Rapid Responses to:

PAPERS: Sarah Schmiege and Nancy Felipe Russo Depression and unwanted first pregnancy: longitudinal cohort study BMJ 2005; 331: 1303 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

The Reality of Post Abortion Trauma

Margaret RJ Cuthill   (28 October 2005)

Biased sample

Fiona K Pinto   (29 October 2005)

Not very significant

Patrick M R Leahy   (31 October 2005)

Re: Not very significant

Heather R Caserta   (1 November 2005)

Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations

David Reardon   (1 November 2005)

Ambivalence and regrets are not the same as Depression

Joan A. Lang Lang   (3 November 2005)

Post Abortion Risk Assessment

Cynthia M. Dudek   (4 November 2005)

The prevalence picture—Could today's water level of a lake tell how much it rained five years ago?

Wenbin Liang   (4 November 2005)

Ambivalence and regret are stressors with their own psychological sequelae

Emily Peterson   (7 November 2005)

Self-Fulfilling Prophecy

Name and address supplied   (11 November 2005)

Debates about our design are beside the point: The Reardon and Cougle findings are invalid and cannot be reproduced with properly coded data.

Nancy F. Russo, Sarah J. Schmiege   (18 November 2005)

Signing the blue form

James Gerrard   (19 November 2005)

Re: Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations

Anne-Marie Rey   (2 December 2005)

Post Abortion Emotional Health

Margaret R. Johnston   (7 December 2005)

Re: Post Abortion Risk Assessment

Douwe A A Verkuyl   (10 December 2005)

Re: Post Abortion Emotional Health

David C, Reardon   (12 December 2005)

Reardon Response to Russo and Schmeige Misleads by Omission and Commission

Brenda N Major, Ph.D.   (14 January 2006)

Depression and unwanted first pregnancy: Methodological issues, additional findings

Nancy F. Russo, Sarah J. Schmiege   (10 February 2006)

Obscuring the Suffering of Women and Compromising Science

Priscilla K. Coleman   (25 March 2006)

