Munk-Olsen et al

Risk of First Time use of Anti-Depressants after First Abortion or Delivery (2018)

Examining the Association of Antidepressant Prescriptions With First Abortion and First Childbirth. Steinberg JR, Laursen TM, Adler NE, Gasse C, Agerbo E, Munk-Olsen T. JAMA Psychiatry. 2018 May 30.

Abstract: IMPORTANCE: The repercussions of abortion for mental health have been used to justify state policies that limit access to abortion in the United States. Much earlier research has relied on self-report of abortion or mental health conditions or on convenience samples. This study uses data that rely on neither.

OBJECTIVE: To examine whether first-trimester first abortion or first childbirth is associated with an increase in women's initiation of a first-time prescription for an antidepressant.

DESIGN, SETTING, AND PARTICIPANTS: This study linked data and identified a cohort of women from Danish population registries who were born in Denmark between January 1, 1980, and December 30, 1994. Overall, 396 397 women were included in this study; of these women, 30 834 had a first-trimester first abortion and 85 592 had a first childbirth.

MAIN OUTCOMES AND MEASURE: First-time antidepressant prescription redemptions were determined and used as indication of an episode of depression or anxiety, and incident rate ratios (IRRs) were calculated comparing women who had an abortion vs women who did not have an abortion and women who had a childbirth vs women who did not have a childbirth.

RESULTS: Of 396 397 women whose data were analyzed, 17 294 (4.4%) had a record of at least 1 first-trimester abortion and no children, 72 052 (18.2%) had at least 1 childbirth and no abortions, 13 540 (3.4%) had at least 1 abortion and 1 childbirth, and 293 511 (74.1%) had neither an abortion nor a childbirth. A total of 59 465 (15.0%) had a record of first antidepressant use. In the basic and fully adjusted models, relative to women who had not had an abortion, women who had a first abortion had a higher risk of first-time antidepressant use. However, the fully adjusted IRRs that compared women who had an abortion with women who did not have an abortion were not statistically different in the year before the abortion (IRR, 1.46; 95% CI, 1.38-1.54) and the year after the abortion (IRR, 1.54; 95% CI, 1.45-1.62) (P = .10) and decreased as time from the abortion increased (1-5 years: IRR, 1.24; 95% CI, 1.19-1.29; >5 years: IRR, 1.12; 95% CI, 1.05-1.18). The fully adjusted IRRs that compared women who gave birth with women who did not give birth were lower in the year before childbirth (IRR, 0.47; 95% CI, 0.43-0.50) compared with the year after childbirth (IRR, 0.93; 95% CI, 0.88-0.98) (P < .001) and increased as time from the childbirth increased (1-5 years: IRR, 1.52; 95% CI, 1.47-1.56; >5 years: IRR, 1.99; 95% CI, 1.91-2.09). Across all women in the sample, the strongest risk factors associated with antidepressant use in the fully adjusted model were having a previous psychiatric contact (IRR, 3.70; 95% CI, 3.62-3.78), having previously obtained an antianxiety medication (IRR, 3.03; 95% CI, 2.99-3.10), and having previously obtained antipsychotic medication (IRR, 1.88; 95% CI, 1.81-1.96).

CONCLUSIONS AND RELEVANCE: Women who have abortions are more likely to use antidepressants compared with women who do not have abortions. However, additional aforementioned findings from this study support the conclusion that increased use of antidepressants is not attributable to having had an abortion but to differences in risk factors for depression. Thus, policies based on the notion that abortion harms women's mental health may be misinformed.

Criticisms

1. The abstract makes clear that the focus of this study is political. As a result, the conclusions are overly broad and exceed the actual analysis and findings support.
2. The study is flawed by the lack of an analysis relative to exposure to miscarriage and other unintended pregnancy losses. Numerous other studies have shown that spontaneous abortion is important bench mark for comparison of outcomes relative to induced abortion.
3. The study is fundamentally flawed by the lack of any analyses regarding the dose effects associated with exposure to multiple abortions . . . and even multiple natural pregnancy losses. Both repeat abortions and miscarriages are associated with higher incidence rates of adverse outcomes. This omission is especially important in regard to longer term analyses, which this study purports to offer by examining first time antidepressant use over five years after a first pregnancy outcome. For example, women who carried a first pregnancy to term, for example, but then had two or three pregnancy losses in the subsequent years, are presented and analyzed as if they had only one pregnancy outcome.
4. Another fundamental flaw is that the study examines only first time antidepressant use. Clearly, an analysis of how long antidepressants were used, once they began to be used, would provide a better understanding of the duration of depressive symptoms before and after a pregnancy outcome. Such an analysis would be especially important to examine the effects of pregnancy (abortion or delivery) one women whose first use of antidepressants was prior to the abortion or delivery. One would hypothesize that if pregnancy outcome has an effect on mental health, than those who already have depression prior to a pregnancy outcome should be analyzed separately to investigate if their symptoms persist for a longer or shorter period, following the pregnancy outcome.
5. The main finding of this paper is that among women without a history of using antidepressants, those who have abortions are over twice as likely to begin using antidepressants in the year following an abortion compared to both women who had no pregnancy and to women who carry a first pregnancy to term. This finding clearly confirms that abortion is a risk factor for higher rates of depression.
6. The rest of the paper is nothing more than an effort to explain away this key finding. The discussion and conclusion of this paper are more focused on speculations to explain away the results than on the results themselves. This demonstrates the clear bias of the authors. Not only do they avoid doing the more in-depth analyses discussed above, but their main finding is essentially dismissed as misleading based on the authors' speculation that the true cause of higher depression rates following an abortion is pre-existing problems that are unrelated to the women's abortions.
7. The authors frequently misstate facts in order to advance false conclusions. For example, they state that among women who have abortions "the increased risk of depression did not change from the year before to the year after an abortion." But their study did not measure depression. It did not measure depression scores among women at multiple times before and after their abortions. They measured only first-time use of an anti-depressant. They do not know if women who first used an anti-depressant prior to abortion got better or worse following their abortions. They do know that if they had done an analysis excluding women with a prior history of depression, those who had abortions were twice as likely to begin using anti-depressants than other women (both never pregnant and those who deliver), but that finding is essentially ignored in favor of the assumption that all women who have abortions are more mentally ill to begin with.
8. The crux of the author's dismissal of their own findings is their argument that "it is possible that mental health problems may lead women to have unintended pregnancies and abortions." In other words, they are claiming that their study proves that women who have abortions are already mentally ill before they have their abortions, which then entirely explains the higher rates of mental illness after abortion. To support this argument, they point to a finding that among women who had abortions the first time use of anti-depressants in the year prior to their abortions was already higher than that for never pregnant women and women who gave birth. But this is a superficial analysis. It assumes that all women who have abortions fit into the same mental health profile. But clearly, first time use of anti-depressants is a very limited measure of mental health. It tells us nothing about the severity, duration, or recurrence of previous or subsequent mental health problems, even of depression. It is limited to "first time antidepressant use," a useful marker but clearly nothing more than a starting point...and the fact that they look no deeper than this starting point, and so quickly dismiss their own findings suggests that a deeper look would undermine their overly broad conclusions, not support them.
9. The study design inappropriately collapses pre- and post- mental health events. The women who had first use of antidepressant prior to their abortion or childbirth should have been analyzed as a separate group to identify if they had an increase or decrease in how long they used antidepressants or had other treatments, such as counseling, in order to measure the relative severity, duration, and recurrence of problems between these two groups. Similarly, women who had a first time use of antidepressants following their first pregnancy outcome should also have been analyzed in a separate comparison to investigate the relative severity, duration, and recurrence of depression and other mental health symptoms.
10. The decline in first-time use beyond several years may be explained by subsequent deliveries (which may heal some portion of the depressive effects) or by the increase in drug or alcohol use following abortion which is well documented. This underscores the weakness of looking at a single variable, such as first time use of antidepressants, in an effort to estimate the overall contribution of abortion to mental health issues.
11. The longer term analysis (beyond one year) of first time antidepressant use, is deeply flawed by the failure to examine the effects of other pregnancy outcomes, subsequent births, abortions, and miscarriages. The whole premise is to investigate the effects of pregnancy outcomes, but women's experiences are not limited to a first pregnancy outcome. Surely, many women who aborted a first pregnancy may have subsequently carried to term, indeed many will have sought out a replacement pregnancy which may have reduced their risk of using antidepressants. Similarly, many women who gave birth may have experienced subsequent miscarriages or abortions, which may have increased their risk of using antidepressants. This demonstrates the superficial nature of these authors' investigation, which was clearly intended to be limited its scope to the point of demonstrating that among the entire population of women having abortions a disproportionate number have a prior history of using antidepressants. That's an interesting fact, but it simply does not prove that abortion doesn't contribute to depression, despite the authors' arguments to that effect.
12. As with prior studies by Munk-Olsen, the author's use one year prior to the pregnancy outcome as a measure which include nine months of pregnancy for women who delivered and just one to three months for women who had abortions. In other record linkage studies, data on pre-pregnancy mental health was examined for one year prior to the estimated date of conception. That same standard should be applied to the analyses of this data set as it would be a better comparison of the women prior to conception.
13. The study was limited to first trimester abortions. Approximately 10% of abortions occur after the first trimester, and it is important to investigate and report on the mental health of later term abortions. The most likely explanation for not including a subset in the analysis for later abortions is that the results for women who aborted were even more negative.
14. There is no breakdown showing the relative risks relative to age groups. Even more notably, the authors excluded women (or at least women's pregnancies) that occurred prior to their 18th birthday. Since there is evidence that younger women have more negative emotional reactions to abortion, this may have been done in order to reduce the number of women in the entire sample who had first antidepressant use after an abortion.
15. The study also excluded women who used antidepressants more than a year prior to their first pregnancy outcome. As previously mentioned, a better design would have been to include women with a history of antidepressant use (any any mental health treatment) into a separate, segregated analysis to investigate whether abortion or childbirth contributed to or reduced the duration of antidepressant use in subsequent years. A count of the number of antidepressants prescriptions filled for each period of time examined, for example, would be a good indicator of whether or not women with prior mental health problems experience more or less depression following abortion or delivery (or miscarriage, which should also have been examined.)
16. Clearly, the question of whether abortion contributes to an increase or decrease in subsequent mental health issues, especially for women with a prior history of mental health issues, is an important one. It is also a question these authors carefully avoid. This is why segregated analyses are necessary for the following groups (a) women with a history of mental illness prior to conception of their first pregnancy, (b) women with no history of mental illness, and (c) all women.
17. In the methods, the authors state that computed a general mental health score using the Charlson Comorbidity Index (CCI) and also compiled three other measures of mental health history based on (a) prior inpatient and outpatient admissions (b) prior use of anypsychotic medications, and prior use of anti-anxiety medications. These same measures could and should have been used to access the subsequent mental health of women in order to identify any change in severity, duration, or new symptoms following pregnancy outcomes. Instead, they limited their use of these measures to only the pre-pregnancy outcome period in order to blame all subsequent mental health problems on pre-pregnancy mental illness.
18. In Table 2, the authors report that across all women (never pregnant, aborting, and delivering) the strongest predictors of first time antidepressant use were a history of psychiatric counseling, a history of taking antianxiety or antipsychotic medications. Notably, they authors did not give a break down of the for each subgroup of women (never pregnant, aborting, and delivering). Instead, we are left to presume that these reported rates applied equally to each subgroup of women, which is highly unlikely. Similarly, the CCI score is reported to show that a higher CCI score is associated with a higher risk of first time use of an antidepressant, but again there is no break down of what these scores were for each group, much less a report of how these CCI scores changed over time following a pregnancy outcome. Again, a prime opportunity was missed (or suppressed) which would have given us a better understanding of the degree of mental illness following abortion or childbirth, especially in women with a prior history of mental illness.
19. A major premise of the authors' analysis is that women neatly fit into two categories: those who abort, those who give birth. More specifically, all women who have abortions are cut from the same cloth. Therefore, a finding of higher first use of antidepressants in the group of women in the year prior to having abortions "proves" that all women having abortions are more prone to depression whether or not they get pregnant and abort. More specifically, the authors are arguing that the elevated rate of first antidepressant use in the first year after an abortion (compared to non-pregnant and delivering women) is entirely explained by pre-existing mental health troubles among women who have abortions. But clearly, this is stretching one small data point into an over generalization. These women are individuals. For example, Ruth and Beth both had abortions. If Ruth first used an antidepressant eight months before her abortion, that tells us nothing about Beth. It certainly doesn't prove that if Beth seeks an antidepressant specifically because she is struggling with unwanted thoughts about her abortion that she is wrong...it is likely she would have sought antidepressants for some other reason because that is the "type of women," like Ruth, that she is. This study is simply too shallow. The analysis would be much more convincing if instead of using a single pre-pregnancy measure (first time use of antidepressants) the authors had used all of their mental health measures, including the CCI score and the annual number of antidepressant prescriptions, to show that while Ruth and Beth had a different start date for their first use of an antidepressant, they had similar CCI scores, for example, and a similar duration for using antidepressants. Even better, the author's should have provided data showing how all of these mental health variables changed over time following each pregnancy outcome (not just first pregnancy outcome, but all subsequent combinations of pregnancy outcomes) in order to help us better judge the duration, complexity, and intensity of subsequent mental health problems associated with pregnancy outcomes.
20. In their discussion, the authors state: "...the risk of depression decreased as more time elapsed after the abortion." In fact, their data only showed a decline in the first time use of antidepressants the farther out one looks from the abortion. But this tells us nothing about the risk of depression over time. If all the women who start using antidepressants following an abortion continue to have depression over two, five or ten years, it is false to assert that "...the risk of depression decreased as more time elapsed after the abortion." In fact, the cumulative number of women who have experienced depression would be always increasing. Only the rate of new cases would be declining, which is quite a different matter. Moreover, it is likely that only a limited percentage of women would be open to taking antidepressants. And others may self medicate with alcohol or drugs. At some point the number of new cases must necessarily decline, but this tells us nothing about he overall incidence rate. That problem could have been easily corrected if the authors had instead used as their outcome variable the number of antidepressants taken per year (or even a measure of dose levels as a proxy for severity). In other words, do women who have a history of abortions take more antidepressant medications per year, on average, (a) compared to women with no history of pregnancy, (b) women who have delivered with no history of abortion, (c) women with a history of miscarriage, and (d) women who have a history of childbirth, abortion, and miscarriage in every combination.
21. The authors claim: "A close look at the data, however, suggests that the higher rates of antidepressant use had less to do with having an abortion than with other risk factors for depression among women who had an abortion." Again, the data that they have selectively chosen to report does reveal that prior psychiatric illness is a predictor for first time use of antidepressants, but it hardly proves that abortion itself does not also contribute to depression. The plain and evident truth is that their own data proves that women who have abortions will begin to use antidepressants at a much higher rate than women who carry to term and compared to women who did not become pregnant. This finding supports the conclusion that abortion may contribute to depression rates. Does it prove that abortion is the sole cause of depression in any specific number of case? No. Are other factors, such as prior psychiatric history, likely to be important? Yes. But it is illogical to argue that just because prior psychiatric history is the strongest predictor of first time antidepressant use that abortion never contributes to the onset, duration, or intensity of depression.
22. The authors also claim: "Taking all of these results together, it is possible that mental health problems may lead women to have unintended pregnancies and abortions, as other research has found," citing two studies, neither of which actually prove that the reason women have unintended pregnancies and abortions is due to pre-existing mental health problems. Again, this is a harsh characterization and overgeneralization of the mental health of women seeking abortions. But even if mental health problems contribute to why some women have abortions, these findings do not even begin to address the critical question of whether or not a mentally ill woman who has an abortion is more likely to improve, get worse, or stay the same, much less whether the abortion itself contributes to any positive or negative trends in mental health.
23. The authors also assert that the protective effects of childbirth do not last, noting that the decline in first use of antidepressants in the year following a first delivery disappears in the span between the second and fifth year and beyond. That's an interesting finding, but to be expected. But also, it underscores another weakness in the study, since their analyses do not include subsequent exposure to abortions or miscarriage. Since their own study shows a doubling of the risk of antidepressant use following an abortion, one would expect women who carried a first pregnancy to term but later had an abortion to have an elevated rate of antidepressant use.
24. In their conclusions, the authors write:"if having an abortion is causally related to depression, one would expect a higher rate of first antidepressant use after the procedure than before the procedure, yet rates of antidepressant use were no higher in the year after having an abortion than in the year before an abortion." Actually, since prior mental health problems are common, a better hypothesis would be to expect the rate, duration, and intensity of depressive symptoms to increase after an abortion. But none of these are reported in the authors' study (even though they had relevant data). Instead, they limited their analyses to first time use of antidepressant, severely limiting the scope of any conclusions that can be made.
25. Additionally they conclude: "if there are lagged effects of abortion, one would expect an increase in the rate of depression over time, but rates of antidepressant use decreased as more time elapsed." Here they seriously misrepresent their own findings. The rate of first time antidepressant use declined. They could have, and should have, reported on the actual rate of antidepressant use (number of doses per year per thousand women exposed to abortion, for example) but they did not.
26. They then write: "Finally, the differences in rates of antidepressant use between women who had an abortion and women who did not have an abortion were substantially reduced when adjusted for earlier mental health conditions, parental mental health conditions, parental educational level, and physical health. This suggests that, compared with women who do not have an abortion, women who have an abortion may be at higher risk of depression after undergoing the procedure because they were at higher risk to begin with." Again, their study was limited to only first time use of antidepressants. But more importantly, even if women having abortions are more prone to experiencing depression due to pre-existing reasons, this does not exclude the possibility, nor justify ignoring the probability, that abortion can trigger, complicate, or prolong mental health issues.
27. Their last word: "Consequently, policies based on the notion that having an abortion harms women’s mental health may be misinformed." Here, the verb "may" is misplaced. The "the notion that having an abortion may harm women's mental health" is actually supported by the very limited data reported in this study. This study verifies that women who have abortions experience higher rates of depression compared to delivering and non-pregnant women. Due to the superficial nature of this study, looking at first time antidepressants alone and lacking any breakdown of subsequent mental health changes among the women who did have pre-existing mental health issues, it sheds no light at all on how much abortion may contribute to the increased risk of mental health problems observed following abortion. Indeed, given the incorrect statements made in the conclusion of this paper, and the poor study design, the assertions made by the authors are clearly misleading or misinformed.

