Yes the research suggests that mental health problems are higher in those who have abortions, but there isn't the data to suggest that it is CAUSED BY ABORTION, and more to the point, that it would somehow be decreased by STOPPING THEM HAVING AN ABORTION. The study doesn't address this, and you can't make the claim because there is frankly no data out there to tell you one way or the other.

    If we were to rein in doctors allowing abortions based on clinical evidence, we would need a simple demonstration that mental health problems DECREASE as a result of removing the choice. This study does not provide that on any level, and your suggestion either shows a lack of understanding of experiments and appropriate controls, or a deliberate misrepresentation of the data for your own agenda.

    The only sensible way for this study to guide policy is to note that women that have abortions are more likely (for whatever reason) to suffer from these problems, so there should be after-abortion care put in to place to prevent this. More to the point, it raises interesting questions about what characteristics make it likely that women who have abortions have mental health problems, as this can help us isolate a vulnerable group and provide care as appropriate. But using this data to suggest that we should rein in abortions is absurd.

    I mean leaving scientific grounds for second, do you really and truly think that taking women who very much want to abort, and making them go through with an unwanted pregnancy, and all the physical, emotional and financial challenges that presents, would make them LESS likely to have mental health problems? This is where having common sense, and not being blinded by religious dogma, helps us to understand and interpret data such as this.

    Also, whilst this is an interesting study, and meta-analyses are always useful, you are of course aware that there is a whole field of research on this issue? And I'm guessing that this is the only paper you've read on it? Policy cannot and should not be made based on individual studies with controversial titles. This is why we ask doctors what we should do for issues such as this. You know, the people that are actually qualified to be consulted on the science behind this. And I can tell you that you would be hard pushed to find one that would be led to the same conclusions as you on this one.

  8. Yes the research suggests that mental health problems are higher in those who have abortions, but there isn’t the data to suggest that it is CAUSED BY ABORTION, and more to the point, that it would somehow be decreased by STOPPING THEM HAVING AN ABORTION. The study doesn’t address this, and you can’t make the claim because there is frankly no data out there to tell you one way or the other.

    Gosh, let's do this again shall we. Ready?

    Contemporary research on abortion and mental health has addressed a number of shortcomings of the earlier work by employing comparison groups with controls for third variables. However, there has also been increased emphasis on incorporating nationally representative samples, prospective designs, controls for prior psychiatric history and comprehensive assessments of mental health outcome measures which in some cases included actual medical records. A secondary objective of this review was to calculate population-attributable risk (PAR) percentages using pooled odds ratios derived from the meta-analysis subdivided by outcome measures. These statistics reflect the incidence of a disorder in the exposed sample (e.g. women who have undergone abortion) that is directly due to the exposure (the abortion procedure).

    Which bit of that is the part which says that there is no evidence that it is the abortion that is the key link? Are you simply ignoring the last sentence? "Directly due to the Exposure".

    Or is that you simply don't want the research to be true?

    • OK I'm having one last attempt to explain this, because you are clearly not getting it.

      "Employing groups with controls for third variables" – Yes they did, but they didn't have the one control you actually need to justify your conclusion, i.e. women who want to abort but are not allowed to. Without this, there is no grounds to suggest that reducing abortions with reduce mental health. At all.

      "Controls for previous psychiatric history" – Yes I've already given you that one (although as an actual scientist, I'm concerned that they haven't given me enough information on how they controlled for it to satisfy me with how).

      And frankly they can say that they have shown it is CAUSED BY all they like, but they haven't. And I think you'll find that these things are relatively easy to push past certain reviewers, so don't use the peer review argument. I realise I'm getting exasperated, and in your defense this isn't the easiest concept to understand, and many scientists fall into the trap, so I can understand non-scientists doing the same. Let me try to explain the problem more clearly:

      They are comparing women that naturally fall into one of a number of groups (e.g. women who have planned pregnancies, women that have unplanned pregnancies they decide to keep, women who decide to abort etc.). Now for ethical reasons, we can't do a "proper" experiment and allocate women randomly, as that would be horrid. The result of this is a problem for you (although not necessarily for them, as their conclusions are slightly more relaxed than yours). The problem is that there is a confounding variable (a variable you aren't interested in, which varied with your independent variable, in this case, decision of whether to abort), which is desire to have a baby (or desire to abort). Note "desire" is seperate from "decision", the importance of which will become clear in a second.

      So women who have planned pregnancies very much want to have a baby, then in decreasing order on the scale of desire you have those with unplanned pregnancies but no abortions, and finally those that aborted (who have the lowest desire to have children). So you see, your groups are defined not only by the choice that they made, but how much they want to have a baby or not.