The Reality of Post Abortion Trauma 28 October 2005
Margaret RJ Cuthill, Director British Victims of Abortion G1 3BU Send response to journal: Re: The Reality of Post Abortion Trauma	As a woman who has experienced two abortions a live birth and has for the last eighteen years worked in the area of Post Abortion Counselling I cannot agree with the results of this study by Professor Nancy Russo. I would suggest that the criteria being used to measure and compare whether depression is worse after abortion than giving birth to an unwanted child is too narrow and variable to allow a true picture of these very different life experiences and their resultant outcomes. Abortion is an unnatural death experience; with it comes grief, remorse, regret, and guilt associated with the choice. When these instinctive emotions/feelings remain unresolved, depression which can cause life-affecting symptoms, outwith the womans control, result. I believe this is just another study to reinforce to the abortion lobby and to persuade women in crisis that abortion is a valid option. I dont have to reinforce that abortion hurts women I see it daily. I look forward to the time when either more hurting women will speak out or someone will ask a different set of questions. Competing interests: Director of Post Abortion Counselling Group
Biased sample 29 October 2005
Fiona K Pinto, Project coordinator Royal College of Physicians, NW1 4LE Send response to journal: Re: Biased sample	In this study, the authors state that "differential exclusion of women from the delivery group on the basis of subsequent abortion creates a bias in favour of finding lower depression in that group." I strongly disagree. Opting for abortion more than once is likely to create a bias in exactly the opposite direction, on the basis that these women regard abortion as a solution, compared to women who had an abortion and subsequently did not choose to terminate another pregnancy. Furthermore, if the authors are truly interested in an unbiased sample of women who have had abortions, then why did they exclude women who had an abortion where the pregnancy was wanted? Surely an unbiased study would look at all women who had abortions not select out those who had abortions where the pregnancy was wanted? For example, David Reardon's study published in the Canadian Medical Association Journal, did not include this bias and found an increased rate of admission for women who had abortions to psychiatric units. Reardon's more comprehensive and objective research proves that all women need more support and alternatives to abortion and information about the possible harmful impact on their health rather than being falsely encouraged by biased samples such as this one that abortion is consequence free. BMJ 2002;324:151-152 ( 19 January ) Primary care Depression and unintended pregnancy in the National Longitudinal Survey of Youth: a cohort study David C Reardon, director of research, Jesse R Cougle, researcher. http://bmj.bmjjournals.com/cgi/content/full/324/7330/151 Psychiatric admissions of low-income women following abortion and childbirth David C. Reardon, Jesse R. Cougle, Vincent M. Rue, Martha W. Shuping, Priscilla K. Coleman and Philip G. Ney CMAJ • May 13, 2003; 168 (10) http://www.cmaj.ca/cgi/content/full/168/10/1253 Competing interests: None declared
Not very significant 31 October 2005
Patrick M R Leahy, Social science student Churchill College, Cambridge, CB3 0DS Send response to journal: Re: Not very significant	The media have unfortunately latched onto this study as something major (not helped by the authors' press release comments). It is anything but. Questions should be raised about the sample. It focuses on women aged 14-21 (p1) which is clearly problematic. These women are likely to be more "benefited" (if anyone is to be) by an abortion as any pregnancy is likely to disrupt educational, career, or other ambitions. Quite clearly, if generalisations about the whole female population are going to be made then a sample of a wider age range is required. The study also excludes "wanted" pregnancies - in fact the authors go so far as to deliberately exclude these individuals (p2). Yet, quite clearly, there might be circumstances where the pregnancy is wanted but an abortion has to be undertaken. The authors justify this by claiming that they are trying to remain as close as possible to the conditions of the previous study. Yet they remain selective about this since, for instance, they did not, contrary to the previous study, exclude women who had had multiple abortions. Both these points make the claim that this is a "nationally representative sample" (p1) rather questionable. I would also have thought it appropriate for the authors to clarify what their measure of depression was - it is not clear from the article. All references refer to: BMJ, doi:10.1136/bmj.38623.532384.55 (published 28 October 2005) Competing interests: None declared
Re: Not very significant 1 November 2005
Heather R Caserta, Homemaker Home:77339 Send response to journal: Re: Re: Not very significant	I just had to respond to this. Age definately makes a difference. For myself, my 1st abortion was at age 19 and my 2nd at age 23. I am now 33 and have just come to a place in my life where it is apparent that I need to deal with these things. If you would've asked me then, I would've agreed that "I was Ok, That it was for the best, That it didn't have any affect on me." Now, 10-15 years later, I am still suffering from those choices. The depression, drug addiction, alcohol abuse...those things numbed the feelings and I kept them hidden. Now my reality is that I have to relive the expereinces, I have to deal with the issues and try to heal. If you ask a woman/child, who has an abortion because of an unwanted pregnancy...it's a relief to most. And maybe for years it still feels that way. But there comes a time when it must be dealt with. Still others suffer daily from their choices. These researchers should attend some of the Post-Abortion web-sites or support groups and really see what abortion does to men & women. Each has a different story...but the outcome is the same. Competing interests: None declared
Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations 1 November 2005
David Reardon, Director Elliot Institute, Springfield, IL 62791 USA Send response to journal: Re: Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations	Dear Editor, I appreciate the opportunity Rapid Response provides to reply the new analysis of the NLSY data regarding associations between long-term depression and outcome of first unintended pregnancy presented by Schmiege and Russo.[1] This study was prepared in response to our own analysis of the same data previously published in the BMJ,[2] with additional data published in a subsequent BMJ letter.[3] Our original short report revealed in a single table that among married women who had an unintended first pregnancy an average of eight years prior an assessment of depression risk, those who had aborted their unintended pregnancy were at a significantly higher risk of depression (OR=1.92 95% CI 1.23 to 2.97)3 than women who carried the unintended pregnancy to term, with controls for income, education, race, age at first pregnancy and a pre-pregnancy score for psychological state using the Rotter internal-external locus of control scale. Stratification by marital status was a central feature of our study as there were no significant differences between women who were unmarried at the time of the depression assessment in regard to outcome of the first unintended pregnancy. The significance of the marital status was discussed in companion pieces and will be further discussed below. The lengthy repetition of our findings above is offered to underscore that in omitting any stratification of their results by marital status, Schmiege and Russo have failed to offer readers a true reanalysis of the data and findings we presented. Though they present four tables and one figure to analyze other aspects of the data using new selection criteria, they have failed to present the most basic evidence necessary to either refute our confirm our prior results. Namely, they have failed to reconstruct our table using data generated from their coding and selection rules, including stratification by marital state and the use of our control variables. It is my hope that in the confident spirit of openness, the authors will publicly affirm their willingness to share their recoded data with our research team and others so that any who are interested can more completely evaluate their new coding approach and may also complete the final reconstruction of our table using this potentially improved coding system. While the study of Schmiege and Russo currently fails to directly address our findings regarding differences in depression rate, which are most evident among women married at the time of the depression assessment, as discussed above, it should be noted that there are a number of other significant flaws and weaknesses in the design, presentation, and discussion presented. First, it should be noted that the authors’ claim in the “What we already know” section stating “Well designed studies have not found that abortion contributes to an increased risk of depression,” is misleading and not supported by their discussion, their citations, or the literature. In fact, the statistical association between abortion and higher depression rates is very firmly established by many well designed studies.[4],[5],[6],[7],[8],[9],[10],[11],[12],[13] Indeed, one of the authors of the this newest study,1 Russo, was herself the lead author of a study of 2,525 women which revealed that women who had abortions had significantly more depression, suicidal ideation, and lower life satisfaction than other women.13 While it is true that Russo argued in that paper13 that this association with depression might possibly be explained by greater exposure to experiences of violence among women who have abortion, the act of simply proposing this hypothesis serves to demonstrate the fact that the irrefutable evidence of an link between abortion and depression requires explanation and further investigation. While the hypothesis that the link between abortion and depression may be explained by some common risk factors certainly deserves additional investigation, the more straight-forward case for a causal link between abortion and depression is strongly supported by the self-attribution of women who have been interviewed about their psychological reactions to abortion[14],[15],[16],[17] and also by clinical experience of counselors who have successfully treated post-abortion depression.17, [18],[19] Furthermore, as suicide is closely associated with depression, some of the best statistical evidence for a causal link between abortion and depression is found in large record based studies linking a dramatic increase risk of suicide[20],[21],[22],[23], suicide attempts[24] and suicidal ideation13 following abortion. Case studies of suicide attempts[25],[26],[27] directly linked to a traumatic abortion experiences strongly support a causal interpretation of these record based findings.22 The preceding discussion of the literature is important because uninformed readers may wrongly infer from the authors’ presentation that our NLSY study2 is the only study linking abortion to depression. In fact, it was simply the first study to compare women who have had abortions to a control group of women who have carried unintended pregnancies to term. The authors’ failure to even partially mention this larger body of literature, including Russo’s own research, may unfortunately lead readers to conclude that since the questions the authors’ are raising about our own NLSY analyses are sufficient enough to prove that there is in fact no link between abortion and depression. Certainly, this has been exactly the impression fostered by numerous news reports sparked by the authors’ new study. For example, both Reuters[28] and WebMD[29] headlined their stories inspired by this study definitively declaring, without any qualifications, “Abortion Does Not Raise Depression Risk.” The articles themselves also and included reporting which reflected numerous other overgeneralizations that go far beyond the data presented. Specifically regarding the authors’ new analysis, it is first important to note that their results do not contradict ours; indeed, they do not even attempt to reconstruct our analysis in regard to stratification by marital status using their recoded coded data. Even what they do report, as noted by Schmiege and Russo, can easily be reconciled with our own findings since any differences in results “can primarily be explained by differences in coding of key variables and sample selection.”1 While the authors argue that their coding of key variables and redefinition of the parameters for sample selection are superior, I am skeptical of this claim. From what they have reported, the first questionable decision they made in recoding the data was the choice to eliminate from the sample women who had abortions who had subsequently indicated that they had wanted their pregnancies at some point before they decided to have an abortion. Ambivalence about pregnancy and abortion is common. I do not doubt that there was a group of women who had swings in “wanting” their pregnancies, which ended in a decision against keeping the pregnancy. Indeed, ambivalent swings from wanting to unwanting a pregnancy is well-known risk factor for emotional turmoil after an abortion.[30] For example, research by Husfeldt and colleagues found that 44% of participants experienced doubts about a decision to abort upon confirmation of their pregnancies, while 30% continued to have doubts on the day of their abortions.[31] Eliminating this subset of women may significantly bias the analysis by eliminating a class of women who have abortions who may be at highest risk of post-abortion depression. Moreover, as the authors note on the last page of their discussion, in most cases the classification of wantedness was made many years after the abortion and may therefore reflect a certain amount of recall bias. Bottom line: these women did in fact have abortions, so they belong in the abortion column. However, I do agree that in a larger data set with better "real time" measures of intended and wantedness, it would be very appropriate to run a separate comparison of women who aborted originally wanted pregnancies with women who aborted pregnancies that were not originally wanted. This data set is simply not so robust, however. Secondly, the authors increased the percentage of control cases identified at risk of depression by deciding to include in the control group women who carried their first pregnancy to term but then had a subsequent abortion. This is very problematic since the measure of depression occurred, on average, eight years after the first pregnancy. In other words, the authors have statistically added any depression associated with abortions of second or third pregnancies to the control group of who delivered a first unintended pregnancy. Indeed, there is some evidence that women who have an abortion after giving birth to living children are at greater risk of emotional sequelae compared to women without children.