Induced First-Trimester Abortion and Risk of Mental Disorder.

Induced First-Trimester Abortion and Risk of Mental Disorder.   Munk-Olsen, T, Laursen TM, Pedersen CB, Lidegaard Ø, Mortensen PB. N Engl J Med 2011;364:332-9.

Abstract

Background
Concern has been expressed about potential harm to women’s mental health in association with having an induced abortion, but it remains unclear whether induced abortion is associated with an increased risk of subsequent psychiatric problems.

Methods
We conducted a population-based cohort study that involved linking information from the Danish Civil Registration system to the Danish Psychiatric Central Register and the Danish National Register of Patients. The information consisted of data for girls and women with no record of mental disorders during the 1995–2007 period who had a first-trimester induced abortion or a first childbirth during that period. We estimated the rates of first-time psychiatric contact (an inpatient admission or outpatient visit) for any type of mental disorder within the 12 months after the abortion or childbirth as compared with the 9-month period preceding the event.

Results
The incidence rates of first psychiatric contact per 1000 person-years among girls and women who had a first abortion were 14.6 (95% confidence interval [CI], 13.7 to 15.6) before abortion and 15.2 (95% CI, 14.4 to 16.1) after abortion. The corresponding rates among girls and women who had a first childbirth were 3.9 (95% CI, 3.7 to 4.2) before delivery and 6.7 (95% CI, 6.4 to 7.0) post partum. The relative risk of a psychiatric contact did not differ significantly after abortion as compared with before abortion (P = 0.19) but did increase after childbirth as compared with before childbirth (P<0.001).

Conclusions
The finding that the incidence rate of psychiatric contact was similar before and after a first-trimester abortion does not support the hypothesis that there is an increased risk of mental disorders after a first-trimester induced abortion. (Funded by the Susan Thompson Buffett Foundation and the Danish Medical Research Council.)

Key Findings

1. Abortion is not associated with any improvement in mental health.  This finding is consistent with previous studies.
   
2. The year following abortion was associated with a higher rate of treatment for some mental health conditions compared to the nine months prior to the abortion, including 1 to 3 months while pregnant. Specifically, relative risk for psychiatric visits involving neurotic, stress-related, or somatoform disorders was 47% and 37% higher for women post-abortion compared to pre-abortion at 2 and 3 months respectively. In addition, psychiatric contact for personality or behavioral disorders was 56%, 45%, 31%, and 55% higher at 3, 4-6, 7-9, and 10-12 months respectively. 
3. The rate of mental health treatments following a first abortion is significantly higher than the rate of mental health treatments following a first delivery. However, the rate of mental health treatment in the nine months prior to abortion (including up to three months coping with pre-abortion stresses) was also higher than the rate of mental health treatments before and after a live birth.
4. Women who have abortions have higher rates of psychiatric treatment (15.2 per 1000 person years) than women who have not been pregnant (8.2 per 1000 person years) and women who deliver their first pregnancy (6.7 per 1000 person years).
5. The findings of this study indicate that compared to other women, the elevated rates of mental health problems associated with abortion occur both before the abortion (during the nine months preceding the date of the abortion) and after the abortion.
6. Because of differences in study design, and because the study did find elevated rates of psychological problems associated with abortion compared to delivering and non-pregnant women, the findings of this study clearly do not contradict previous studies linking abortion to elevated rates of psychological problems. The study does contribute to the literature, however, in that it provides evidence that some mental health problems associated with abortion may arise from pre-abortion stresses, during the time frame of discovering an unintended pregnancy, facing the pressures and concerns associated with making the abortion decision, and undertaking the abortion itself.
7. Clinicians should be alert to the fact that there are higher rates of mental health treatments are sought by women in the year following a first abortion compared to the year following a first delivery.
8. The researchers in this study conclude that the higher rates of mental health treatments provided to women who have abortions is not causally due to abortion, but rather due to a self-selection bias causing women with a propensity to require higher rates of mental health care to be more likely to have abortions.  (This conclusion is a good bit at odds with the study's design since it actually excluded women with a history of inpatient care prior to their first pregnancy, in other words, it was designed to look at the most mentally healthy set of women).  Even if this speculation were confirmed, it raises the interesting question: Is abortion "attractive" to the mentally unstable as an act of healing or as an act of self-destruction?  The lack of any evidence of benefit and the abundance of evidence of continued or increased mental health problems after an abortion suggest that it is not an act of healing.
   