      Now what you seem to be saying is that because mental health issues are higher in the abortion groups to the others, we'd be better off reining it in, because if some women had to go through with the pregnancy they would slot into the other group, and therefore have reduced chance of mental disorders as a result. The crucial difference of course is that unlike people from the other group, they DO NOT want a baby, and DO NOT want to make the decision. Therefore you are not comparing like with like, and could get drastically different results (which I suspect you would). You are saying that it is the abortion that causes mental turmoil. It is quite evident from the above that our groups are defined by i) the presence or absence of abortion, but crucially also ii) the desire to terminate their child (and the multitude of complex variables tied up in that little nugget). Therefore we cannot tell whether it was the abortion procedure itself, or the fact that due to circumstances they wanted to terminate their fetus.

      Just think about it sensibly – the women in the unplanned group that DIDN'T abort actually made the informed decision themselves to go ahead with the pregnancy. The women in the other group dislike the idea of pregnancy so much they chose to abort. It's a basic flaw in experimental design and that is why you cannot make the causal claim.

      One explanation for example could be that it is the fact that they got pregnant against their wishes, and very much didn't want the baby that caused the future problems, not the abortion itself as such. If this were the case, and you then made them go ahead with the pregnancy, the result would be that the mental health problems in this group would be higher again even than when we allowed them to have abortions.

      I will say it one more time. You do not have the data to back up your suggestion, and in continuing with your argument, you are merely showing up your lack of knowledge on experimental design.

      • You are asking for a practically gold standard piece of research for this particular group – that's almost impossible to achieve unless under the current regime you want to force some women to carry their pregnancies and then to track them. Given that that's not going to happen we have to settle for the next best comparison and that's a substantial group of women who carried an unwanted pregnancy.

  9. And as to your last comment, why on earth would I not want to believe it? If their conclusions followed logically from their experimental design, that would be fine – as I pointed out, there are sensible policy decisions we can make if this were true.

    However, the part you are quoting is the authors interpretation of their data, and I'm certainly not going to automatically take their interpretation as being true. You know, scientists often don't read the conclusions section of research articles, or if they do often skim it for the simple reason that one should make their own interpretations of the data presented, not just believe what somebody else says as being true. More to the point, you quoting them at me is not a logical argument, it is in fact a fallacy (the argument from authority). For me the only important thing is the data, and yet again, the data doesn't support your conclusion. I suggest you stick to theology.

  10. P.S. Just read something above which I missed earlier. This proved conclusively? Really? You describe yourself as a statistician, I assume you mean you did an undergraduate in the subject rather than an actual statistician? Because you would be absolutely slaughtered in the scientific community for saying something like that!

    This gives some indication of the link, but not the one that you are making. And there is every reason to criticise the study, because of the confounding variable affecting the group selection as described above.

    As I say, stick to theology :-)

    • I'm pretty convinced it's conclusive. Even at the margins of the confidence intervals it's damning.

      You reject the controls for third variables that the meta-analysis employs (and yes you're right that some papers cited have them and some don't) but you provide no evidence to back up this claim. Your asking for "women who want to abort but are not allowed to" leaves me pondering whether you would ever actually countenance the necessary work needed to actually get those figures (i.e. actually deliberately refusing to do abortions so we would have the outcomes of such a sub-population). If you don't want to countenance such an approach to getting the results then I suggest you drop the issue.

      I made my living before the Vicar Factory from Stats as it happens.

  11. Statistics are very necessary, but all they are doing is backing up what should be blindingly obvious. Women cannot get away scot-free with violating their most basic and primeval instinct to nurture and protect their young.

    Whilst it is true that our present 'culture of death' has brutalised us to some extent, abortion is after all the taking of human life. Even women who say they have not been adversely affected by an abortion should add 'for now'. I would suspect that sooner or later the guilt would catch up with most women. Those who do not care about the taking of human life could, in other circumstances, be described as psychotic!

    • Jill –

      Following your logic, surely a woman who WANTS to have an abortion, and thereby violate "their most basic and primeval instinct to nurture and protect their young.", is in your own terms mentally disordered i.e. it is not the *act* of abortion itself that makes them mentally ill?

      Psychosis has a particular medical meaning, different from the kind of Antisocial Personality Disorder inability to emphathise or care about harm that you seem to be identifying. And surely the size of the pro-life lobby suggests that, even if women *should* feel guilt over abortion, many of them do just regard the foetus as a disposable clump of sells? If the pro-choice lobby conceded that abortion *is* murder then surely there wouldn't be much of a debate left!