[32],[33] Comparing the mental health of women who aborted a first unintended pregnancy to a group including women who aborted subsequent pregnancies simply does not help to clarify interpretation of the data. Instead, it only serves to muddy the data and increase the likelihood that any statistical comparisons will not detect significant results. Our own decision to exclude from the control group women with a subsequent abortion was intended to reduce the confounding effects of multiple pregnancies. Indeed, our restriction of focus on the outcome of first unintended pregnancies was entirely driven by the fact that it is extremely difficult to control for multiple pregnancy experiences, especially with such a limited data set. One can well make the case that in addition to eliminating cases which had subsequent abortions we should also have eliminated cases where women had subsequent unintended pregnancies. Or, most ideally, to eliminate the many confounding effects of multiple pregnancies with different outcomes and different levels of wantedness, one could limit the comparison to women who never had a second pregnancy. Unfortunately, such a restrictive approach would have severely limited the number of cases that could be drawn from the data set and would also have made results even less generalizable since most women will have multiple pregnancies. Still, narrowing the inclusion criteria in such a fashion would have been far more rational than expanding the criteria, as Schmiege and Russo have done to add women exposed to abortion to the control group of women who did not abort their first pregnancy. Another difficulty in evaluating this new study is that the author’s post-1979 analysis, which was intended to parallel our analysis which used the 1979 Rotter control variable, identified 38% fewer cases of women exceeding the depression cutoff score. We reported 196 cases among delivering women to their 111 cases, and 80 cases above the cutoff for women who aborted compared to their identification of 55 cases.3 Having not yet had an opportunity to examine the researchers’ data, I have no idea how coding changes resulted in such a significant drop in the number of cases identified by the authors. In any event, this loss of cases would clearly reduce statistical power and therefore serves to increase the probability that their study would find no significant results. Another problem with the authors’ paper revolves around a secondary analysis concerning the very high rate of non-reported abortions in the NLSY data set. Compared with national averages, only 40% of the expected number of abortions are reported by women interviewed in the NLSY. [34] As noted in our original short report, since shame, secrecy, and thought suppression regarding an abortion are all associated with greater post­abortion depression, anxiety, and hostility,[35] the fact that 60% of abortions are concealed from NLSY interviewers would almost certainly dilute our ability to accurately measure the association between abortion and depression. While Schmiege and Russo do not call readers attention to the 60% concealment rate, they do attempt to address the questions raised by the high concealment problem by undertaking an examination of depression scores relative to whether or not women completed a private abortion history assessment. The fundamental, unaddressed problem with this approach is that it assumes that only those women who did not turn in the abortion history assessment were engaged in concealing past abortions. Absent evidence to the contrary, willingness to accept and return an abortion history card does not demonstrate a willingness to fully disclose one’s abortion history. Furthermore, the authors’ conclusion that women who are prone to depression are more likely than women not prone to depression to complete an abortion history card and report an abortion is not supported by a reasonable explanation for why women would behave in this way. Are there actually any studies showing that persons prone to depression are more willing to disclose past abortions? A citation to this effect would be most helpful. Even if it were true, as they claim, that “women who are willing to disclose abortion are also more willing to disclose stigmatizing mental health problems, such as depression,” the converse does not necessarily follow: namely, that women who are unwilling to disclose abortions experience less depression. In my reading of this study, the only contribution the authors actually make to an understanding of our earlier published results is found in their identification of the fact that by electing to use a control variable for psychological state before the first pregnancy, the 1979 Rotter scale, we lost 80 percent of the teenage pregnancies, due to the age distribution of the NLSY cohort prior to 1980. The authors are therefore correct that our findings may not be generalizable to minors. If we had noticed this at the time, we would have eliminated all cases of teen pregnancies and thereby narrowed our conclusions. But nothing in the authors’ paper reverses our findings regarding the group of women that were selected. At most, it simply narrows our findings to mostly non-teenage women, a group which makes up the majority of women who have abortions. At this point it is useful to recap the importance of our findings in regard to marital status. In short, we found nearly a doubling of risk for depression among married women at the time of the depression assessment who aborted an unintended first pregnancy an average of eight years previously as compared to married women who carried a first unintended pregnancy to term. This doubling of risk was found after removing the effects of education, race, age at first pregnancy, and pre-pregnancy psychological state as measured by 1979 administered Rotter score.3 As noted, our results for unmarried women were not significant. In other words, we found that women who aborted a first unintended pregnancy and were still single an average of eight years later had a similar risk for depression compared to single mothers who gave birth to a first unintended pregnancy—even though they were not faced with the same challenges of being a single parent.. While single women who had abortions had a very similar depression rate compared to married women who had aborted their first pregnancy, among women who delivered their unintended first pregnancy, those who were unmarried in 1992 were twice as likely to be at risk of depression compared to those who were married. In other words, abortion did not reduce the risk of depression for any of the groups we examined, but marriage did – unless the married woman had previously aborted her first unintended pregnancy. As shown in the table reproduced below, when cases for married and unmarried women are combined, the association between abortion and depression is statistically significant, but not strongly. It would therefore be expected that the kinds of changes in coding employed by Schmiege and Russo discussed above would easily dissolve the significance found for all women. This is why a reanalysis using the control variables we employed with stratification for marital status is particularly important, since the more robust significance found in the comparison for married women may well have persisted even after their realignment of classification rules. Women scoring in range for high risk of depression (CES depression score >15) who had their first abortion or unintended childbirth between 1980 and 1992 Women with unintended births but no subsequent abortions Women who had an abortion Adjusted odds ratio Total No (%) at high risk Total No (%) at high risk Adjusted odds ratio (95% CI)* Unmarried† 253 91 (36) 129 37 (29) 0.88 (0.54 to 1.43) Married† 530 101 (19) 164 43 (26) 1.92 (1.23 to 2.97) In first marriage 443 78 (18) 131 35 (27) 2.23 (1.36 to 3.64) All women 783 192 (25) 293 80 (27) 1.39 (1.02 to 1.90) *Adjusted for family income, education, race, age at first pregnancy, and 1979 Rotter score. †When CES depression questionnaire was administered in 1992. This brings us to still another problem. The authors observe that given their selection criteria, women who gave birth to an unintended pregnancy prior to 1980 had significantly higher risk of depression compared to women who reported aborting their first pregnancy prior to 1980 and compared to women who had a first unintended pregnancy after 1980, whether they carried to term or had abortions (Table 1).1 For reasons not discussed, the authors decided in this case to report these results without any controls for age, race, income, or education, much less marital status, which proved to be so important in our own analysis. This omission of all control variables is especially odd since several control variables are used in Table 2. 1 Unfortunately this inconsistent use of control variables should have been avoided since it might arouse the suspicion of some readers that the authors have selectively used or omitted control variables to produce results which would most appear to support their hypotheses. Yet another problem with the authors pre-1980 sample is that it heavily loaded with women who were younger at the time of there first pregnancy and older at the time of the depression assessment in 1992. Comparing these women to the post-1980 group is very problematic without any controls for age. Moreover, interpretation of these results is further complicated by the fact that women in the NLSY under age 20 at the time of their pregnancies have the highest concealment rate for abortion among all age groups.34 It is therefore impossible to know how many women classified as having delivered their first unintended pregnancy prior to 1980 actually had a prior, unreported abortion. The importance of this confounding effect is underscored by literature which demonstrates that many women who have abortions, especially minors, will become pregnant again, within two years, [36],[37],[38] and many will then carry the subsequent pregnancy to term. Therefore, in their discussion of this data point, the authors not only fail to control for available confounding variables, such as age, they also miss the opportunity to discuss how their findings should be cautiously interpreted in light of the NLSY high concealment rates. A more accurate evaluation of the points raised above will be possible once others have had a chance to more carefully understand and evaluate the new coding system employed by Schmiege and Russo. As Russo, is a honored member of the American Psychological Association (APA), I am confident the authors will publicly confirm their willingness to share their recoded data with myself and other researchers as required by rules for sharing data set forth by the APA.[39] As stated above, at a minimum the newly coded data must be used to reconstruct of our original table stratified by marital status. This brings me back to the original purpose of our study which was to investigate the hypothesis that the previously observed increased risk of depression following abortion might be inconsequential compared to equal or higher rates of depression that might be associated with giving birth to an unintended child.4 While our original short report did not provide room for this discussion, I will offer it here. In a follow-up study of 442 women who had abortions, researchers tracked depression scores using the Brief Symptom Inventory (BSI) one hour post-abortion, one month post-abortion, and two years post-abortion.4,5 At the two year follow-up, approximately 50% of the women either refused to participate in the follow-up evaluation or could not be contacted. Among those who did participate in the two year post-abortion assessment, depression scores were significantly higher than their one hour post-abortion scores, though higher one hour post-abortion scores were also significantly predictive of higher depression scores two years later. 5 In addition to these important findings, the researchers found that 24.5% of the women remaining in their sample at the two-year followup had scores above the cutoff for clinical depression on the BSI depression scale.5 Curiously, rather than registering alarm, the researchers’ erroneously asserted that the depression rate detected in their study was only slightly over that of American women in general by reference to a study of national prevalence conducted by Blazer, Kessler, McGonagle, and Swartz,[40] which indicated a 20% lifetime prevalence rate of major depression among women 15-35 years of age. The reason this was assertion was erroneous is that the researchers mistakenly compared their scores for depression in the most recent month to Blazer’s findings regarding lifetime prevalence rates. Fortunately, Blazer’s group also reported the prevalence of current (30 day) major depression for females aged 15-24 and 25-34, as 8.2% and 4.3% respectively.40 Thus, when the proper comparison is made for most recent month depression rates, these follow-up abortion studies4,5 actually found that depression rates two years after abortion were 3 to 5 times higher among women who have had an abortion compared to the general population of similarly aged women. In my opinion, this finding that one-fourth of women two years after their abortion had high depression scores should have motivated the researchers to encourage more detailed pre-abortion screening and post-abortion counseling. But curiously, the authors appeared to generally dismiss the importance of their own findings on the basis of the hypothesis that giving birth to an unwanted pregnancy would likely incur equal or greater psychological price, 4 a theme echoed by Schmiege and Russo.1 This is an extremely important hypothesis, but it is also a hypothesis, which up to that point, had never been tested. In the most recent, comprehensive review of the literature on emotional reactions to abortion and future research priorities,[41] the absence of studies examining psychological adjustment following an unintended pregnancy using control groups (comparing those who abort to those who carry to term) is identified as a major shortcoming of the existing literature. Our original study appears to be the first to employ this very appropriate control group.1 We have continued to pursue this direction with two other studies that have compared the psychological wellbeing of women who carry unintended pregnancies to term with women who have abortions: one finding higher risk of long-term generalized anxiety disorder among women who abort[42] and the other finding higher rates of substance abuse.