   
   

Limitations

1. Although the researchers asserted that they were comparing women having a first trimester abortion to women giving birth, the study actually mixed the groups. Delivering women who had a previous abortion were included in both the birth group and abortion group. Similarly, women who gave birth to one or more children and then had a subsequent abortion were also included in both groups. The design necessarily dilutes all effects by mixing aborting women and delivering women. In all similar record linkage studies, women with both abortion and childbirth experiences as a third group, one which can be then be compared to both the group of women having only a history of abortion and those having only a history of childbirth.

   a) Several other other studies have already shown that a history of abortion increases the risk of psychiatric problems during subsequent pregnancies that are delivered. For women with a history of abortion, preparing to give birth to a first live born child can itself trigger unresolved emotional issues regarding a prior abortion.

   b)Munk-Olsen's failure to segregate the delivering women in this study who previously had abortions would necessarily increase the overall average of psychiatric issues attributed to childbirth even though a significant portion of the effects may actually have been due to the prior abortions.

2. The pre-event (abortion or childbirth) measure of mental health was limited to only 9 months.  This period was chosen to include (on average) only the time that a delivering woman was pregnant.  It therefore totally ignores the pre-pregnancy mental health of delivering women. By contrast, for women who had abortions this nine month pre-event window includes six to eight pre-pregnant months plus one to three months during which the woman discovered she was faced with an unintended pregnancy and was deciding on an abortion.
3. The sample included women with outpatient psychiatric care, but excluded women with a prior history of inpatient psychiatric care. That would be fine, but it then in regard to the outcome variable, it treats both first time outpatient and inpatient psychiatric care as identical. That is inconsistent. The results should have been reported to show first time outpatient and first time inpatient results both separately and combined. It would also have been preferable to report results separately for women with no prior history of outpatient psychiatric care and those with various levels of outpatient care . . . since clearly women with a history of 40 or more outpatient treatments are different than women with no history of outpatient treatments, but for the sake of this study they are treated the same with no effort to account for these differences.
4. While it is important to control for prior mental health, it would have been far preferable to use the the full life history of psychiatric treatments for women in both groups, or at least a five year history prior to the pregnancy. Such far more complete data exists in the Danish database. It was available to the researchers.  Why it was not used is unclear.
5. At the very least, the control period should have been based on a period of time estimated to have been before the women in both groups became pregnant, as was done in the Reardon(2003) analysis of California medical records, which Munk-Olsen was seeking to replicate and refute. Furthermore, while the researchers did exclude women who had any prior history of inpatient psychiatric care, they did not control for prior history of outpatient psychiatric care.  In other words, we do not know if there were any differences in the rates of outpatient treatment between the two groups.
6. The study excluded women who died (including death from suicide) prior to the end of the 12 month follow-up. Given numerous studies showing higher rates of suicide and deaths from accidents following abortion, this design decision may skew the results.
7. The study considered only a single psychiatric treatment.  It did not measure or weight repeated treatments, which might be used as a measure of the severity and duration of mental health problems.
8.  The study put women in both categories.  Women who had both an abortion and delivery appeared in both groups.  This may have confounding effects.  To eliminate these confounding effects the authors should have limited the study to first pregnancy outcome, and included miscarriage and other natural losses as a third group.
9. Curiously the researchers did not follow the research approach used in previous Danish Study Hospitalization for Mental Illness among Parents after the Death of a Child which examined mental health over a much longer period than just 12 months. Nor did they use a one year exclusion of women with prior hospitalization as was done in the record based study of low income women in California. This tendency to vary from prior research methods, using arbitrary periods like nine months arouses suspicion that selection criteria were chosen to magnify, or diminish, results in a fashion that best supported the authors' hypothesis.
10. The authors controlled for "presence or absence of a history of mental disorders in the parents" of the pregnant girls or women. No justification was given for this. Without showing the results before and after this control, the effect is unclear.
11. The paper failed to cite the only previous record linkage studies on abortion and mental health, a study of 56,741 low income women in California, all of which found significantly higher rates of treatments for mental health and sleep disorders following abortion. This is very significant since these are the best studies with which to compare results and also the ones which should have informed study design; specifically, Munk-Olsen should have equaled or exceeded the methodological approach previously employed. Instead, she chose an inferior approach, controlling for less time prior to the pregnancy outcome and examining psychiatric outcomes for a shorter period following pregnancy outcomes.
12. In addition to ignoring previous record linkage studies, the author's discussion merely compared their results to previous interview based studies, in what was clearly an attempt to suggest that "all" previous studies suffered from selection bias and that, therefore, their record linkage study was superior to all the previous literature (even though they were actually avoiding the comparison of their results the three record linkage studies described above).
13. Bottom line: the control for prior mental health history in this study is very weak and may even lead to misleading results. Additional problems associated with this choice to limit prior mental health issues to this nine-month window prior to pregnancy outcome are detailed below:
    • For women carrying to term, the nine month window does not include any pre-pregnancy time. Women who are excited about having a baby may be less likely to seek mental health.  This may is why the incidence rate is so low 3.9%.  With only a nine-month pre-event window, the authors are even excluding those women who may have been experiencing anxiety or other problems as they struggled to become pregnant.  So this measure of mental health during a first pregnancy is a poor indicator of mental health prior to a first delivery, much less prior to first pregnancy.
    • Given the lack of more complete controls for mental health history suggested above, we must assume that since both groups had no prior history of outpatient mental health care prior to their first pregnancies, the mental health of both groups was very similar prior to becoming pregnant.  The three fold increase in mental health treatments prior to abortion (14.6 per 1000 yrs) compared to women who gave birth (3.9 per 1000 yrs) would therefore appear to be most likely explained by stress these women faced discovering they were faced with an unplanned pregnancy and, in many cases, the concurrent disruption of relationships with male partners, parents, employers and others.  In addition, abortion women were more likely to involved in unstable and possibly abusive relationships prior to their pregnancies.  It is not surprising, then, that many of these women facing the stress of abortion decision-making sought psychiatric advise at a much more elevated rate than they had in the past.
    • It is a common failure in abortion research to fail to identify the difference between pre-pregnancy mental health and pre-abortion mental health. Proponents of the abortion and mental health (AMH) connection generally support the view that abortion experience includes pre-abortion mental health stressors occur between the impregnation and abortion.  Discovery of the pregnancy, negative interactions with loved ones over the pregnancy, moral struggles, maternal stress, morning sickness, stress related to career and education decisions, and numerous other factors associated with the pregnancy can all contribute to escalating levels of stress.  Indeed, AMH proponents often describe abortion as both a stress reliever (because it resolves some of these pre-abortion stresses) and a stress inducer (because it creates new stresses or calcifies existing stresses).  This study's failure to distinguish between pre-pregnancy mental health history and post-conception/pre-abortion mental health history prevents any application of the findings to the more central question of when, if ever, an abortion to be beneficial.
    • AMH deniers often suggest that higher rates of post-abortion mental health problems, which are consistently found in the literature, are due to higher rates of pre-existing mental health problems unrelated to the abortion or the unintended pregnancy.  This hypothesis has not yet been tested due to the failure to distinguish between pre-pregnancy mental health and pre-abortion mental health.  This study, again, fails to make that distinction, which is especially unfortunate since, being a record based study, the necessary data was available to the researchers. 
      

14. The only confounding variables for which the authors use controls are age and number of pregnancies. Controls for marital status and socioeconomic status are missing, even though such data is generally available in record based studies.
15. All women who had a history of psychiatric inpatient treatment more than 9 months prior to the abortion were excluded from the study. Many studies suggest that these women are at heightened risk for post-abortion mental health problems. Additional research should be done to look only at this subset of women to determine if women with a prior history of significant mental health problems (pre-pregnancy) are likely to have a lower or higher rate of mental health problems after abortion or after childbirth.
16. Women who experience repeat abortions are likewise not considered. Approximately half of all abortions, at least in the United States, are for women with a prior history of abortion. Numerous studies indicate an elevated risk of mental health problems associated with multiple abortions. This study fails to shed any light on this important issue. In fact, by including women with a history of multiple abortions in the comparison group of women at the time of their first pregnancy, it is likely this inflated the psychiatric contact rate of women giving birth. In short, the failure to examine psychiatric contact rates associated with repeat abortions means the conclusions cannot be applied to women having multiple abortions.
17. The limited follow-up to just one year after the pregnancy outcome is also an unfortunate limitation. There is evidence that elevated rates of post-abortion reactions persist for at least four years. Conversely, post-partum reactions tend to occur within the first few months following a delivery and the mental health benefits of childbirth may therefore be underestimated by examining treatment rates within only one year.
18. The study did not examine whether the individual women who had psychiatric treatment prior to the abortion were at higher or lower risk of additional psychiatric treatment after the abortion.  For example, did pre-abortion mental health screening/counseling help to reduce subsequent risk of mental health treatment?  Or was it a predictor of higher rates of subsequent mental health treatment?  This is an important issue not answered by the study.
19. The authors' conclusions are not consistent with the data. Specifically, the authors conclude: "our study shows that the rates of a first-time psychiatric contact before and after a first-trimester induced abortion are similar."  The identification of some "similar" rates, however, does not change the fact that their data also showed several statistically significant higher rates for specific mental disorders following abortion (see table 1). It would have been reasonable for their main conclusion to have been "our study shows higher rates of treatment for psychological illness  Additional study, and better controls, may reveal even more significant differences.  In addition, the authors statement that "This finding does not support the hypothesis that there is an overall increased risk of mental disorders after first-trimester induced abortion," oversimplifies and misrepresents the hypotheses presented by AMH proponents.  Even more importantly, the findings of this study do not contradict even this oversimplified hypothesis.  Instead, the conclusion seems crafted to be a sound-bite that obfuscates rather than clarifies the issues of concern.
20. The study results are slanted by the exclusion of women most likely to have the most severe reactions to abortion, namely (a) women aborting after the first trimester, (b) those with a prior history of psychiatric hospitalization, and (c) women who died from suicide or accidents (elevated risk taking behavior or unrecognized suicide) within the year following their pregnancies.
21. The Danish records are so complete that nearly all of these problems could have been addressed through better analyses. The fact that the researchers did not attempt to replicate the better design found in previous studies, but instead incorporated these methodological problems into their study design suggests that they did so in order to slant their result toward their intended conclusions.
22. The study was funded by a grant from the Susan Thompson Buffet Foundation which is primarily focused on funding population control efforts with funding of Planned Parenthood International and other groups promoting abortion in developing countries.