  12. Given that you are utterly incapable of understanding my point (which really isn't that much of a complicated one), I will indeed drop the issue and stop wasting my time.

    Given your aptitude in the realm of statistics, I'm starting to get some idea of why you turned to God.

  13. Ah, Ryan, you have hit the nail on the head! This is the very reason the pro-aborts, and those with a vested interest in abortion, are so terrified of independent counselling, and the handing out of very reasonable information on mental illness and increased risk of breast cancer, etc, following abortion.

    Abortion became popular because women were misinformed. I know – I was there at the time! We were told that the foetus was not a human being, that it was merely a 'blob' (A longed-for baby is never a foetus or a blob, it is a baby, as I am sure Peter will tell you. Language is a powerful tool.) There was no question of there being any harm to the mother.

    Having had itching ears (due to the sexual revolution) and soaked up this misinformation, we have filtered this mindset down through several generations, who now cannot see anything wrong with it – in theory! Remember, we once thought slavery was fine, and that black people had no souls.

    The truth, though, has a habit of getting out in the longer term. I think this is already happening. We know much more now about the development of the child in the womb, and about the psychological and physical damage to the mother. The Evan Harrises of this world are fighting a losing battle, against nature and against God.

    • I think overemphasis on experiential arguments for Christian morality are a mistake; lots of people who (e.g.) take drugs and sleep around are happy, whereas "take up your Cross" means exactly that. For example, masturbation-turns-you-mad rhetoric was and is obvious nonsense (which doesn't negate *moral* objections to the act of course), and even the Catechism of the Catholic Church curiously doesn't invoke psychology as an ideological framework that would condemn masturbation (http://www.scborromeo.org/ccc/para/2352.htm)

      Would you agree that many people in the pro-choice side do, in good conscience, believe the foetus to just be a clump of cells? Or do they believe at some level that it is murder and just convince themselves otherwise? Given that many (most?) on the pro-choice side are hardly Christian, is it really justifiable is assuming such guilt? As evidence, look at the use of the morning after pill. Even those with who are (at best) uneasy with abortion would not regard the use of it as perfectly analogous to abortion, whereas presumably, in your terms, it is a human life?

  14. So, let's say that the government decides, on the basis of the BJP paper presented here, to order a moratorium on all abortions beyond 24 weeks, except where the pregnancy presents a *physical* risk to the mother, pending further research.

    It may be shot down as conjecture, but would that really lead to an across-the-board reduction in the incidence of substance abuse and suicides among those women who are affected by the ban? Or would thie incidence of depression largely increase among those who might be forced to go through with a pregnancy against their will?

    Apparently, the findings, it is claimed, show that, on average, pregnant women who carry an unwanted child to full term end up preserving their mental health from likely harm.

    • Apparently, the findings, it is claimed, show that, on average, pregnant women who carry an unwanted child to full term end up preserving their mental health from likely harm.

      Yes, they reduce their mental health risks by a third just by carrying the baby rather than aborting it.

  15. I think that we've now had too much focus on gathering proper inferences from the study alone. The data should now be compared to the other statistic that you mentioned.

    One might concede that doctors may identify specific cases in which the mental health risk is actually attenuated by abortion. In the UK, it is even probable that it will take some time to establish that mental deterioration is caused by abortion, rather than just a concomitant. (Although, that is still very worrying).

    What amazes me, in the light of this study, is the staggering 97.7% of abortions in 2010 that doctors authorised in the UK to prevent the risk of mental health injury as defined in the 1967 Abortion Act.

    Is merely the prospect of increased stress consititute an increased risk to mental health? Where is that defined? How is the type of injury described in section 1(a) distinguished medically from the grave and permanent injury described in section 1(b)?

    What this paper should prompt is a thorough review of the process for assessing mental health risk in the case of abortion. Recent calls for an anti-depressant prescription review were prompted by a much smaller sample size and a less comprehensive study.

  16. One last thought. I read the following about mental injury: 'To amount in law to "nervous shock" (a.k.a. mental injury), the psychiatric damage suffered by the claimant must extend beyond grief or emotional distress to a recognised mental illness' (parenthesis mine).

    So, 97.7% of the abortions last year were approved in order to avert recognised mental illnesses, rather than just emotional distress. Are doctors required to identify the actual illness that might otherwise result?

    If they are not and merely authorise the abortion, I'm sorry, but that fact alone beggars belief in the current system!

Leave a Reply to Peter OuldCancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.