[43] While the observation of higher rates of depression, anxiety, and substance abuse among women who have had abortions compared to similar women who have carried unintended pregnancies to term is itself significant, the body of research using these control groups is admittedly very sparse. But even if these studies did not produce evidence of significantly higher adverse effects associated with abortion, what is most telling is that they have failed to produce any evidence of the benefits widely claimed for abortion. Indeed, it is most remarkable that even though Schmiege and Russo themselves fail to find any statistically significant benefit in reduced risk of depression associated with abortion--despite their best efforts to reconfigure the selection criteria--they instead proclaim that “Abortion may be indirectly associated with a lower risk of depression through beneficial effects on education, income, and control of family size.” This statement has been welcomed by reporters covering this study as an affirmation of commonly held assumptions, but in fact it is based on nothing more than hopeful inferences. Indeed, our analysis of the same data, after controlling for education and income, demonstrates there is no indirect beneficial effect on reducing depression. Furthermore, at least for women who subsequently marry, abortion is significantly associated with an increased risk of depression. When interpreted in the light of the large body of literature discussed earlier, this positive association does not appear to be incidental and nothing presented in Schmiege’s and Russo’s analysis contradicts this finding. In fact, the NLSY is a very weak sample for studying psychological adjustments of abortion. Not only does it have a 60% concealment rate, but only a few questions related to psychological state and reproductive health are collected, and these are collected only once every several years because the NLSY is actually designed to study labor experiences, not health. That it includes any health variables is fortunate, but those it does have are only sufficient to catch a hint of what women who have abortions actually experience. The possibility of using the NLSY for abortion research was first identified by Nancy Russo, a co-author of the present study.1 In her first analysis of the NLSY, Russo examined of self-esteem scores collected at two points in the NLSY cycle.[44] In that study, Russo concluded that that the lack of any statistically significant decline in self-esteem among women admitting having had an abortion compared to delivering women proved that abortion does not have “a substantial and important impact on women’s well-being.” However, none of the major inadequacies of the NLSY data set, in particular the high concealment rate, were discussed. Indeed, Russo instead assured readers that the NLSY contained sufficient “size and variability in the critical variables” to detect any negative health effects of abortion and therefore her finding of no negative impact on self-esteem provided a solid basis for concluding that there are not significant negative effects of abortion on women’s “well-being.” While I would strenuously disagree with Russo’s claim that the NLSY is a sufficiently sensitive tool for reaching definitive conclusions regarding the potential risks and benefits of abortion,44 I do agree it provides at least a few useful variables that can be used to get a glimpse at how at least some women may adjust to abortion and the birth of unintended pregnancy. Indeed, it is precisely because the NLSY is a very imperfect research tool that Schmiege and Russo’s approach, as well as our own, is subject to so much second guessing. If this current disagreement regarding analytic methods and conclusions proves anything, it proves the need for a far better data set that would be made available to a larger number of researchers for a much deeper and conclusive analysis than is currently possible. Therefore, while I don’t expect Schmiege and Russo to embrace the criticisms I have raised above, it is my sincere hope that they, and anyone else interested in the issue of women’s health, will embrace our recommendation9,41 for federal funding of major longitudinal cohort study carefully designed to fully examine the associations between obstetric history and emotional well-being. A properly designed study with annual assessments and interviews that would be made available to all researchers, much as is currently done with the NLSY, would enable researchers to gauge the interactions among mental health and childbirth, parenting, adoption, abortion and miscarriage of wanted, unwanted, planned, and mistimed pregnancies. In addition, this collection of data would allow researchers to better investigate the effects and interactions of marriage, divorce, single parenting, multiple partners, domestic violence, PMS and menstrual irregularities, contraceptive practices, and similar factors related to reproductive and mental health and to wide variety of psychological, medical, and social issues that are uniquely related to women. Even if Schmiege and Russo were to object to everything else I have written above, I am optimistic that they will agree to the importance of a longitudinal study such as we have described. Such a study will finally produce the comprehensive data needed to reach incontrovertible conclusions regarding the positive and negative health effects of unintended pregnancies and abortion and deserves the support of all researchers interested in women’s health. CITATIONS (Links to full articles are provided where known) [1] Schmiege S, Russo NF. Depression and unwanted first pregnancy: longitudinal cohort study. [2] Reardon DC, Cougle JR. Depression and unintended pregnancy in the National Longitudinal Survey of Youth: a cohort study. British Medical Journal. 2002; 324:151-2. http://bmj.bmjjournals.com/cgi/reprint/324/7330/151 [3] Reardon DC, Cougle JR. Depression and unintended pregnancy in young women: Authors Reply. BMJ. 2002;324:1097. http://bmj.bmjjournals.com/cgi/eletters/324/7330/151#top [4] Major B, Cozzarelli C, Cooper ML et al: Psychological responses of women after first trimester abortion. Arch Gen Psych, 2000; 57(8): 777-84. [5] Cozzarelli, C., Major, B., Karrasch, A., & Fuegen, K. (2000). Women’s experiences of and reactions to antiabortion picketing. Basic and Applied Social Psychology, 2000;22:265-275. [6] 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. http://www.cmaj.ca/cgi/reprint/168/10/1253 [7] Gould NB. Postabortion depressive reactions in college women. Journal of American College Health Association. 1980; 28, 316-320. [8] Moseley DT, Follongstad DR, Harley H, Heckel RV. Psychological factors that predict reaction to abortion. Journal of Clinical Psychology. 1981; 37, 276-279. [9] Cougle J, Reardon DC, Coleman PK. Depression associated with abortion and childbirth: A long-term analysis of the NLSY cohort. Medical Science Monitor, 1003; 9(4), CR105-112. http://www.medscimonit.com/pub/vol_9/no_4/3074.pdf [10] Bradley CF. Abortion and subsequent pregnancy. Canadian Journal Psychiatry 1984; 29:494. [11] Soderberg H, Janzon L, Sjoberg NO: Emotional distress following induced abortion: A study of its incidence and determinants among adoptees in Malmo, Sweden. Eur J Obstetr Gyn Reprod Biol, 1998; 79: 173-8. [12] LO Linares et al. Predictors of repeat pregnancy outcome among black and Puerto Rican adolescent mothers. Developmental and Behavioral Pediatrics 13(2):89, 1992. [13] Russo N, Denious JE. Violence in the lives of women having abortions: Implications for policy and practice. Professional Psychology Research and Practice, 2001); 32:142-150. [14] Coleman PK, Nelson ES.The quality of abortion decisions and college students' reports of post-abortion emotional sequelae and abortion attitudes. Journal of Social and Clinical Psychology, 1998; 17, 425-442. [15] Rue VM, Coleman PK, Rue JJ, Reardon DC. Induced abortion and traumatic stress: A preliminary comparison of American and Russian women. Med Sci Monit, 2004 10(10): SR5-16. http://www.medscimonit.com/pub/vol_10/no_10/4923.pdf [16] Congleton, G. & Calhoun, L. (1993). Post-abortion perceptions: A comparison of self-identified distressed and non-distressed populations. International Journal of Social Psychiatry 1993; 39, 255-265. [17] Burke T, Reardon DC. Forbidden grief: the unspoken pain of abortion. Springfield (IL): Acorn Books, 2002. [18] Gould NB. Postabortion Depressive Reactions in College Women. J.Am. College Health Association. 1980; 28:316-320. [19] Franco K, et al. Anniversary Reactions and Due Date Responses Following Abortion, Psychother Psychosom 1989; 52:151-154, 1989. [20] Gissler M. Hemminki E, Lonnqvist J. Suicides after pregnancy in Finland: 1987-94: register linkage study. British Medical Journal 313:1431-4, 1996. http://bmj.bmjjournals.com/cgi/content/full/313/7070/1431 [21] Gissler M, Berg C, Bouvier-Colle MH, Buekens P. Injury deaths, suicides and homicides associated with pregnancy, Finland 1987-2000. Eur J Public Health. 2005 Oct;15(5):459-63. [22] Reardon DC, Strahan TW, Thorp JM, Shuping MW. Deaths associated with abortion compared to childbirth: a review of new and old data and the medical and legal implications. The Journal of Contemporary Health Law & Policy 2004; 20(2):279-327. http://www.afterabortion.org/research/DeathsAssocWithAbortionJCHLP.pdf [23] Reardon DC, Ney PG, Scheuren FJ,, Cougle JR, Coleman, PK, Strahan T Deaths associated with pregnancy outcome: a record linkage study of low income women. Southern Medical Journal. 95(8):834-841, August 2002. http://www.afterabortion.org/research/DeathsAssociatedWithAbortion.pdf [24] Morgan CM, Evans M, Peter JR, Currie C: Mental health may deteriorate as a direct effect of induced abortion. Br Med J, 1997; 314: 902. http://bmj.bmjjournals.com/cgi/content/full/319/7205/318 [25] Tischler C, Adolescent suicide attempts following elective abortion. Pediatrics, 1981; 68(5): 670-1 [26] Greenglass ER. Therapeutic abortion and psychiatric disturbances among women. Canadian Psychiatric Association Journal. 1976;21:453-459. [27] Garfinkle B, Hoberman H, Parsons J, Walker J. Stress, Depression and Suicide: A Study of Adolescents in Minnesota (Minneapolis: University of Minnesota Extension Service, 1986). [28] http://news.yahoo.com/s/nm/20051028/hl_nm/abortion_depression_dc_1 (accessed 10/28/05) [29] http://www.foxnews.com/printer_friendly_story/0,3566,173880,00.html (accessed 10/28/05) [30] Major B, Cozzarelli C: Psychological predictors of adjustment to abortion. J Soc Iss, 1992; 48: 121-142. [31] Husfeldt C, Hansen SK, Lyngberg A, Noddebo M, Pettersson B. Ambivalence among women applying for abortion. Acta Obstetricia et Gynecologia Scandinavica, 1995; 74, 813-817. [32] Lask B. Short‑term Psychiatric Sequelae to Therapeutic Termination of Pregnancy. Br J Psychiatry, 1975; 126: 173‑177. [33] Peppers LG, “Grief and Elective Abortion: Implications for the Counselor,” Disenfranchised Grief: Recognizing Hidden Sorrow, ed. Kenneth J. Doka (Lexington Books: Lexington MA, 1989), pp.135‑146. [34] Jones EF, Forrest JD: Under reporting of abortion in surveys of U. S. women: 1976 to 1988. Demography, 1992; 29: 113-26. [35] Major B, Gramzow RH. Abortion as stigma: cognitive and emotional implications of concealment. J Pers Soc Psychol. 1999;77:735-45. [36] Horowitz NH. Adolescent mourning reactions to infant and fetal Loss. Soc. Casework 1978; 59 (1978): 551. [37] Wheeler SR “Adolescent Pregnancy Loss,” Loss During Pregnancy or the Newborn Period, edited by J.R. Woods Jr. and J.L. Woods, 1997. [38] H. Cvejic et. al., “Follow-up of 50 adolescent girls 2 years after abortion,” Canadian Medical Association Journal, 116:44, 1997. [39] American Psychological Association, Ethical Standards for Reporting and Publishing of Scientific Information. http://www.apa.org/journals/authors/openletter.pdf [40] Blazer DG, Kessler RC, McGonagle KA, Swartz MS. The prevalence and distribution of major depression in a national community sample: The National Comorbidity Survey. American Journal of Psychiatry. 1994; 151, 979-986. [41] Coleman PK, Reardon DC, Strahan T, Cougle JR. The psychology of abortion: a review and suggestions for future research. Psychology and Health 2005; 20(2):237-271. [42] Cougle JR, Reardon DC, Coleman PK. Generalized anxiety following unintended pregnancies resolved through childbirth and abortion: a cohort study of the 1995 National Survey of Family Growth. J Anxiety Disord. 2005;19(1):137-42. [43] Reardon DC, Coleman PK, Cougle JR. Substance use associated with unintended pregnancy outcomes in the National Longitudinal Survey of Youth. Am. J. Drug and Alcohol Abuse. 2004; 26(1):369 - 383. [44] Russo NF, Zierk K: Abortion, childbearing, and women's wellbeing. Prof Psychol: Research and Practice, 1992; 23: 269-80. Competing interests: None declared
Ambivalence and regrets are not the same as Depression 3 November 2005
Joan A. Lang Lang, Professor & Chair, Dept of Psychiatry Saint Louis University Send response to journal: Re: Ambivalence and regrets are not the same as Depression	The scientific study of the psychological sequelae of such important life events as abortion or carrying to term an unwanted pregnancy is of great importance. Not only do doctors and therapists need to know as much as possible in order to help our patients, but we also are often asked to support or criticize public policies, up to and including legislation. When we do this, integrity demands that we do so thoughtfully and in keeping with whatever evidence is available, not just on the basis of our personal opinions, however deeply held those may be. The authors are to be commended on this careful analysis, and for neither under nor overplaying their results. They conclude that the existing evidence does not support the assertion that there is a link between abortion and depression. They do not attempt to argue that this lack of evidence supports any assertion that abortion (or carrying to term) of an unwanted pregnancy is without any sequelae. In my experience, most women who have struggled with the decision of whether to keep or to abort such an unwanted pregnancy have many feelings during and after the time of decision making. But few if any have depression because of the abortion per se. Regret and wishes that things could have been different do not in themselves constitute depression. Where depression is found, it is usually either a preexisting condition, or related to the circumstances (such as an unsupportive partner and/or lack of family support). In the case of an unwanted pregnancy that is kept, certainly the outcome may be in net a happy one for the mother, but it can also be a postpartum depression and difficulties bonding with the infant, among other things. Competing interests: None declared
Post Abortion Risk Assessment 4 November 2005