Data Clarification Request Refused

An email request was made to Munk-Olsen on 6/15/2015

Dear Dr. Munk-Olsen,

I am studying your 2011 paper "Induced First-Trimester Abortion and Risk of Mental Disorder."

In your paper you make two statements:

During the period from 1995 through 2007, a total of 84,620 girls and women had a first-time first-trimester induced abortion. Of these girls and women, 868 (1.0%) had a first psychiatric contact (either inpatient or outpatient psychiatric care) during the 9 months before the abortion, as did 1277 (1.5%) within the 12 months after the abortion.

During the same study period, a total of 280,930 girls and women gave birth to their first live-born child. Of these girls and women, 790 (0.3%) had a first-time psychiatric contact within the 9 months preceding delivery, as did 1916 (0.7%) from 0 through 12 months post partum.

It is my understanding that the two above groups include women who had both an abortion and a live birth.  Would you please complete the following two statements for those who had first an abortion and then a birth and for those who had first a birth and later an abortion:

During the same study period, a total of _________  girls and women had a first-time first trimester abortion of their first pregnancy and subsequently gave birth to their first live-born child.   Of these girls and women, ___ (___%) had a first psychiatric contact (either inpatient or outpatient psychiatric care) during the 9 months before the abortion, as did ____(___%) within the 12 months after the abortion.  Subsequently, of these girls and women, _____ (___%) had a first-time psychiatric contact within the 9 months preceding delivery, as did 1916 (__._%) from 0 through 12 months post partum.

During the same study period, a total of _________  girls and women gave birth to their first live born child and had no history of prior abortion and also had a first-time first trimester abortion of at least one subsequent pregnancy.   Of these girls and women, _____ (___%) had a first-time psychiatric contact within the 9 months preceding delivery, as did 1916 (__._%) from 0 through 12 months post partum.  Subsequently, of these girls and women, ___ (___%) had a first psychiatric contact (either inpatient or outpatient psychiatric care) during the 9 months before the abortion, as did ____(___%) within the 12 months after the abortion. 

Thank you for this clarification.

On 7/18/15 Munk Olsen replied:

Thank you for your email. I am pleased that you find our paper interesting. However, answering your questions would require substantial work from several people here at our center, and for this reason I cannot provide them to you. 

Sincerely,

Trine Munk-Olsen

Note: Most researchers routinely reply to such requests with the additional details. It literally takes only seconds to run a tabulation like that requested. Munk-Olsen's refusal to provide any details was clearly motivated by a fear of providing data that would lend itself to any interpretation of the data beyond what she has decided to reveal.

Critique by David C. Reardon

Given the excellence of this data set, it is unfortunate that the researchers failed to consider the methodology employed in prior record based studies of abortion and mental health (for example, Reardon DC, 2003 and Coleman PK, 2002) and instead made choices which tend to confound rather than clarify the issues at hand.

Perhaps the biggest problem is the failure to examine pre-pregnancy mental health. This was done in prior studies which controlled for psychiatric admissions for at least one year prior to the calculated date of conception (Reardon DC, 2003). Instead, Munk-Olsen chose as a base line for mental health a nine month period prior to pregnancy outcome, a period covering the entire period of pregnancy for delivering women and approximately three months of pregnancy and six months pre-pregnancy for women who aborted.

This poorly explained decision unfortunately introduces elements of an apples versus oranges comparison.

This is problematic for two key problems. First, it totally ignores the pre-pregnancy mental health of delivering women. Women who are excited about having a baby may be less likely to seek mental health. With only a nine-month pre-event window, the authors are even excluding those women who may have been experiencing anxiety or depression if they were struggling to become pregnant.

Secondly, for aborting women, it includes two to three months of a period in the women's lives which is likely highly stressful since these women are, it must be presumed, facing discovery of an unplanned pregnancy and potential conflicts over this with partners, parents, and others. This nine month window therefore not only fails to provide a base line for mental health prior to the subjects' pregnancy but mixes, several months of pre-pregnancy mental health with one or more months of post-pregnancy, pre-abortion stress.

Third, it is a variation from Munk-Olsen's own prior methodology employed in a 2006 study of mental health problems following childbirth in which all women who ever had prior history of inpatient treatment were excluded . . . not just those treated within nine months of abortion or delivery. Why this change, unless it was necessary to shift the results in a fashion that would reduce the difference between treatment rates before and after abortion?

These study design criteria are not only unprecedented in similar studies but simply contrary to the stated objective of controlling for prior mental health. This failure in design is even more puzzling given the fact that the available data set included mental health information for the entire life of the women in the study. Given the availability of all prior mental health treatment dates, it seems self evident that the researchers should have created a scale for exposure to prior mental health treatments prior to the estimated conception date of each woman's first pregnancy covering at least one full year, preferably five years, and perhaps for each subject's entire life.

Given the fact that the authors only exclusion criteria was a history of inpatient psychiatric treatment, we must assume that since the mental health of both groups was relatively similar prior to becoming pregnant. The three fold increase in mental health treatments prior to abortion (14.6 per 1000 yrs) compared to women who gave birth (3.9 per 1000 yrs) would therefore appear to be most likely explained by stress these women faced discovering they were faced with an unplanned pregnancy and, in many cases, the concurrent disruption of relationships with male partners, parents, employers and others.

In addition, aborting women were more likely to involved in unstable and possibly abusive relationships prior to their pregnancies. It is not surprising, then, that many of these women facing the stress of abortion decision-making sought psychiatric advise at a much more elevated rate than they had in the past.

The lack of pre-pregnancy mental health measure, in itself, renders it impossible to draw general conclusions regarding from this study. But there are also numerous additional problems with the study design that further confound the results and interpretation:

• The study excluded women who are most likely to have the most severe reactions to an abortion, namely those who already had a prior history of abortion, those with a prior history of inpatient mental health treatment. It also excluded women who died prior to one year after the pregnancy event, thereby excluding women who committed suicide, even though a record linkage study from neighboring Finland found a six fold higher rate of suicide in the year following abortion compared to the year following childbirth (Gissler M, 1996).

• The study failed to keep women in the two groups separate. Women who had both an abortion and delivery were included in both parts of the analysis groups. This is especially problematic given the evidence women with a prior history of abortion have more stress during and after subsequent pregnancies (Coleman PK, 2002). As a result, by putting women with a history of abortion in the group of delivering women, the "control" group may be adulterated with the very post-abortion effects this study professes to be exploring. The authors should have limited the study to first pregnancy outcome, and should also have included first pregnancies ending in miscarriage, still birth, and other natural losses as a third group.

• The outcome variable used for this study was limited to only a single psychiatric treatment. It did not measure or give any weight to repeated treatments or multiple mental health problems in an effort to evaluate the severity or duration of mental health problems associated with pregnancy outcomes. It should also be noted that the authors' conclusions are not consistent with their data. Specifically, the authors conclude: "our study shows that the rates of a first-time psychiatric contact before and after a first-trimester induced abortion are similar." The fact that many of the psychiatric contact rates during the nine months preceding an abortion were similar to the contact rates in the year following an abortion were statistically insignificant does not alter the fact that, even with all of the problems in the methodology identified above, their data also showed several statistically significant higher rates for specific mental disorders.

Specifically, as seen in table 1, contact rates for neurotic, stress-related or somatoform disorders were higher in every two month period following an abortion, and was statistically significant in two of the six periods. Similarly, for personality or behavioral disorders, the relative risk was higher for four of the six periods and significantly higher for two of the six periods. In other words, when one actually examines the table of reported findings, the conclusion that "in some cases the rate of a first-time psychiatric contact is significantly higher following an abortion compared to the nine months preceding an abortion" is at least equally true, and arguably more accurate.

Finally, I would note that at least one of the authors, Laursen, was also a co-author of an excellent record linkage study from Denmark examining mental illness among parents of deceased children that was also published in the New England Journal of Medicine.(Li J, 2005) That study design examined long term mental health effects, beyond one year, and controlled for exposure to multiple losses, and had none of the selection bias problems identified in this Munk-Olsen study.

This exact study design could have, and should have, been used simply by plugging the date of exposure to induced abortion into the same fields used for date of exposure to the death of a child. By using a proven study design in this way, Munk-Olsen and Laursen could have avoided the impression (in this case, the very profound impression) that this new study design (with a nine month pre-event control period and intermixing of subjects into both the abortion group and delivery group, et cetera) was artificially constructed precisely to minimize the number of statistically significant findings associating abortion with increased mental health treatments. In conclusion, this study does nothing to impute previous record linkage studies showing elevated rates of psychiatric treatments following induced abortion.

First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital

First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital in women with a history of treated mental disorder. Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard O, Mortensen PB. Arch Gen Psychiatry. 2012 Feb;69(2):159-65. doi: 10.1001/archgenpsychiatry.2011.153.

Context Mental health problems are associated with women's reproductive decisions and predict poor mental health outcomes after abortion and childbirth.

Objectives To study whether having a first-trimester induced abortion influenced the risk of psychiatric readmission and compare findings with readmission risk in women with mental disorders giving birth.

Design Survival analyses were performed in a population-based cohort study merging data from the Danish Civil Registration System, the Danish Psychiatric Central Register, and the Danish National Hospital Register from January 1,1994, to December 31, 2007.

Setting Denmark.

Participants: All women born in Denmark between 1962 and 1992 with a record of 1 or more psychiatric admissions at least 9 months before a first-time first-trimester induced abortion or childbirth.

Main Outcome Measure: Readmission at a psychiatric hospital with any type of mental disorder from 9 months before to 12 months after a first-time first-trimester induced abortion or childbirth.

Results: Relative risk (RR) for readmission risk 9 to 0 months before a first-trimester induced abortion was 0.95 (95% CI, 0.73-1.23) compared with the first year after the abortion. This contrasts with a reduced risk of readmission before childbirth (RR, 0.56; 95% CI, 0.42-0.75) compared with the first year post partum. Proximity to previous psychiatric admission in particular predicted rehospitalization risks in both the abortion and the childbirth group.

Conclusions: Risk of readmission is similar before and after first-time first-trimester abortion, contrasting with a marked increased in risk of readmission post partum. We speculate that recent psychiatric episodes may influence women's decisions to have an induced abortion; however, this decision does not appear to influence the illness course in women with a history of treated mental disorders.

:Comments:

• Most of the limitations described above regarding Munk-Olsen's 2011 study also apply to this extended analysis.
• This study was limited to 2,838 women with a history of having inpatient treatment for psychiatric illness at some time prior to the nine months prior to their abortions. In other words, it is limited to women who most likely have the most mental health problems.
• Why Munk-Olsen does not run the same analyses for women with no prior mental health problems is not explained. While it is certainly relevant to look at how women with pre-existing mental health issues fare following abortion or delivery, it is equally important to look at how women with no pre-existing mental heatlh issues fare after their first pregnancy, whether aborted or delivered. The decision to not publish any results relevant to women without a prior history of mental illness is almost certainly due to the fact that her results showed a difference between aborting and delivering women. Otherwise, she would have published evidence that there was no difference in order to bolster her argument that pre-pregnancy mental health alone explains the higher rates of mental illness among aborting women.
• Her main conclusion is that the rate of women being readmitted for psychiatric care after an abortion is similar to the rate during the nine months before an abortion proves that abortion does not change the rate of psychiatric admissions among women. But in fact, she does not measure the number of readmissions each woman faces, but only whether she is admitted once. Nor does she look to see if the admissions are for longer periods or for any other measure of whether their psychiatric illnesses have become worse or better. Instead, she simply speculates that these women are most likely to have chosen abortions because of their pre-existing psychiatric issues.
• Without any attempt to measure the frequency of subsequent treatments beyond a first time readmission, or any other measure of severity, plus with the limitation to only one year follow-up (and no reporting of suicides), her conclusion that abortion "does not appear to influence the illness course in women with a history of treated mental disorders" is totally unsupported by the very limited evidence she choose to share.