Cynthia M. Dudek, Medical Insurance Agent 23800 W. 10 Mile Rd., Suite 180; Southfield, MI 48034 Send response to journal: Re: Post Abortion Risk Assessment	There is a need for each of us to be seekers of truth and to realize that no one has all of the answers. We need to commit ourselves to the Hippocratic Oath, and to do no harm. As a medical insurance agent, I would ask others to consider the fact that abortion is a causal element and an avoidable risk factor for many consequences. Those consequences may be compounded, as is the case for women who are exposed to estrogen as a result of a pregnancy and are unable to bring a child to term. There are women who will never go through the differentiation stage of breast development and thereby not receive the benefits of a reduced risk of breast cancer. The details matter when assessing risk. We need to consider age at first birth, family history, previous history of breast cancer, age at menarche, parity, contraceptive use, BRCA1 and BRCA2 genes, diet, breastfeeding history, obesity in postmenopausal women, and more. Abortion consequences result in women needing more medical attention. It impacts a woman’s future insurability. It impacts the health of their future children. All of these elements are factors in the high cost of medical insurance. Women deserve better! Cynthia Dudek Planning Resources, Inc. cdudek.planningresources@rc.net Competing interests: None declared
The prevalence picture—Could today's water level of a lake tell how much it rained five years ago? 4 November 2005
Wenbin Liang, taking master of public health Curtin University of Technology Send response to journal: Re: The prevalence picture—Could today's water level of a lake tell how much it rained five years ago?	Dear Editor, In the paper, the assessment of depression in 1992 was an estimation of the prevalence of depression among the population of interest.[1] The prevalence was at least determined by the incidence of depression in the population, the duration of depression, and the mortality rate of patients. Therefore, even if abortion for “an unintended first pregnancy” is a strong risk factor for depression, it may only highly increase the incidence of depression for the first few years after abortion. A measurement on prevalence can hardly reflect any effect of abortion on the incidence of depression. Nevertheless if an association was showed to be significant, there was still no information for the relationship between abortion and the risk of becoming depression. For example if “abortion people” and “live birth people” had the same incidence rate, and “live birth people” were more likely to be cured in a shorter time, then the depression prevalence for the “abortion people” would be higher than the depression prevalence for the “live birth people”. Nevertheless if “abortion people” had a higher incidence of depression and they were more likely to be cured in a much shorter time then the depression prevalence for them would be similar to or even lower than the depression prevalence for the “live birth people”. Even if depression is considered as a disease that can never be cured, and patients would never die. Prevalence may still fail to refect the true relationship between abortion and the risk of depression, because abortion for “an unintended first pregnancy”, could be a component cause that was only involved in some types of sufficient causes. Over a long period the effect of abortion on prevalence would be easily diluted by the effect of other sufficient causes. For example, assuming that abortion could increase the incidence of depression in the first 2 years after abortion by 100% among a hypothetical population with 10000 people: 5000 people choosing abortion, 5000 people choosing to give birth, and the incidence of depression was 5 per 1000 person-year among people choosing to give birth. So the incidence of depression for “abortion people” is 10 per 1000 person-year for the first two years. In the first two years there are (5/1000) *2*5000=50 cases among “live birth people”, (10/1000)*2*5000=100 cases among “abortion people” The prevalence ratio is 0.5. 10 years later the prevalence ratio of depression is approximately {[50+(5/1000) *8*5000]/5000}/{[100 +(5/ 1000)*8*5000]/5000}=0.8333—We could hardly observe any relationship here, as the effect has been diluted. Reference 1. Schmiege S., Russo N., Depression and unwanted first pregnancy: longitudinal cohort study. BMJ, doi:10.1136/bmj.38623.532384.55 (published 28 October 2005) Competing interests: None declared
Ambivalence and regret are stressors with their own psychological sequelae 7 November 2005
Emily Peterson, Blogmistress, the http://afterabortion.blogspot.com Send response to journal: Re: Ambivalence and regret are stressors with their own psychological sequelae	I understand Dr. Joan Lang Lang's point that experiencing ambivalence and regret after an abortion is not the same thing as being depressed. I don't think anyone would disagree with that. It's not the issue that this study takes up. I imagine that persistent feelings of ambivalence and regret about a major life decision (whatever that major life decision was), left unresolved, would themselves be life stressors that might impair a person's emotional health. Whether this is so would be a valuable area for research. I appreciate Dr. Lang Lang's belief that in her practice, when a woman has (a) had an abortion and (b) is depressed, that it turns out that it is never the case that the depression is related to the abortion. Other therapists have reached very different conclusions about some of their post-abortive clients. These anecdotes are interesting but of course not dispositive. I appreciate the work and analysis put into this current study. Disputes about which variables to count and who to put into what population aside, my main concern about the study is that it only measures one adverse psychological state: depression. People react differently to life stressors. Some people develop anxiety symptoms, some people develop substance abuse issues, and so on. A study that looks at the full range of adverse psychological states and behaviors that might potentially ensue after a difficult life decision such as abortion would be of greater interest than the current study. Competing interests: None declared
Self-Fulfilling Prophecy 11 November 2005
Name and address supplied, N/A N/A Send response to journal: Re: Self-Fulfilling Prophecy	"Each has a different story...but the outcome is the same." This claim is preposterous. I had an abortion several months ago. Abortion is never pleasant, but it hasn't had a negative effect on my physical & emotional health. It's not like I don't "deal with the issues;" I've often reflected upon my experience. The only thing that's made me upset about the experience is other people's negative reaction, like Caserta's and Cuthill's. When people expect me to have all sorts of problems, I feel like maybe I should have problems and wonder if something's wrong with me for not having problems. It's not like I got pregnant on purpose to have an abortion, yet people seem to want to guilt trip women like me. Leading women to think they'll be screwed up after abortion isn't a good way to decrease the number of abortions. It most probably increases the number of screwed up women. But I suspect that making women feel bad is not a problem to people who think that abortion is a sin that should cause women to feel, as Cuthill says, "grief, remorse, regret, and guilt." The number of women on post-abortion counseling or websites are not enough support the claim that most post-abortion women are emotionally disturbed. That number, however large, is only a small fraction of the total number of women who've aborted, which is millions upon millions. According to my human sexual behavior textbook, the majority of women who have had an abortion do not regret their choice and are not depressed. The results of this study by Schmiege and Russo are in concordance. So, the emotional problems of women on post-abortion counseling and websites seem to be the exception, not the rule. I'm psychologically fine perhaps because I have a promising future and because my supportive boyfriend (who's also my future husband & father to my future kids) and my liberal college setting don't make me feel like I should be disturbed. My suggestion is that we stop telling women that it is inevitable or "instinctive" to feel "grief, remorse, regret, and guilt" and be depressed after abortion. Instead, we should provide women with the support and information to prevent future unplanned pregnancies and help them with career development or life planning so that they can look forward to their future, not dwell on their past. Career development seems to help, since "higher mean education and income and lower total family size" are associated with a lower risk of depression. Also, aspirations for higher education and a job outside of the home seem to decrease the likelihood of teenage pregnancy. Help women; don't make us suffer. Competing interests: None declared
Debates about our design are beside the point: The Reardon and Cougle findings are invalid and cannot be reproduced with properly coded data. 18 November 2005
Nancy F. Russo, Regents Professor Arizona State University; Tempe, AZ85282, Sarah J. Schmiege Send response to journal: Re: Debates about our design are beside the point: The Reardon and Cougle findings are invalid and cannot be reproduced with properly coded data.	The rapid responses to our article[1] have raised a variety of issues. Here we address the methodological issues that have been raised by David Reardon [2] and others, as noted.  To avoid redundancy, theoretical, ethical, and clinical issues raised across the responses will be addressed in subsequent responses.  In his response, Reardon takes the opportunity to repeat the results presented in his original article (co-authored with Jesse Cougle), arguing that we have “failed to present the most basic evidence necessary to either refute or confirm our prior results.”  He states “their results do not contradict ours; indeed, they do not even attempt to reconstruct our analysis in regard to stratification by marital status using their recoded coded data. Even what they do report… can easily be reconciled with our own findings since any differences in results ‘can primarily be explained by differences in coding of key variables and sample selection’.”   1. Results based on miscoded data are not in need of refuting and cannot be reconciled. In our attempt to take a collegial approach, we have perhaps not been clear about the basic methodological message of our article. The previous results were based on miscoded data such that first unintended pregnancies and their outcomes (pregnancy versus abortion) were not accurately identified. We are not talking about a difference in interpretation of how best to code a particular variable, e.g., as when researchers need to decide whether to use all possible categories for marital status or to create a variable that only has 2 categories (i.e., married and unmarried). We are talking about the use of codes to identify the variable the former authors themselves claimed to have measured.   Results based on miscoded data are not in need of confirmation or refutation, as results based on such data are invalid and meaningless. As stated in our article, the NLSY dataset is large and complex, with over 3,000 variables related to pregnancy. As can be seen by the coding syntax appended to this response, identification of first pregnancy is particularly complex, as it involves combining variables across multiple survey years. Even the most competent researcher may have difficulty in writing codes that accurately specify the variable “first unintended pregnancy”.  Because of the enormous room for error in choosing the proper variables and coding of variables in this large dataset, we chose to rely on the variable selection strategy and coding language provided to us by an expert in the use of the survey: a member of the NLSY staff. Reardon suggests that we should make our codes available so that our findings can be checked. The codes, which are in SAS, can be found at the end of this response. We are confident that the codes produce valid results, i.e., that they create the variable that we are claiming to measure. The coding was provided to us by NLSY staff and the logic was rechecked several times. However, if there is a problem with this coding language, we welcome learning about it. As embarrassing as that would be, our goal is to produce accurate research findings. We welcome constructive feedback on how to achieve that goal.   2.  There was no need to reproduce previously reported tables based on results from analyses that we considered inappropriate for the research question. Our study was designed to test the hypothesis that was offered (but tested with miscoded data) in the earlier study.  In doing so, we found several design assumptions that we did not consider appropriate for testing that hypothesis.  We therefore did not incorporate them in our study, but did explain the rationale for our decisions.  We also noted that we indeed did conduct analyses to parallel those of the previous study by examining pregnancies limited to 1980 and later, but did not find significant results. Typically journals have limited space and are not receptive to publishing nonsignificant findings that are based on inappropriate analyses, so we did not include them in our paper.  Nonetheless, we are happy to present these analyses here and to elaborate the rationale for our decisions. Most of the design issues relate to the claim that our sample is biased because: (a) we did not exclude those in either the abortion or delivery group who had subsequent abortions after the first pregnancy (submitted by Fiona Pinto); (b) we did not stratify our results by marital status  (submitted by David Reardon); (c) we excluded abortions where the pregnancy was reported as “wanted” (submitted by Patrick Leahy);and (d) the age range of our sample was limited (submitted by Patrick Leahy).   2a. If the goal is to generalize the results to women having a first unintended pregnancy, exclusion of women with multiple abortions from either the delivery group or the abortion group is inappropriate.  Exclusions of women with multiple abortions from only the delivery group are doubly inappropriate.  Whether one agrees with this view is irrelevant, however, because the results are the same in any case. Women dealing with an unwanted first pregnancy need information that can help them assess how choosing one outcome versus another (i.e., delivery vs. abortion) will contribute to a change in risk for physical and mental health outcomes. We could have eliminated women having multiple abortions from both groups, but without a crystal ball, women who will have multiple abortions cannot be separated from other women. Consequently, information based on such a sample has no useful medical purpose.  But regardless of whether one agrees with our judgment, as we clearly stated on p. 2 of the manuscript and as can be seen in the following tables, the conclusions did not differ when women were excluded on the basis of subsequent abortion, regardless of whether that exclusion was from both the abortion and delivery groups, or simply the delivery group.   Logistic and OLS regression results if all women with subsequent abortions are excluded:                                                No. (%) exceeding            Unadjusted                  Adjusted                                                                           CES-D cut-off                OR (95% CI)            OR (95% CI)    Dichotomous CESD: Delivery Group                         185/646 (28.6)            .66 (.48 to .90)        1.14 (.76 to 1.70)*                        Abortion Group                          67/322 (20.8)                 P = 0.009                     P = 0.54                                                             Mean (SD)               Unadjusted                    Adjusted                                                                                            Beta (S.E)                   Beta (S. E)     Continuous CESD: Delivery Group                            11.8 (9.9)                    -2.0 (.65)                      .36 (.78)*            Abortion Group                             9.8 (9.1)                    P = 0.002                      P = 0.64       *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                                   Logistic and OLS regression results if only women in the delivery group with subsequent abortions are excluded:                                                  No. (%) exceeding            Unadjusted                  Adjusted                                                                            CES-D cut-off                OR (95% CI)            OR (95% CI)                Dichotomous CESD: Delivery Group                         185/646 (28.6)             .82 (.63 to 1.08)      1.17 (.82 to 1.65)*                    Abortion Group                        119/479 (24.8)                  P = 0.16                     P = 0.39                                                              Mean (SD)               Unadjusted                    Adjusted                                                                                            Beta (S.E)                   Beta (S. E)     Continuous CESD: Delivery Group                            11.8 (9.9)                  -1.01 (.60)                    .26 (.70)*             Abortion Group                            10.8 (9.9)                    P = 0.09                      P = 0.71        *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                                   2b. Results do not differ by current marital status. Arguments could be made about whether to stratify by marital status the year depression was measured or the year the unwanted pregnancy occurred, but  they are irrelevant here because the findings are consistent across marital status.  We stratified by marital status in an earlier iteration of the paper, but dropped the table because it didn’t contribute information of significance. The following table presents the same findings found in our Table 2, but is now stratified on the basis of “married” versus “unmarried” in 1992.  As can be seen, the findings are consistent across marital groups for both our logistic and OLS regression results.   Updates of Table 2, stratified by marital status in 1992                                                   Logistic Regression predicting CES-D cut-off scores                                                    No. (%) exceeding            Unadjusted                   Adjusted                                                                           CES-D cut-off                   OR (95% CI)            OR (95% CI)       Full Sample Unmarried Women: Delivery Group                         133/369 (36.0)              .83 (.58 to 1.18)       1.32 (.83 to 2.09)#                    Abortion Group                        71/223   (31.8)                     P = 0.30                      P = 0.25      Married Women: Delivery Group                         85/393  (21.6)                .81 (.55 to 1.21)       1.08 (.67 to 1.76)#                    Abortion Group                        46/251   (18.3)                    P = 0.31                      P = 0.75       All Women: Delivery Group                         220/768 (28.6)              .82 (.64 to 1.07)       1.19 (.85 to 1.66)*                    Abortion Group                        119/479 (24.8)                    P = 0.14                     P = 0.30 Post 1979 Pregnancies only   Unmarried Women: Delivery Group                          65/200 (32.5)                .91 (.55 to 1.51)      1.39 (.71 to 2.71)&                    Abortion Group                         33/108 (30.6)                        P = 0.73                  P = 0.34        Married Women: Delivery Group                          45/256  (17.6)               .98 (.55 to 1.72)      1.31 (.69 to 2.49)&                     Abortion Group                        21/122   (17.2)                       P = 0.93                  P = 0.42        All Women: Delivery Group                         111/457 (24.3)              .97 (.67 to 1.40)       1.33 (.84 to 2.10)†                   Abortion Group                          55/232 (23.7)                       P = 0.87                 P = 0.23        Note. Pregnancy outcome coded "0" for delivery and "1" for abortion; higher CES-D scores indicate greater levels of depression. Odds ratios greater than 1.0 would thus indicate a higher risk for depression in the abortion group *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                               # Adjusted for race, age at first pregnancy, education, and family income           & Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, education, and family income  † Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, and 1992 marital status, education, and family income                  OLS Regression predicting CES-D continuous scores                                                    Mean (SD)                        Unadjusted                         Adjusted                                                                                                 Beta (S.E)                       Beta (S. E)                                       Full Sample Unmarried Women: Delivery Group                             13.6 (10.4)                          -.63 (.89)                   1.55 (1.13)#                    Abortion Group                            13.0 (10.6)                           P = 0.48                       P = 0.17             Married Women: Delivery Group                               10.1 (9.0)                          -1.31 (.72)                   -.55 (.82)#                    Abortion Group                              8.8 (8.7)                             P = 0.07                        P = 0.51            All Women: Delivery Group                            11.8 (9.95)                         -1.07 (.58)                     .38 (.68)*                Abortion Group                             10.8 (9.9)                           P = 0.06                       P = 0.58   Post 1979 Pregnancies only   Unmarried Women: Delivery Group                             13.0 (10.1)                          .35 (1.28)                   1.80 (1.63)&                    Abortion Group                            13.4 (11.7)                           P = 0.78                       P = 0.27             Married Women: Delivery Group                               8.8 (8.0)                            -.37 (.90)                     .56 (1.00)&                     Abortion Group                              8.4 (8.4)                             P = 0.68                         P = 0.58           All Women: Delivery Group                             10.7 (9.4)                            .07 (.78)                      .92 (.87)† Abortion Group                            10.8 (10.4)                           P = 0.93                       P = 0.29   Note. Pregnancy outcome coded "0" for delivery and "1" for abortion; higher CES-D scores indicate greater levels of depression.  Positive beta coefficients would thus indicate a higher risk for depression in the abortion group. *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                               # Adjusted for race, age at first pregnancy, education, and family income           & Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, education, and family income  † Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, and 1992 marital status, education, and family income    2c.  In a study of unwanted pregnancy outcome, it is inappropriate to include wanted pregnancies in the sample; it is doubly inappropriate to exclude wanted pregnancies from only the delivery outcome group.   The stated purpose of our study was to examine the relationship between pregnancy outcome and depression of unwanted first pregnancies.  We chose this as our research question and designed our analyses accordingly.             But regardless of one’s opinion about the appropriateness of the particular design strategy, as we state in the paper only 15 women (of a total sample of n=1247) who terminated their first pregnancy by abortion reported that pregnancy as wanted. Although we excluded these women because of design considerations, redoing the analyses with these 15 women included produces no change in the results.   2d. The age range of our sample was not limited to 14-21 years of age.  Our description of the age of the sample appears to be misinterpreted by Patrick Leahy. The age range of women was 14-21 in 1979, the year the study began. The year of first pregnancy ranged from 1970 to 1992, and the age of women at the first pregnancy ranged from 12-33.   In sum, there may be debates about our design decisions, which we defend.  However, the most important point here is that these design issues do not affect the outcome of the results when data are properly coded.   3. The critique of our paper, as well as the public discussion of research on the relationship of abortion and mental health outcomes, has exhibited a persistent failure to distinguish between correlation and causation in the interpretation of results.   This failure is clearly illuminated in Reardon’s statement:  “First, it should be noted that the authors’ claim…stating “Well designed studies have not found that abortion contributes to an increased risk of depression,” is misleading and not supported by their discussion, their citations, or the literature. In fact, the statistical association between abortion and higher depression rates is very firmly established by many well designed studies.”[italics ours] “ Reardon then cites a number of studies, some more well designed than others, but none that have established that abortion contributes to an increased risk of depression.  Indeed, he goes on to describe an article by Russo and Denious [3}, repeating their findings that women who had abortions had significantly more depression, suicidal ideation, and lower life satisfaction than other women.  He then describes those researchers as arguing that “this association with depression might possibly be explained by greater exposure to experiences of violence among women who have abortion” and asserts “the act of simply proposing this hypothesis serves to demonstrate the fact that the irrefutable evidence of a link between abortion and depression requires explanation and further investigation.” What he fails to say, is that Russo and Denious did not “simply propose” this hypothesis.  They tested it and found that when exposure to violence and partner variables were controlled, no relationship was found between abortion and the negative mental health outcomes measured. Reardon then goes on to cite as evidence that abortion causes depression “self-attributions of women,” “clinical experience of counselors,” “case studies,” and statistical evidence based on studies that have not included the basic controls needed to warrant causal conclusions.  None of this is credible scientific evidence that abortion increases risk for depression.  Many studies, including our own, have found correlations between abortion and a host of mental health outcomes. Given that abortion typically occurs in the context of unwanted pregnancy, this association is to be expected. The error is in focusing on abortion rather than the conditions that lead to risk for unplanned and unwanted pregnancy. As Russo and Denious pointed out, having a history of childhood sexual abuse and exposure to intimate violence is associated with risk for unplanned and unwanted pregnancy, whether or not such pregnancy ends in abortion or delivery.  In this context, focusing on the effects of abortion rather than trying to understand the relationship of abortion to mental health outcomes can lead to misattribution of the effects of childhood physical and sexual abuse, intimate partner violence, and other adversities. We will deal with issues related to discrepancies between scientific, clinical, and anecdotal evidence raised by the various responses as well as in issues relating to media coverage and future research needs, including the need for research on the effects of underreporting, in separate replies. For now however, we stand by the statement that there is no credible scientific evidence that abortion increases risk of depression.   [1] Schmiege S, Russo NF. Depression and unwanted first pregnancy: longitudinal cohort study. [2] Reardon DC, Cougle JR. Depression and unintended pregnancy in the National Longitudinal Survey of Youth: a cohort study. British Medical Journal. 2002; 324:151-2. http://bmj.bmjjournals.com/cgi/reprint/324/7330/151 [3]  Russo N, Denious JE. Violence in the lives of women having abortions: Implications for policy and practice. Professional Psychology Research and Practice, 2001); 32:142-150.   SAS SYNTAX libname in 'g:\british'; OPTIONS SYSPARM='.'; OPTIONS NOCENTER; OPTIONS LS=80;   DATA a; SET in.nlsy_1110; rename R1017500=wantp183; rename R1325400=wantp184; rename R1522056=outp1_84; rename R1702800=wantp185; rename R1892758=outp1_85; rename R2002900=wantp186; rename R2259858=outp1_86; rename R2579200=wantp188; rename R2879700=outp1_88; rename R3188900=wantp190; rename R3409800=outp1_90; rename R3792500=wantp192; rename R4009468=outp1_92;   data b; set a; if R0214800=2; proc freq; tables wantp183 wantp184 outp1_84 wantp185 outp1_85 wantp186 outp1_86 wantp188 outp1_88 wantp190 outp1_90 wantp192 outp1_92; data c; set b; p1_out=.; if outp1_92>-4 then p1_out=outp1_92; if outp1_90>-4 then p1_out=outp1_90; if outp1_88>-4 then p1_out=outp1_88; if outp1_86>-4 then p1_out=outp1_86; if outp1_85>-4 then p1_out=outp1_85; if outp1_84>-4 then p1_out=outp1_84;   wanted=.; if p1_out>-4 and p1_out<6 then do; if outp1_84>-4 and wantp183>=-3 then wanted=wantp183; else if outp1_84>-4 and wantp184>=-3 then wanted=wantp184; else if outp1_85>-4 and wantp185>=-3 then wanted=wantp185; else if outp1_86>-4 and wantp186>=-3 then wanted=wantp186; else if outp1_88>-4 and wantp188>=-3 then wanted=wantp188; else if outp1_90>-4 and wantp190>=-3 then wanted=wantp190; else if outp1_92>-4 and wantp192>=-3 then wanted=wantp192; else if wanted=. and outp1_92 in (2 3 4 5) then wanted=999; end;   data d; set c; if wanted=3 or wanted=4 then do; if p1_out = 1 then pregout = 0; end; if wanted ne 1 then do; if p1_out=4 then pregout = 1; end;   data cesd; set d; x = n (R3894900, R3895000, R3895100, R3895200, R3895300, R3895400, R3895500, R3895600, R3895700, R3895800, R3895900, R3896000, R3896100, R3896200, R3896300, R3896400, R3896500, R3896600, R3896700, R3896800);   if x = 20 then cesdcont = sum (R3894900, R3895000, R3895100, R3895200, R3895300, R3895400, R3895500, R3895600, R3895700, R3895800, R3895900, R3896000, R3896100, R3896200, R3896300, R3896400, R3896500, R3896600, R3896700, R3896800);   if cesdcont>=0 then do; cesddich=.; if cesdcont>15 then cesddich=1; else if (cesdcont> 0) or (cesdcont<16) then cesddich=0; end; proc freq; tables pregout cesddich pregout*cesddich; run;   proc format; value p1_out 1='live birth' 2='miscarriage' 3='stillborn' 4='abortion' 5='still pregnant';   value pregout 0='live birth' 1='abortion';   value wanted 1='wanted to be preg' 2="didn't matte" 3='no, not then' 4='no, never' 999='not asked, 1992';   value cesddich 0='below cutoff' 1='above cutoff'; run;     Competing interests: None declared