Letter to Editor By David Reardon Identifying Methodological Problems and Requesting Additional Analyses

These letters were previously available at the links shown. As of January 2017, it appears that Archives of General Psychiatry have moved the content to a yet to be discovered location.

Correction of Study Design Problems Would Bring More Clarity to Abortion Mental Health Issue. Reardon DC. Arch Gen Psychiatry. February 23, 2012

As one of the authors of three record linkage studies examining abortion and mental health(1)(2)(3), I applaud the quality of data being accessed by the Munk-Olsen team(4)(5). I am disappointed that their review and discussion ignored our studies. But I’m even more disappointed by their study design decisions which tend to obscure rather than clarify the key issues in this field.

For example, because negative reactions to abortion increase with time(6) our study examined psychological treatments rates for four years. Munk-Olsen unnecessarily limited follow-up to one year.

Furthermore, our studies employed a uniform baseline, controlling for any psychiatric admissions prior to conception. By contrast, Munk-Olsen introduces a dissimilar comparison overlaying nine months of pregnancy for delivering women with a nine months mix of pregnancy and no pregnancy for aborting women.

The weakness of this dissimilar comparison is underscored by the finding that recent admission is the strongest predictor of subsequent readmissions(4). This creates a front loading problem as evidence by 14% of those assigned to the abortion group being readmitted within six months, prior to becoming pregnant. Worse, because the study design censored these cases after their first readmission, the design totally conceals any information about whether this “extra fragile” group (readmitted within 180 days) faced additional readmissions after their abortions.

It is my hope Munk-Olsen will ameliorate this effect by reporting via a published letter results of a reanalysis restricted to the 429 "reasonably stabilized" women whose previous discharge was at least one year prior to the 21-month period examined.

Yet another design problem was the decision to include 952 women (13%) into both the abortion and childbirth groups. Since abortion is associated with a higher risk of mental illness during or after subsequent pregnancies,(7)(8)(9)(10) such double assignment of women to both groups would tend to obscure rather than clarify differences.

This excellent data set can and should be put to better use. Future analyses should show segregated results for women divided into five groups: women with any history of hospitalization for mental health, and women with 0, 1-5, 5-10, and >10 outpatient mental health treatments, either lifetime or within 5 years prior to the estimated conception date of their first pregnancy. With these groupings, analyses should examine inpatient or outpatient treatment rates in the year prior to pregnancy and for one through five years following first pregnancy outcome (including birth, abortion, miscarriage, and other losses). This study design would more clearly identify relative risks of mental illness associated with various pregnancy outcomes for groups of women with various levels of predisposition to mental illness.

Finally, the opportunity to use these data to provide some definitive answers to questions relating reproductive experiences to mental health would be enhanced by including on the research team experts on both sides of this controversial research field, such as Priscilla Coleman and David Fergusson on one side and Brenda Major and Nancy Russo on the other. Such collaboration would increase confidence in both the study design and the interpretation of results.

References

(1) Reardon DC, Cougle JR, Rue VM, Shuping MW, Coleman PK, Ney PG. Psychiatric admissions of low income women following abortion and childbirth. Can Med Assoc J. 2003; 168(10):1253-7.

(2) Coleman PK, Reardon DC, Rue VM, Cougle JR.State-funded abortions vs. deliveries: A comparison of outpatient mental health claims over five years. American Journal of Orthopsychiatry, 2002; 72(1):141–52.

(3) Reardon DC, Coleman PK. Relative treatment rates for sleep disorders and sleep disturbances following abortion and childbirth: a prospective record-based study. Sleep. 2006 Jan;29(1):105-6.

(4) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard O, Mortensen PB. First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital in women with a history of treated mental disorder. Arch Gen Psychiatry. 2012 Feb;69(2):159-65.

(5) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard Ø, Mortensen PB. Induced first-trimester abortion and risk of mental disorder. N Engl J Med. 2011 Jan 27;364(4):332-9.

(6) Major B, Cozzarelli C, Cooper ML, Zubek J, Richards C, Wilhite M, Gramzow RH. Psychological responses of women after first-trimester abortion. Arch Gen Psychiatry. 2000 Aug;57(8):777-84.

(7) Coleman PK, Reardon DC, Cougle JR. Substance use among pregnant women in the context of previous reproductive loss and desire for current pregnancy. Br J Health Psychol. 2005 May;10(Pt 2):255-68.

(8) Coleman PK, Reardon DC, Rue VM, Cougle J. A history of induced abortion in relation to substance use during subsequent pregnancies carried to term. Am J Obstet Gynecol. 2002 Dec;187(6):1673-8.

(9) Stotland NL. Abortion: social context, psychodynamic implications. Am J Psychiatry. 1998 Jul;155(7):964-7.

(10) Burke T, Reardon DC. Forbidden Grief. The Unspoken Pain of Abortion. Springfield, IL: Acorn Books; 2002.

Munk Olsen's Response

In her response to Reardon's letter, Munk-Olsen made the following responses

1. She admitted that the limitation to a one year follow-up was arbitrary.
2. She admitted that the use of nine months prior to pregnancy outcome, rather than nine months prior to pregnancy, meant the two groups "may not be directly compatible across this 9 months period." In this case, "compatible" can only be interpreted to be a synonym for "comparable" meaning the two groups are not properly comparable during this time period as Reardon argued. Obviously, the whole premise of the study is that the design should have comparable measures for each group. Psychiatric treatment rates before the conception dates of women in both groups was available, but not used. This was almost certainly not provided precisely because a proper comparison did not advance Munk-Olsen's preferred outcome.
3. In regard to the suggestion that the study should have examined not just first time psychiatric admissions following a birth or abortion among women with prior psychiatric admissions, but instead should have also looked at how many times women were treated following their pregnancy outcomes to determine if abortion might increase the severity or frequency of post-abortion psychiatric episodes, Munk-Olsen rejected this request insisting that "The problem of counting re-admissions is that several psychiatric contacts could be caused by the same psychiatric disease/episode, and that it is difficult to discriminate between several separated events, and several contacts for the same event." If a quick re-analysis of the data had shown that there was no increased frequency of admissions, Munk-Olsen could easily have reported this in her response to disprove Reardon's concern. Instead, she simply defended her dodge, essentially admitting "The evidence may show that their are more frequent re-admissions for psychiatric inpatient treatment for women who have abortions, but since we can't be certain that doesn't mean something else, we choose not to report that data."
4. Regarding the problem of mixing women who had both abortions and births into both groups, Munk-Olsen simply declares that it is a fact that some women have both experiences. True. But that does not justify bad research methodology. It is also a fact that it very easy to split women exposed to different pregnancy experiences into more than two group, which is a common practice that has been used by many other researchers. If she had simply used three groups (1. first pregnancy was aborted, 2. first pregnancy was delivered, and 3. one of first two pregnancies aborted and the other delivered) the effects associated with two pregnancies with different outcomes would not distort the effects associated with a single pregnancy outcome. Instead, of providing data from a new analysis showing that Reardon's concerns, if addressed, do not change her results, she instead insists that additional analyses are not necessary (to show to anyone else, I guess). Indeed, she asserts that controlling for parity (number of children born before women's first abortions) somehow remedies the problem, though she give no explanation how or why this should be so.

Critique by Priscilla Coleman

The study has numerous problematic elements and the results cannot be trusted for the reasons described below.

1) The sample is limited to women who had a first abortion or birth between 1994 and 2007. The oldest women in the population for whom the investigators had data were born in 1962, and many of the oldest women surely experienced their first abortion or birth well before 1994 (since they would have been 32 at the time), yet they are not included in the analyses and no explanation is provided for their exclusion.

2) There is a serious problem with the structuring of the birth and abortion groups that the authors openly acknowledge: out of the total sample of 8,131 women, 952 (nearly 12%) were in both groups! In order to conduct clean comparisons, these women should absolutely have been removed prior to conducting the analyses.

3) The authors only analyzed a very small fraction of all women with histories of mental illness, the most extreme forms requiring in-patient hospitalization. For every woman hospitalized for a mental illness after abortion, there are potentially 1000s of women suffering from disorders who received out-patient services, were never treated, of self-medicated with substances. The sample is very narrowly constructed.

4) The only control variables employed are age, calendar period, time since previous admission, reproductive history, and parental history of mental disorders. There were no controls for variables demonstrated in previous studies to be associated with the choice to abort and with post-abortion mental illness including marital status, education level, religion, income, relationship history variables including abuse, planning of the pregnancy, reasons women chose to undergo abortion, and pressure to abort, among other variables.

5) The authors conducted correlational analyses and inappropriately made inferences of causality, For example, in the first sentence in the conclusion section of the article they state: “In the present study, we found that first-time first-trimester induced abortion does not influence the risk of readmission to psychiatric facilities.” Such a statement is not permitted with the use of variables that cannot be controlled (like abortion status), particularly when so few control variables are incorporated.

6) Follow-up was limited to 12 months after the pregnancies were resolved. Prospective studies have shown that many women experience mental health problems associated with an abortion years after the procedure. By only measuring readmission for one year, women who have delayed responses, sometimes triggered by a later pregnancy, are not included in the analyses. The data are available in the Danish registries and there is no valid reason for cutting off the follow-up period so early.

Related Studies Where Munk-Olsen Continues to Reject Requests for Relevant Analyses

Munk-Olsen has published a substantial number of subsequent studies related to pregnancy and mental health but has neglected to report on effects associated with abortion . . . when it suits her. When requests have been made for her to report on the effects of abortion or to undertake any additional analyses which might shed further light on the issue, she has refused . . . a position which seems to lend credibility to the concern that she is seeking to conceal results which might reveal shortcomings in her published analyses and conclusions.

Perinatal psychiatric episodes: a population-based study on treatment incidence and prevalence.

Perinatal psychiatric episodes: a population-based study on treatment incidence and prevalence. Munk-Olsen T, Maegbaek ML, Johannsen BM, et al. Transl Psychiatry. 2016;6(10):e919. doi:10.1038/tp.2016.190.

This is another study in which Munk-Olsen attempts to blame live birth as the primary cause of emotional problems after delivery but chooses to ignore prior pregnancy losses as a significant risk factor. Again, she has refused requests for reanalysis, including the following submitted to her via PubMed:

This analysis of perinatal psychiatric episodes by Munk-Olsen T, 2016(1) is flawed by the decision to ignore the effects of prior pregnancy losses on mental health during and after subsequent pregnancies. This omission reflects a pattern of selective reporting by the lead author and should be corrected by publication of new analyzes.