Signing the blue form 19 November 2005
James Gerrard, General Practitioner Windmill Health Centre, Mill Green View, Leeds LS14 5JS Send response to journal: Re: Signing the blue form	In the United Kingdom, terminations of pregnancy are performed under the Abortion Act after two doctors have signed certificate A, "the blue form". In the vast majority of cases the form is signed under section C, indicating that continuing the pregnancy would involve greater risk to the physical or mental health of the pregnant woman than if the pregnancy were terminated. The study by Schmiege and Russo appears to show that delivering an unwanted first pregnancy is not associated with a greater risk of depression than terminating it. Other studies indicate that there is a lower risk of mortality during pregnancy and until one year after birth compared to women without a recent pregnancy(1). These findings seem to offer support to any UK doctor who feels unable to sign the blue form in good faith. (1) Lewis G, editor. Why Mothers Die 2000-2002. The Sixth Report of the Confidential Enquiries into Maternal Death in the United Kingdom. London: RCOG Press;2004 Competing interests: I am a GP with a conscientious objection to termination of pregnancy. I do not sign the blue form.
Re: Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations 2 December 2005
Anne-Marie Rey, volunteer Abortion-information Send response to journal: Re: Re: Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations	David Reardon claims that their study "was simply the first study to compare women who have had abortions to a control group of women who have carried unintended pregnancies to term". This is wrong. The best prospective study comparing women who had an abortion with women who carried an unplanned pregnancy to term, with a 10 years follow up and 6'000 women in each group, was the one of A.C. Gilchrist et al "Termination of pregnancy and psychiatric morbidity", British Journal of Psychiatry 1995; 167:243-248. They found no increase in the relative risk of psychiatric morbidity in the abortion group. Curiously enough Reardon does not include this study in his long list of literature... Nor does he mention the Danish study of H.P. David "Post-abortion and post-partum psychiatric hospitalization", In: Abortion: medical progress and social implications. Pitman, London 1985 (Ciba Found.Symposium 115) p.150-164. The risk of admission was about the same for women who had an abortion and those who had given birth: 12/10'000. Competing interests: Responsible for the Website Abortion-information http://www.svss-uspda.ch
Post Abortion Emotional Health 7 December 2005
Margaret R. Johnston, Exec Dir. Southern Tier Women's Services Vestal NY 13850 Send response to journal: Re: Post Abortion Emotional Health	Every day I listen to women who are trying to make a good pregnancy decision for their lives. If, after consideration, they decide to end their pregnancy, I help them, and I also do my best to be on the alert for any of the risk factors that researchers like Dr. Nancy Russo and Dr. Brenda Major have identified in their ground breaking work. We have discussions with women about their ability to cope with their decisions, what tangible emotional support they have, their religious or spiritual concerns, and the social factors that lead them to believe that having a child is not the right choice for this time in their lives. I find it ironic that people who are opposed to abortion, like David Reardon, are actively encouraging a climate of stigma, judgment, and shame which is surely not helpful to anyone's sense of emotional health. If there is a way to study the perception of societal stigma and subsequent emotional health it would be very illuminating. There is, perhaps, a misconception that only women who are prochoice get abortions, but on some days, it seems that up to 1/3 of all of the women seeking abortion state that they have previously "not believed in abortion" or in some other way not been in favor of the practice of abortion. And yet, after careful consideration of their choices and the consequences to them and their families, they choose abortion. And for the most part, they seem to find some resilience in their response to this change of perspective in spite of the extremely negative attitude toward abortion in public discourse. It may be a comfort to people in the research world that providers, and others in helping professions, are committed to Post Abortion Emotional Health, and are trying to use available scientific evidence to ensure positive outcomes for our patients. We have no interest in denying that some women have negative emotional outcomes. We know that with such a major life decision as whether or not to bear a child, there are bound to be strong feelings and reactions. We also know that some women have far fewer coping abilities, less social support, or pre-existing mental problems that make this more challenging. We are trying to figure out how to help these women which sometimes means encouraging them to take more time for their decision, consult family or clergy, or seek professional help from a therapist. Women have been choosing to end pregnancies since the beginning of recorded history. As an abortion provider and as someone who cares deeply about women, my goal is to help women make good decisions for their lives and help them achieve good emotional health. The conflict about abortion seems to focus on whether abortion is right or wrong but it leaves out the people involved. If the Abortion Wars did not dominate this discussion of research, think how much help we could offer real women and men. Competing interests: None declared
Re: Post Abortion Risk Assessment 10 December 2005
Douwe A A Verkuyl, Gynaecologist Hoogeveen Hospital, Send response to journal: Re: Re: Post Abortion Risk Assessment	The medical insurance agent C.D. Dudek threatens women who have an abortion with an increased risk of breast cancer. There is no proof for this assertion I think, please quote the paper. On the other hand never to feed breast and never to become pregnant while having 480 ovulations in your life will increase the risk of ovarian and breast cancer, probably. Perhaps she should dissuade women to become nuns. Competing interests: None declared
Re: Post Abortion Emotional Health 12 December 2005
David C, Reardon, Director Elliot Institute, Springfield, IL Send response to journal: Re: Re: Post Abortion Emotional Health	I am disappointed that the BMJ editors chose to publish Ms. Johnston’s baseless accusation (“Post Abortion Emotional Health”) that I am “actively encouraging a climate of stigma, judgment, and shame.” Clearly Ms. Johnston either knows nothing about me or is so interested in perpetuating a denigrating stereotype of those “who are opposed to abortion” that she doesn’t care if her judgmental comments are actually true. In fact, beginning with the publication of my first book in 1986 I have been a well-known and active leader in efforts to promote a healing environment for those who have had abortions. A centerpiece of this activity has been an unremitting effort to create cultural changes that will reduce or eliminate external causes of “stigma, judgment, and shame,” especially among those who oppose abortion. One of my most succinct articles on this issue is “Beyond the Politics of Abortion” (http://www.afterabortion.org/hope/arti2.htm) In that article I advise those who hold the pro-choice view to not dismiss the feelings of loss, grief, and regret that many women feel after an abortion. I also admonish those who identify themselves as pro-life that they have an obligation to develop empathy for the circumstances that drive women to consider abortion. In particular, pro-life Christians should remember that only those who have not sinned are qualified to “throw the first stone”—and none have yet been found who qualify. In promoting empathy for the circumstances of women who choose abortions, I frequently remind those who oppose abortion on moral grounds that no one who “has not been there” knows what pressures or circumstances might lead them to consider having an abortion that was previously unthinkable. (Ms. Johnston has noted that she has observed this in her own practice. Indeed, there is evidence that at least 70 percent of American women seeking abortions are doing so either in violation of their moral conscience or in conflict with their maternal desires. In my opinion, these conflicted feelings prior to abortion play a major role in negative reactions post-abortion.) In short, I frequently admonish Christian’s—who by definition are nothing more than sinners who seek and rely on God’s mercy they see manifested in the life of Jesus—to remember the maxim “there but the grace of God go I.” In doing so, they will quickly remember that “loving the sinner while hating the sin” means that they should take the lead in helping women and men who are plagued with unresolved grief after a past abortion to recover emotional and spiritual wellbeing. In the same context I have always opposed any efforts (which in my experience are actually very rare) to provoke “stigma, judgment, and shame” since I believe such tactics are firstly un-Christian and secondarily counterproductive. If one checks with any of the hundreds of post-abortion ministries in America, most of which are led by women who have had abortions, I believe one will discover that my efforts to reduce stigma and promote a healing climate are widely appreciated and may have contributed to the growth of the post-abortion healing movement. Competing interests: None declared
Reardon Response to Russo and Schmeige Misleads by Omission and Commission 14 January 2006
Brenda N Major, Ph.D., Professor of Psychology University of California, Santa Barbara Send response to journal: Re: Reardon Response to Russo and Schmeige Misleads by Omission and Commission	As a scientist who has conducted a number of studies of women’s emotional responses following abortion, I am compelled to respond to David Reardon’s reply [1] to Russo and Schmeige’s [2] re-analysis of the NLSY data regarding the purported association between abortion of a first unwanted pregnancy and depression. Reardon’s reply misrepresents scientific knowledge regarding this association and misleads the reader in a number of ways, a few of which I will address here. First, Reardon falsely implies that “a causal link between abortion and depression is strongly supported,” and cites studies of women who have been interviewed about their psychological reactions to abortion to support his claim. The studies he cites, however, are not based on representative samples of women who have had abortions, but rather on biased samples of women who have self-identified as feeling depressed resulting from abortion. Responses of these women cannot be generalized to the general population of women who have abortions. Even if these samples were representative, however, responses to these interviews would not establish a causal link between abortion and depression. Strong evidence exists that people can perceive a causal connection between events in their lives when none in fact exists. The infamous witchcraft trials of Salem, MA, in which young women came to believe that physical symptoms they were experiencing were caused by witchcraft, provide a compelling example of this fallacy. Notably, when representative samples of women who have had an abortion are interviewed about their psychological reactions to abortion, most (but not all) report being satisfied with their decision, feel they were benefited more than harmed by their decision, and state that they would make the same decision again had they to do it over again. [3] Second, Reardon misleads the reader by falsely claiming that studies showing a higher rate of suicide among women who have had an abortion than among women who have had a child “strongly support a causal interpretation.” This conclusion is wrong. Evidence of a correlation between two things says nothing about whether one causes the other. We know for a fact, for example, that there is a positive association between the number of churches in a city and the number of bars in a city. Does the fact that this association has been found repeatedly mean that alcohol drives people to church? Or that church drives people to drink? No. The most likely explanation for this association is that it is spurious – caused by a third unmeasured variable that is related to both number of churches and number of bars in a city, such as city-size. There is substantial evidence that a third variable (or more) also underlies the observed association between abortion and suicide. A likely candidate for this third variable is a woman’s general mental health prior to her becoming pregnant. A longitudinal study in which individuals were interviewed as teenagers (age 15-16) and again nine years later as young adults (age 24-25) revealed that women who had experienced depression as a teenager were more likely to have abused drugs, dropped out of school, engaged in deviant behavior, and been hospitalized for psychiatric reasons during the subsequent nine years compared to women who had not been depressed as teenagers. [4] Women who experience psychiatric problems as teenagers, compared to those who do not, may also be more likely to engage in unprotected sex, be in unstable relationships, and have unintended pregnancies, all of which are associated with abortion. Third, Reardon misleads readers when describing the results of my own research. My colleagues and I conducted a longitudinal study in which we interviewed 442 women just prior to having a first trimester abortion of an unintended pregnancy, and again immediately after, one month after, and two years after the abortion. [2] In describing results of this study, Reardon states that “among those who did participate in the two year post- abortion assessment, depression scores were significantly higher than their one hour post-abortion scores.” What Reardon fails to report, however, is that depression scores in this sample were significantly lower at all times after the abortion (immediately after, one month after, and two years after) than they were prior to the abortion. Reardon also does not mention that the best predictor of women’s depression scores two years after their unplanned pregnancy/abortion