Giannandrea SA, 2013 and Gong X, 2013, and others have shown that prior fetal loss increases the risk psychiatric disorders.(2,3) There is even a dose effect. Multiple losses, either from miscarriage or induced abortion, predict elevated rates of postpartum anxiety.(2) History of abortion is associated with higher rates of mental illness(4,5) and a two to tenfold increased risk of alcohol or illegal drug use during subsequent term deliveries.6 Notably, the heightened risk of mental illness following fetal death is confirmed by one of Munk-Olsen’s own studies.(7)

In light of the literature, Munk-Olsen T, 2016's conclusion that it is not possible to “predict which women will become ill postpartum” is false.(1) Prior fetal loss is risk factor. The data should have been analyzed with attention to this known risk factor. Confirming this effect could motivate improved screening to identify women who may benefit from additional care.

Sadly, the failure to address fetal loss history does not appear to be an oversight. Rather, in light of Munk-Olsen’s refusal to provide details regarding the impact of abortion history on outcomes reported in several of her studies,(8–10) there appears to be a pattern of obfuscating the effects of pregnancy loss on mental health while overstating the effects of childbirth.

For example, in her own fetal death study, Munk-Olsen T, 2014 she ignored the precedent of other researchers to report on the effects of both miscarriage and induced abortion.2 Instead, her results show only the effects of miscarriage. Even more oddly, she actually did use induced abortion as a control variable, thereby admitting its relevance,(7) but withheld all statistics showing how abortion effected the results--clearly to deny critics of her previous studies findings that confirm their criticisms.(8,9)

This pattern of obfuscating methodological choices was also evident in Munk-Olsen's only two abortion studies.(4,5) In those studies, there is no separate analyses given for women without a prior history of mental disorders, exactly the opposite of what she does elsewhere(7,11) in order to eliminate the confounding effects of prior mental illness. The resulting blend of healthy and ill women was further muddied by mixing women who had one or more abortions prior to their first deliveries into the “control group” of women giving birth.(4,5) This cross-adulteration prevented a true comparison of the mental health of women who abort their first pregnancies and those who deliver their first pregnancies–a clean and straight forward comparison such as wad done in Reardon DC, 2003.(12) Moreover, when a colleague requested a simple count of the number of women in the sample who had both abortions and deliveries and the percentage of these who had psychiatric contact, Munk-Olsen emailed the dismissive response: “answering your questions would require substantial work from several people here at our center, and for this reason I cannot provide them to you.”(13)

Munk-Olsen’s methodology across studies is also inconsistent. In some studies she has shown treatment rates for psychological conditions prior to pregnancy outcome for one(7) or even two years.(11) But in her abortion studies, she used only a nine-month period, covering the entire time women delivering were pregnant while providing a muddy mix of pre-conception time and pregnant time for those who had abortions. In response to criticism on this point, Munk-Olsen admitted this "may not be directly compatible."(5) But isn't the whole point to find methodological choices that make groups "directly comparable"? The precedent of using a period of time prior to conception for the mental health baseline is clearly preferable.(12)

Despite my concerns regarding selective reporting, it remains my hope that Munk-Olsen will respond to this letter with results, not just excuses, showing the equivalent of Figures 1-31 revised to show treatment rates segregated for (a) women with no history of pregnancy loss, (b) women with history of one miscarriage, (c) women with a history of one abortion, and (d) women with a history of multiple fetal losses.

Based on previous research, we expect that such an analysis will support the recommendation that women with a history of fetal loss may benefit from additional counseling services during and after subsequent pregnancies. It is inexplicable why Munk-Olsen would refuse to comply with this request.

References

(1) Munk-Olsen T, Maegbaek ML, Johannsen BM, Liu X, Howard LM, di Florio A, et al. Perinatal psychiatric episodes: a population-based study on treatment incidence and prevalence. Transl Psychiatry [Internet]. Nature Publishing Group; 2016 Oct 18 [cited 2016 Nov 5];6(10):e919. Available from: http://www.nature.com/doifinder/10.1038/tp.2016.190 PMID: 27754485

(2) Giannandrea SAM, Cerulli C, Anson E, Chaudron LH. Increased risk for postpartum psychiatric disorders among women with past pregnancy loss. J Womens Health (Larchmt) [Internet]. 2013;22(9):760–8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24007380 PMID: 24007380

(3) Gong X, Hao J, Tao F, Zhang J, Wang H, Xu R, et al. Pregnancy loss and anxiety and depression during subsequent pregnancies: data from the C-ABC study. Eur J Obstet Gynecol Reprod Biol [Internet]. Elsevier; 2013 Jan [cited 2016 Nov 8];166(1):30–6. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23146315 PMID: 23146315

(4) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard Ø, Mortensen PB. Induced first-trimester abortion and risk of mental disorder. N Engl J Med. 2011;364(4):332–9. PMID: 22310504

(5) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard O, Mortensen PB. First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital in women with a history of treated mental disorder. Arch Gen Psychiatry [Internet]. 2012 Feb [cited 2015 Jul 16];69(2):159–65. PMID: 22310504

(6) Coleman PK, Reardon DC, Rue VM, Cougle J. A history of induced abortion in relation to substance use during subsequent pregnancies carried to term. Am J Obstet Gynecol. 2002;187(6):1673–8. Coleman PK, 2002

(7) Munk-Olsen T, Bech BH, Vestergaard M, Li J, Olsen J, Laursen TM. Psychiatric disorders following fetal death: a population-based cohort study. BMJ Open. 2014;1–6. < PMID: 24907247>

(8) Reardon DC. Lack of pregnancy loss history mars depression study. Acta Psychiatr Scand. 2012;126(2):155. PMID: 22616564

(9) Reardon DC. Postpartum mental health study flawed by fetal loss omission. Scand J Prim Health Care [Internet]. 2015 Oct 2 [cited 2016 Nov 9];33(4):318–9. PMID: 26683289 Available from: http://www.tandfonline.com/doi/full/10.3109/02813432.2015.1111710

(10) Munk-Olsen T, Laursen TM, Meltzer-Brody S, Mortensen PB, Jones IR. Psychiatric disorders with postpartum onset: possible early manifestations of bipolar affective disorders. 2011;1–7. Available from: http://dx.doi.org/10.1001/archgenpsychiatry.2011.157

(11) Munk-Olsen T, Pedersen HS, Laursen TM, Fenger-Grøn M, Vedsted P, Vestergaard M. Use of primary health care prior to a postpartum psychiatric episode. Scand J Prim Health Care [Internet]. Informa Healthcare Stockholm; 2015 Jun 24 [cited 2015 Aug 1];33(2):127–33. PMID: 22147807 Available from: http://informahealthcare.com/doi/abs/10.3109/02813432.2015.1041832

(12) Reardon DC, Cougle JR, Rue VM, Shuping MW, Coleman PK, Ney PG. Psychiatric admissions of low-income women following abortion and childbirth. Cmaj. 2003;168(10):1253–6. PMID: 12743066

(13) Email: Subject:"Question: NEJM Abortion and Risk of Mental Disorder" Dated 7/8/2015

Bipolar Affective Disorders

'Psychiatric disorders with postpartum onset: possible early manifestations of bipolar affective disorders. Munk-Olsen T, Laursen TM, Meltzer-Brody S, Mortensen PB, Jones I. Arch Gen Psychiatry. 2012 Apr;69(4):428-34. doi: 10.1001/archgenpsychiatry.2011.157. Epub 2011 Dec 5.'

CONTEXT: Childbirth has an important influence on the onset and course of bipolar affective disorder, and it is well established that there may be a delay of many years before receiving a diagnosis of bipolar disorder following an initial episode of psychiatric illness.

OBJECTIVE: To study to what extent psychiatric disorders with postpartum onset are early manifestations of an underlying bipolar affective disorder.

DESIGN: Survival analyses were performed in a register-based cohort study linking information from the Danish Civil Registration System and the Danish Psychiatric Central Register.

PARTICIPANTS: A total of 120,378 women with a first-time psychiatric inpatient or outpatient contact with any type of mental disorder excluding bipolar affective disorder.

MAIN OUTCOME MEASURES: Each woman was followed up individually from the day of discharge, with the outcome of interest being an inpatient or outpatient contact during the follow-up period with a first-time diagnosis of bipolar affective disorder.

RESULTS: A total of 3062 women were readmitted or had an outpatient contact with bipolar affective disorder diagnoses. A postpartum onset of symptoms within 0 to 14 days after delivery predicted subsequent conversion to bipolar disorder (relative risk = 4.26; 95% CI =3.11-5.85). Approximately 14% of women with first-time psychiatric contacts during the first postpartum month converted to a bipolar diagnosis within the 15-year follow-up period compared with 4% of women with a first psychiatric contact not related to childbirth. Postpartum inpatient admissions were also associated with higher conversion rates to bipolar disorder than outpatient contacts (relative risk = 2.16; 95% CI = 1.27-3.66).

CONCLUSIONS: A psychiatric episode in the immediate postpartum period significantly predicted conversion to bipolar affective disorder during the follow-up period. Results indicate that the presentation of mental illness in the early postpartum period is a marker of possible underlying bipolarity.

Request for Additional Analysis (Refused)

BiPolar Disorder Study Neglects Prior Research and Adequate Controls. Reardon DC.

Dear Editor,

The Munk-Olsen team’s study of elevated risk of bipolar affective disorder following psychiatric illness in the first month after a delivery is potentially important.(1) Unfortunately, the study neglected to control for the effects of prior pregnancy outcomes on bipolarity.

This omission is striking given the fact that Munk-Olsen has used the same data to publish two studies on abortion and subsequent psychiatric treatment.(2)(3). She is also familiar with the three similar record linkage studies we have published in regard to a population of 56,751 low income women in California.(4)(5)(6)

One of these latter studies revealed that women with a history of abortion were three times more likely (OR 3.0, 95% CI 1.5-6.0) to be hospitalized for bipolar disorder than women who carried to term during the four years following pregnancy outcome.(4) Our study also found that women who had abortions were 2.6 times more likely to be hospitalized for psychiatric treatment than were women delivered.(4) Similarly, Munk-Olsen has also found higher rates of psychiatric contact for each of the first 12 months following an abortion compared to delivery.(2) All of these facts have a direct bearing on the present study(1) and should have been addressed in the study design.

It is therefore tremendously baffling . . . if not suspicious . . . that the present study(1) did not include additional analyses relative to other pregnancy outcomes: abortion, miscarriage, and other pregnancy losses. Clearly, a history of pregnancy loss may impact the rates of postpartum depression following a live birth.(7)(8)(9) The failure to consider and control for pregnancy loss history is a major methodological weakness in this new study.

Both the American Psychological Task Force on Abortion and Mental Health and the Royal College of Psychiatry have called for more research regarding abortion and mental health. Yet studies such as this one continue to be published without information about the effects of pregnancy loss on the outcome, even when the researchers have access to complete reproductive histories. Whether investigation of these effects is being neglected due to lack of insight, or whether results are being redacted for ideological reasons, is unclear.

Journal editors and peer reviewers should heed the call for more research on associations between abortion and mental health by requesting that every study regarding reproductive outcomes and mental health should include segregated results allowing for comparisons relative to pregnancy outcome: live birth, miscarriage, abortion, and other losses.

It is my hope that Munk-Olsen will correct this oversight in the near future. Reanalysis should include segregated results allowing comparisons between delivering women with no history of pregnancy loss, women with a history of one abortion, women with a history of two or more abortions, and women with a history of one or more miscarriages.

It is my fear, however, that the politicization of abortion has long caused and continues to perpetuate a trend toward distortions in study design which are intended to either conceal or to obfuscate the associations between abortion and negative mental health outcomes.

References

(1) Munk-Olsen T, Laursen TM, Meltzer-Brody S, Mortensen PB, Jones I. Psychiatric disorders with postpartum onset: possible early manifestations of bipolar affective disorders. Arch Gen Psychiatry. 2011 Dec 5. [Epub ahead of print]

(2) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard O, Mortensen PB. First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital in women with a history of treated mental disorder. Arch Gen Psychiatry. 2012 Feb;69(2):159-65.

(3) Munk-Olsen T, Laursen TM, Pedersen CB, Lidegaard Ø, Mortensen PB. Induced first-trimester abortion and risk of mental disorder. N Engl J Med. 2011 Jan 27;364(4):332-9.

(4) Reardon DC, Cougle JR, Rue VM, Shuping MW, Coleman PK, Ney PG. Psychiatric admissions of low income women following abortion and childbirth. Can Med Assoc J. 2003; 168(10):1253-7.

(5) Coleman PK, Reardon DC, Rue VM, Cougle JR.State-funded abortions vs. deliveries: A comparison of outpatient mental health claims over five years. American Journal of Orthopsychiatry, 2002; 72(1):141–52.

(6) Reardon DC, Coleman PK. Relative treatment rates for sleep disorders and sleep disturbances following abortion and childbirth: a prospective record-based study. Sleep. 2006 Jan;29(1):105-6.

(7) Burke T, Reardon DC. Forbidden Grief. The Unspoken Pain of Abortion. Springfield, IL: Acorn Books; 2002.

(8) Stotland NL. Abortion: social context, psychodynamic implications. Am J Psychiatry. 1998 Jul;155(7):964-7.

(9) Giannandrea SA, Cerulli C, Anson E, Chaudron LH. Increased risk for postpartum psychiatric disorders among women with past pregnancy loss. J Womens Health (Larchmt). 2013 Sep;22(9):760-8. doi: 10.1089/jwh.2012.4011.

Oversight of Bipolar Disorder Link to Abortion Demonstrated by Subsequent 2015 Study

Munk-Olsen's failure to account for prior abortion history in this study is underscored by a new study in the Journal of Affective Disorders, in which researchers found that 42.4 percent of the women with bipolar disorders had a history of abortion compared to only 13.5 percent of the control group. (Marengo E, Martino DJ, Igoa A, Scápola M, Fassi G, Baamonde MU, Strejilevich SA. Unplanned pregnancies and reproductive health among women with bipolar disorder. J Affect Disord. 2015 Jun 1;178:201-5.)

---

Bipolar Affective Disorders 2

Risk of Postpartum Relapse in Bipolar Disorder and Postpartum Psychosis: A Systematic Review and Meta-Analysis. Wesseloo R, Kamperman AM, Munk-Olsen T, Pop VJ, Kushner SA, Bergink V. Am J Psychiatry. 2015 Oct 30:appiajp201515010124.

OBJECTIVE: Women with a history of bipolar disorder, postpartum psychosis, or both are at high risk for postpartum relapse. The aim of this meta-analysis was to estimate the risk of postpartum relapse in these three patient groups.

METHOD: A systematic literature search was conducted in all public medical electronic databases, adhering to the PRISMA guidelines. Studies were included if they reported postpartum relapse in patients diagnosed with bipolar disorder and/or a history of postpartum psychosis or mania according to DSM or ICD criteria or the Research Diagnostic Criteria.

RESULTS: Thirty-seven articles describing 5,700 deliveries in 4,023 patients were included in the quantitative analyses. The overall postpartum relapse risk was 35% (95% CI=29, 41). Patients with bipolar disorder were significantly less likely to experience severe episodes postpartum (17%, 95% CI=13, 21) than patients with a history of postpartum psychosis (29%, 95% CI=20, 41). Insufficient information was available to determine relapse rates for patients with bipolar disorder and a history of postpartum episodes. In women with bipolar disorder, postpartum relapse rates were significantly higher among those who were medication free during pregnancy (66%, 95% CI=57, 75) than those who used prophylactic medication (23%, 95% CI=14, 37).

CONCLUSIONS: One-third of women at high risk experience a postpartum relapse. In women with bipolar disorder, continuation of prophylactic medication during pregnancy appears highly protective for maintaining mood stability postpartum. In women with a history of isolated postpartum psychosis, initiation of prophylaxis immediately after delivery offers the opportunity to minimize the risk of relapse while avoiding in utero medication exposure.

---

Anti-Depressants Study

Prevalence of antidepressant use and contacts with psychiatrists and psychologists in pregnant and postpartum women. Acta Psychiatr Scand. 2012 Apr;125(4):318-24. doi: 10.1111/j.1600-0447.2011.01784.x. Epub 2011 Nov 25. Munk-Olsen T, Gasse C, Laursen TM.

Abstract

OBJECTIVE: We aimed to study prevalence of antidepressant drug use from 12 months prior childbirth to 12 months postpartum and to compare the prevalences with those in a group of women of similar age who did not give birth. We additionally studied prevalences of contacts with private practicing psychiatrists and psychologists during a similar time period.

METHOD: Our study population comprised of pregnant women, and their controls were drawn from a 25% sample of the entire Danish population. Information on redeemed prescriptions for antidepressants and referrals to psychiatrists and psychologists was extracted. The outcome measure was period prevalence calculated in 3-month intervals from 12 months before childbirth to 12 months postpartum.

RESULTS: In the 2-year observation period around childbirth, 2733 (3.17%) women had one or more prescriptions for an antidepressant and 935 (1.18%) and 1399 (1.76%) were referred to consultations with a psychiatrist or psychologist, respectively. Women giving birth had a markedly lower use of antidepressants compared to controls, with the largest observed difference during third trimester of pregnancy (0.6% vs. 2.20%).

CONCLUSION: We found that the prevalence of redeemed prescriptions for antidepressants decreased during pregnancy and increased postpartum. Similar patterns were observed for contacts with private practicing psychiatrists and psychologists.

Request for Additional Analysis (Refused) Lack of pregnancy loss history mars depression study. Reardon DC. Acta Psychiatr Scand. 2012 Aug;126(2):155; author reply 155-6. doi: 10.1111/j.1600-0447.2012.01880.x. Epub 2012 May 23.

Dear Editor,

The Munk-Olsen team’s study of antidepressant use and psychological treatments before, during, and after childbirth(1) unfortunately omits any control for the effects of prior pregnancies and any comparison to other pregnancy outcomes.

A history of pregnancy loss (including induced abortion or miscarriage) can be a stressor that may arouse unresolved feelings during and following subsequent pregnancies.(2)(3)(4) This is also evidenced by findings that women with a history of abortion are significantly more likely to self-medicate with drugs or alcohol during a subsequent pregnancy than women without a history of this pregnancy loss.(5)

The Munk-Olsen study would have been much more informative if it had included (a) controls for prior pregnancy outcomes, and (b) parallel analyses showing the treatment rates a year before and after other pregnancy outcomes, including abortion, miscarriage, and other natural losses.

The failure to provide this additional analysis is most striking given the fact that Munk-Olsen has used the same data to publish a much more nuanced comparison of psychiatric treatment rates among women who have abortions and women who carry to terms.

Both the American Psychological Task Force on Abortion and Mental Health and the Royal College of Psychiatry have called for more research regarding abortion and mental health.

In a broad sense, efforts to study the interactions between reproductive health and mental health will never be reliable as long as researchers ignore or suppress analyses which fail to encompass the full reproductive history of women, including both voluntary and involuntary pregnancy losses.

Journal editors and peer reviewers should be alert to this problem and should heed the call for better research by demanding that every study regarding reproductive outcomes and mental health should include segregated results allowing for direct comparison of outcome variables relative to the entire range of pregnancy outcomes: live birth, abortion, miscarriage, and other natural losses. By pressing researchers to address and report on these related pregnancy issues, reviewers and editors will helping to advance more rigorous investigation of all of these issues. Without such requests for more detailed analyses, study designs can easily be fashioned to avoid or minimize the investigation of controversial issues.

It is my hope that Munk-Olsen will address the research imperatives raised herein by expanding the study design presented in the present paper.(1) Reanalysis should include segregated results allowing comparisons between: (a) delivering women with no history of pregnancy loss, (b) delivering women with a history of one abortion, (c) delivering women with a history of two or more abortions, (d) delivering women with a history of one miscarriages, (e) delivering women with a history of two or more miscarriages, and treatment rates for (f) women with no prior pregnancy whose first pregnancy is aborted, and (g) women with no prior live birth who have two or more abortions or other losses.

David C. Reardon, Ph.D.

Elliot Institute

References

(1) Munk-Olsen T, Gasse C, Laursen TM. Prevalence of antidepressant use and contacts with psychiatrists and psychologists in pregnant and postpartum women. Acta Psychiatr Scand. 2011 Nov 25. doi: 10.1111/j.1600-0447.2011.01784.x. [Epub ahead of print]

(2) Reardon DC, Cougle JR, Rue VM, Shuping MW, Coleman PK, Ney PG. Psychiatric admissions of low income women following abortion and childbirth. Can Med Assoc J. 2003; 168(10):1253-7.

(3) Burke T, Reardon DC. Forbidden Grief. The Unspoken Pain of Abortion. Springfield, IL: Acorn Books; 2002.

(4) Stotland NL. Abortion: social context, psychodynamic implications. Am J Psychiatry. 1998 Jul;155(7):964-7.

(5) Coleman PK, Reardon DC, Cougle JR. Substance use among pregnant women in the context of previous reproductive loss and desire for current pregnancy. Br J Health Psychol. 2005 May;10(Pt 2):255-68.

---

Psychiatric disorders following fetal death: a population-based cohort study.

Psychiatric disorders following fetal death: a population-based cohort study. Munk-Olsen T, Bech BH, Vestergaard M, Li J, Olsen J, Laursen TM. BMJ Open. 2014 Jun 6;4(6):e005187. doi: 10.1136/bmjopen-2014-005187.

OBJECTIVES:Women have increased risks of severe mental disorders after childbirth and death of a child, but it remains unclear whether this association also applies to fetal loss and, if so, to which extent. We studied the risk of any inpatient or outpatient psychiatric treatment during the time period from 12 months before to 12 months after fetal death.

PARTICIPANTS: A total of 1,112,831 women born in Denmark from 1960 to 1995 were included. In total, 87,687 cases of fetal death (International Classification of Disease-10 codes for spontaneous abortion or stillbirth) were recorded between 1996 and 2010.

PRIMARY AND SECONDARY OUTCOME MEASURES: The main outcome measures were incidence rate ratios (risk of first psychiatric inpatient or outpatient treatment).

RESULTS: A total of 1379 women had at least one psychiatric episode during follow-up from the year before fetal death to the year after. Within the first few months after the loss, women had an increased risk of psychiatric contact, IRR: 1.51 (95% CI 1.15 to 1.99). In comparison, no increased risk of psychiatric contact was found for the period before fetal death. The risk of experiencing a psychiatric episode was highest for women with a loss occurring after 20 weeks of gestation (12 month probability: 1.95%, 95% CI 1.50 to 2.39).

CONCLUSIONS: Fetal death was associated with a transient increased risk of experiencing a first-time episode of a psychiatric disorder, primarily adjustment disorders. The risk of psychiatric episodes tended to increase with increasing gestational age at the time of the loss.

EDITOR'S NOTES:

1. Once again, even though other researchers have shown that abortion and miscarriage are both associated with higher rates of subsequent psychiatric disorders. See for example: Increased risk for postpartum psychiatric disorders among women with past pregnancy loss. Giannandrea SA, Cerulli C, Anson E, Chaudron LH. J Womens Health (Larchmt). 2013 Sep;22(9):760-8. Munk-Olsen's team chose to use a different definition of pregnancy loss so as to exclude reporting on the effects associated with induced abortion.
2. This analysis actually does indicate that the results were modified by "controlling for induced abortions" but meticulously avoids showing readers how induced abortions impacted the results. This actually involves an amazing dance to avoid disclosure of her findings. Munk-Olsen sees that induced abortion history is a significant factor and needs to be used as an adjustment of the results but she does not want anyone else to see how abortion history is affecting the results. So, not only does she not give results associated with abortion, she simply adds a note at the bottom of the tables indicating that the results were adjusted for numerous factors, including abortion history.
3. This study used a 12 month period before the date of fetal death (not 9 months) and also excluded women with psychiatric contact before the pregnancy outcome "as well as women with records of spontaneous abortions and stillbirths prior to the start of the followup period." In other words, unlike her abortion studies, in this study she limited the study to the healthiest set of women to reduce the number of confounding factors. by contrast, in her abortion study, she introduces as many confounding factors as possible.

Does childbirth cause psychiatric disorders? A population-based study paralleling a natural experiment. Munk-Olsen T, Agerbo E. Epidemiology. 2015;26(1):79-84. Background: Childbirth is associated with increased risk of first-time psychiatric episodes, and an unwanted pregnancy has been suggested as a possible etiologic contributor. To what extent childbirth causes psychiatric episodes and whether a planned pregnancy reduces the risk of postpartum psychiatric episodes has not been established.

Methods: We conducted a cohort study using data derived from Danish population registers, including all women having in vitro fertilization (IVF) treatment and their partners with recorded information in the IVF register covering fertility treatments in Denmark at all public and private treatment sites from January 1994 to December 2005. We compared parents and childless persons to examine whether childbirth is directly associated with onset of first-time psychiatric episodes, with incidence rate ratios (risk of first psychiatric inpatient or outpatient treatment) as the main outcome measures.

Results: The incidence rate for any type of psychiatric disorder 0-90 days postpartum was 11.3 per 1000 person-years (95% confidence interval = 8.2-15.0), and 3.8 (3.4-4.3) among women not giving birth. IVF-treated mothers had an increased risk of a psychiatric episode postpartum, incidence rate ratio = 2.9 (2.0-4.2) compared with the risk of psychiatric episodes in childless women. Risk of psychiatric episodes from 90 days postpartum and onwards was decreased (incidence rate ratio = 0.9 [0.7-1.0]).

Conclusions: Using a study design paralleling a natural experiment, our results showed that childbirth is associated with first-time psychiatric disorders in new mothers, indicating that a planned pregnancy does not reduce risks of or prevent postpartum psychiatric episodes.

NOTES: This was a weak study in many respects since IVF mothers may dissimilar to other delivering women in many ways. They may, for example, have a higher rate of prior abortions or miscarriages which are associated with elevated rates of post-partum depression in subsequent pregnancies. Notably, the authors report that they did control for prior history of abortion but they failed to report the rates of abortion or anything about how abortion affected the results . . . thereby concealing the effects of abortion even though the fact that they controlled for abortion signals that it had a significant effect on post-partum treatment rates.

More precisely, while the study found that treatment rates for depression were higher in the first 90 days following IVF deliveries, compared to the rates for not pregnant women, there was no difference beyond the first 90 days.

In the discussion, the authors conclude that since women seeking IVF treatment want their babies, "our results indicate that pregnancy intention does not reduce the risks of having a severe psychiatric episode requiring treatment at a psychiatric facility after a delivery." But if this is true, delivering an intended pregnancy does not reduce the risk of postpartum depression, then this is evidence that delivering an unintended pregnancy does not increase postpartum depression either since intention is not a predictor of postpartum depression.

---

First Time Psychiatric Hospitalization -- Changing the Yard Stick

It's notable that a 2006 study of postpartum psychological illness done by Munk-Olsen does not match up with the results she reported for women giving birth in her 2011 study published in the New England Journal of Medicine (NEJM).

In the 2006 study she reports first time psychiatric hospitalization rates for women following a first live birth (table 1) that are much, much lower than what she reported in her 2011 abortion/childbirth study (figure 1).

For example, in the second month after delivery, she reports an admissions rate of 2.87 per 1000 person years in the 2006 study, but a just under 12 per 1000 person years in her 2011 NEJM study (compared to about 18 per 1000 person years for women who had abortions).

What can explain this huge difference when she is supposedly reporting the same statistic for the same group of women in two different studies?

Well, in the 2006 study she excludes all women with a history of inpatient psychiatric care prior to their first delivery. That's a very reasonable approach if you want to measure how childbirth may impact emotionally healthy women. It is the same approach that could have and should have been used in the 2011 study.

But no, in the 2011 study she changes her protocol for unexplained reasons. In the 2011 study she suddenly decides to exclude only those women who had inpatient psychiatric care in the nine months prior to the abortion abortion or delivery which she is charting. She didn't choose one year, but nine months. Moreover, she didn't choose nine months or one year prior to the estimated conception date, but instead, prior to the outcome date...that means delivering women were pregnant almost the entire time and aborting women were pregnant for two to three months of that 9-month period which is pivotal for her comparisons.

Clearly, the exact same protocol use in the 2006 study could have been done. But no, in the 2006 study, she simply decides "pregnancies ending in abortion (both elective abortions and miscarriages) were not included in the study." This itself is very odd, since a well known 2003 study by Reardon looked at psychiatric admissions after not only childbirth, abortion, and miscarriage, but also at combinations of these, for women who experienced both live births and voluntary and involuntary pregnancy losses. Normally, researchers try to build on prior research and improve the methodology employed, not water it down to show fewer results.

In any case, Munk-Olsen not only failed to replicate Reardon's protocol which gives a breakdown by exposure to all the various pregnancy outcomes, but she actually chose to water down her own established protocol to avoid reporting on the healthiest women (those without inpatient psychiatric care prior to their first delivery).

There's only one explanation for this. She needed to shift her approach in order to produce results that were more easily interpreted to advance her preconceived goal of blaming the higher rates of mental health problems following abortion on pre-existing mental health problems. Toward this end, she abandoned her 2006 methodology which employed women with the least compromised prior mental health (no prior inpatient psychiatric history) and instead chose to include women with prior psychiatric inpatient care.

By including women with more prior history of severe mental illness, she could blame the differences observed in the 2011 study on prior mental health problems. At the same time, she could avoid reporting on the psychological effects associated with abortion among women with no known history of prior psychiatric inpatient care (the same control group used in Reardon's 2003 study). Her desire to hide the impact of abortion on women with the least history of severe mental illness is underscored by her second 2011 study, published in Archives of General Psychiatry in which she specifically focuses on the women with prior history of mental health problems but fails to show how the same analysis applies to women without a prior history of mental health problems.

Notably, in both of her abortion studies, she also ignored all requests to examine whether or not the women with prior mental health problems had more frequent need for mental health care after their abortions. In other words, she claims that the higher rate of mental health problems following abortion may be entirely explained by higher rates of prior mental health problems among women who abort, but she also refuses to examine if these same women have more frequent or more severe (or less frequent or less severe) mental health issues after their abortions. She simply asserts that abortion has no impact on mental health without measuring the frequency or severity of mental health crises before and after abortions.

Other Munk-Olsen Studies with Better Methodology

Use of primary health care prior to a postpartum psychiatric episode. Munk-Olsen T, Pedersen HS, Laursen TM, Fenger-Grøn M, Vedsted P, Vestergaard M. Scand J Prim Health Care. 2015 Jun;33(2):127-33. doi: 10.3109/02813432.2015.1041832. Epub 2015 Jul 15.A

OBJECTIVE: Childbirth is a strong trigger of psychiatric episodes. Nevertheless, use of primary care before these episodes is not quantified. The aim was to study the use of general practice in Denmark from two years before to one year after childbirth in women who developed postpartum psychiatric disorders.

DESIGN: A matched cohort study was conducted including women who gave birth in the period 1996-2010. Women were divided into four groups: (i) all mothers with postpartum psychiatric episodes 0-3 months after birth, n = 939; 2: All mothers with a postpartum psychiatric episode 3-12 months after birth, n = 1 436; and (iii) two comparison groups of mothers, total n = 6 630 among 320 620 eligible women.

SUBJECTS: Women born in Denmark after 1 January 1960, restricting the cohort to women who gave birth to their first singleton child between 1 January 1996 and 20 October 2010.

MAIN OUTCOME MEASURES: The main outcome measures were consultation rates, consultation rate ratios, and rate differences.

RESULTS: Women who developed a psychiatric episode after childbirth had higher GP consultation rates before, during, and after the pregnancy. Women with a psychiatric episode 0-3 months postpartum had 6.89 (95% CI 6.60; 7.18) mean number of consultations during pregnancy, corresponding to 1.52 (95% CI 1.22; 1.82) more visits than the comparison group.

CONCLUSION: Women with a postpartum psychiatric episode had higher use of GP-based primary health care services years before the childbirth, and in this specific group of patients childbirth itself triggered a marked increase in the number of GP contacts postpartum.

:NOTES:

• In this study Munk-Olsen excluded women with any history of prior outpatient or inpatient psychiatric treatment prior to childbirth. This should also have been done in her abortion studies.
• Her analysis did not examine the effects of prior pregnancy loss on post-partum depression, even though research by Giannandrea (2013) has shown that women with a history of miscarriage and/or induced abortion are at higher risk of postpartum psychiatric problems, including depression, anxiety, and PTSD.
• Her analysis included tabulations of the number of average number of consultations women had with their physicians displayed over a monthly basis for over three years. In response to a query for this type of measurement of frequency of contact for psychiatric care after abortion, she had previously claimed it was too hard to do and might be misleading.
• These criticisms were raised in a letter to the editor which was published. Munk-Olsen declined the opportunity to respond.