was whether or not they had a history of experiencing major depression prior to the pregnancy for which they were seeking abortion. Twenty-six per cent of the women in this sample had experienced an episode of major depression at some time in their life prior to the pregnancy, based on their responses to a modified version of the Diagnostic Interview Schedule (DIS) [5]. This rate exceeds the lifetime prevalence of major depression among a nationally representative sample of women age 15-24 in the US (20.8%), as determined by the National Comorbidity Survey (NCS) [6, 7]. Significant correlates for major depression in the NCS included being female, poorly educated, unmarried (single/divorced/widowed/ or never married), and poor. [6] Since all of these are characteristic of women seeking to terminate an unintended pregnancy, it is not surprising that lifetime major depression is more common among this population than it is among the general population. The high prevalence of lifetime major depression among women seeking to terminate an unintended pregnancy supports the argument that psychiatric problems prior to an unintended pregnancy are a likely “third variable” that explains the higher incidence of mental health problems (e.g., depression, suicide, drug abuse) observed among women after an abortion. Fourth, Reardon misrepresents the prevalence of major depression among women who have an abortion compared to the prevalence in the general population. We found that 20% of the women in our sample had experienced an episode of major depression in the 24 months after their abortion, based on their responses to the DIS [3]. This rate is comparable to the 12 month (16%) and lifetime (20.8%) rate of major depression found among women age 15-24 in the general population [6,7]. People with a prior history of depression have substantial risks of depression chronicity and 12-month recurrence. [8] In the NCS, the probability that a young woman (15-24) who had experienced an episode of major depression sometime in her life would experience another episode in the last 12 months was 78%. [8] Thus the incidence of major depression among women after abortion is consistent with their relatively high rate of major depression prior to becoming pregnant. Curiously, in describing the results of our study Reardon did not report the percentage of women in our sample who met criteria for diagnosis of major depression using the DIS. Instead, he reported the percent of women in the sample who scored higher than clinical cut-offs on the depression subscale of the Brief Symptom Index (BSI) two years postabortion. [9, 10]. Reardon then erroneously compared this percentage (24.5%) to the percentage of women aged 15-24 and 25-34 in the NCS who had experienced an episode of major depression during the preceding thirty days (8.2% and 4.3% respectively) to make the claim that “depression rates two years after abortion were 3 to 5 times higher among women who have had an abortion compared to the general population of similarly aged women.” This comparison of prevalence of depressive symptoms to prevalence of major depression is inappropriate. Although symptom scales such as the BSI are useful screening tools, they are not standardized diagnostic psychiatric instruments. Further, they are known to yield a very high rate of false positives [11]. In contrast, the DIS was designed to diagnose whether a patient's symptoms meet the criteria for major depression as described in the Diagnostic and Statistical Manual (DSM) of Mental Disorders (DSM-III-R). [12] Prevalence of major depression in the NCS was determined using an expanded version of the DIS [6, 7]. Thus prevalence rates based on these two instruments can be meaningfully compared. Doing so yields a strikingly different picture than that painted by Reardon. In sum, Reardon’s reply to the Russo and Schmeige article misleads the reader by misrepresenting scientific findings. 1. Reardon reply 2. Schmiege, S., Russo, NF. Depression and unwanted first pregnancy: Longitudinal cohort study. BMJ 2005;331:1303 (3 December), doi:10.1136/bmj.38623.532384.55 (published 28 October 2005) 3. Major B, Cozzarelli C, Cooper ML, Zubek J, Richards C, Wilhite M, Gramzow RH. Psychological responses of women after first-trimester abortion. Archives of General Psychiatry, 2000, 57:777-784. 4. Kandel DB., Davies M. (1986) Adult sequelae of adolescent depressive symptoms. Archives of General Psychiatry 1986; 43: 255-262. 5. Robins LN, Helzer JE, Croughan J, Ratcliff KS. National Institute of Mental Health Diagnostic Interview Schedule: its history, characteristics, and validity. Archives of General Psychiatry 1981;38:381-389. 6. Blazer DG, Kessler RC, McGonagle KA, Swartz MS The prevalence and distribution of major depression in a national community sample: The National Comorbidity Study. American Journal of Psychiatry 1994; 151:979- 986 7. Kessler, RC, McGonagle, KA, Zhao, S, Nelson, CB, Hughes,M, Eshleman S, Wittchen H-U, Kendler, KS: Lifetime and 12 month prevalence of DSM III-R psychiatric disorders in the United States: results from the National Comorbitiy Survey. Archives of General Psychiatry 1994; 51:8-19. 8. Kessler RC, McGonagle KA, Swartz M, Blazer DG, Nelson CB. Sex and depression in the National Comorbidity Survey I: Lifetime prevalence, chronicity and recurrence. Journal of Affective Disorders 1993; 29: 85- 96. 9. Derogatis LR, Melisaratos N. The Brief Symptom Inventory: An introductory report. Psychological Medicine. 1983, 13:596-605. 10. Cozzarelli C, Major, B, Karrasch A, Fuegen K. Women’s experiences of and reactions to antiabortion picketing. Basic and Applied Social Psychology, 2000; 22:265-275. 11. Roberts RE, Lewinsohn PM, Seeley JR. Screening for adolescent depression: A comparison of depression scales. Journal of the American Academy of Child and Adolescent Psychiatry. 1991; 30:58-66. 12. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association, 1994:339-345. Competing interests: None declared
Depression and unwanted first pregnancy: Methodological issues, additional findings 10 February 2006
Nancy F. Russo, Regents Professor Department of Psychology, Arizona State University, 871104, Sarah J. Schmiege Send response to journal: Re: Depression and unwanted first pregnancy: Methodological issues, additional findings	The rapid responses to Schmiege & Russo1 as well as emails directed to us personally have raised several methodological points that we address here with updated findings from additional analyses.   The goal of our research has been to examine the outcomes of unwanted first pregnancy (abortion or delivery) and risk of depression and to explain discrepancies with previous research by Reardon & Cougle2 that were based on the same data set. We concluded that the Reardon & Cougle finding of a correlation between abortion and depression was a result of miscoding the predictor variable, unwanted first pregnancy. In addition to examining the possibility of an abortion-depression correlation as hypothesized in the first study, we investigated potential indirect effects of abortion via effects of education, income, and reduced family size. We also conducted analyses that failed to find evidence that underreporting abortion may be masking an abortion-depression relationship.   The Reardon & Cougle study is now being used to as evidence that “abortion causes depression” despite the fact that the findings are correlational in nature and based on miscoded data. We have addressed the issue of invalidity of miscoded data in our previous response.3   Note that we do not generalize our results beyond unwanted first pregnancy nor do we suggest that our measure of depression risk encompasses all possible negative outcomes associated with terminating vs. delivering a first unwanted pregnancy.  Such generalizations were inappropriate for the previous study’s findings and are inappropriate for ours as well.     We emphasize the following methodological points:   Differential exclusion is an unacceptable sampling practice Pinto4 “strongly” disagrees with our statement that “differential exclusion of women from the delivery group on the basis of subsequent abortion creates a bias in favour of finding lower depression in that group," arguing that such an exclusion is likely to create bias in the opposite direction.    This statement reflects a lack of understanding of the research process. Differential exclusion – whatever direction the resulting bias may take – violates basic tenets of scientific research. In our study of unwanted first pregnancy, when the data are coded such that unwanted first pregnancies are truly identified, the difference between the abortion and delivery groups is not statistically significant.  These findings contrast with the previous erroneous findings of Reardon & Cougle2 no matter what sampling strategy is used – including when women with multiple abortions are excluded only from the delivery group and also when additional unwanted first pregnancies in the delivery group are identified increasing the power of the statistical analyses (see Tables 1-5b below).   The exception to this finding is when multiple abortions are excluded from both abortion and delivery groups and analyses are not statistically adjusted for covariates. In that case, only 20.7% of women in the abortion group exceed the CES-D cut-off score compared to 28.1% of the delivery group (P = .01).  This finding is not congruent with the claim that abortion on the first pregnancy poses a higher risk for depression for all women with unwanted first pregnancies, as erroneously reported by & Reardon & Cougle2 (see Table 5a).    Designs should reflect hypotheses being tested Pinto4 further states “if the authors are truly interested in an unbiased sample of women who have had abortions, then why did they exclude women who had an abortion where the pregnancy was wanted?” Leahy5 raises a similar question.   Both fail to recognize that when conducting research on the outcome of first unwanted pregnancy, it is inappropriate to include outcomes from wanted pregnancies in the analysis.  When a women challenged by an unwanted pregnancy evaluates her options, she needs to know their relative risks – i.e., what involves more risk, terminating or delivering that pregnancy.        Women who have wanted pregnancies rarely seek abortion.  In additional analyses based on expanding sampling as described above, we found only 17 women seeking abortion reported their unintended pregnancy was wanted; of that group, 5 (29.4%) exceeded the CES-D cut-off (mean = 12.2; S.D. 10.8). Although the number is too small for meaningful statistical analysis, it is comparable to the 28.3% of women exceeding the CES-D cut off score found in the delivery group, and higher than the 25.1% figure found for the abortion group.   Women with wanted pregnancies are more likely to seek abortion because of reasons related to maternal or fetal health, which may be discovered when the pregnancy at a more advanced stage of gestation, when there is more pain and physical risk involved. The NLSY data set does not permit differentiation of therapeutic from elective abortions. The nature and severity of risks involved for women who face therapeutic abortion differ markedly from those women who seek elective abortion of unwanted pregnancies (the vast majority of the latter occurring at early stages of pregnancy when medical risks are substantially lower than those related to childbirth).   In sum, although the relationship of abortion to mental health in women who terminate wanted pregnancies is important to study, it is not appropriate to generalize from one group to the other, and was not the focus of our study.   An additional 17 women in the abortion group reported “didn’t matter” in response to the unwantedness question. Although there is a strong argument for assuming that their choice of abortion should put them in the group of women who “did not identify their pregnancy as wanted,” their mean depression level was slightly lower than that of the abortion group overall. This number of women is so small whether they are included or excluded doesn’t change the results. However, including them does lower the mean depression levels in the abortion group (not significantly), so to avoid the appearance that we manipulated our sampling to produce specific outcomes, we did not include them in the expanded analyses. Between the analyses provided in our initial paper and the expanded analyses presented here we believe we have demonstrated that the lack of detecting a significant abortion-depression relationship as was initially reported in Reardon & Cougle is due to the miscoding of data in the original study and not because of design and sampling consideration in our study.    Research designs that seek to compare abortion and delivery groups but do not account for unwantedness of pregnancy do not provide an adequate test of the hypothesis that abortion is associated with increased risk for psychological disorder.   Abortion is confounded with unwanted pregnancy. As Major9 points out, there are a variety of conditions that lead to increased risk for both psychological disorder and unplanned and unwanted pregnancy, so that it is to be expected that women who have a history of abortion will have poorer mental health profiles than other women. Indeed, when confounded studies find no association with abortion and negative indicators, it raises the possibility that abortion might have protective mental health effects under some conditions   As Rey6 points out, Gilchrist et al.7 (1995) has conducted the most well-designed prospective study of the unplanned pregnancy outcome, and she is accurate in reporting those researchers found no increase in the relative risk of psychiatric morbidity in the abortion group.   However, additional findings from that study are relevant here.  First, previous psychiatric illness predicted subsequent psychiatric illness. That is, risk for psychiatric illness after abortion or delivery was predominately determined by history of illness before pregnancy. Findings also varied with history. Of interest here is the finding that among women with either a past history of non-psychotic illness or no history of psychiatric illness, women who delivered had a significantly higher likelihood of having a psychotic episode than those who had an abortion. The pattern was similar for women with a previous history of psychosis (although the difference was not statistically significant). 7   These findings, which involve actual psychiatric diagnoses, are much more relevant to “blue card decisions”8 of UK doctors than U.S. findings dealing with cut-off scores on a scale indicating risk for depression that has a high false positive rate.   The findings of this study stand in contrast to those of a number of studies that confound abortion and unwanted pregnancy and then report abortion to be correlated with a variety of problems. Such studies are simply not well-designed for the question of whether abortion itself increases risks for negative psychological outcomes apart from its association with unwanted pregnancy.   Thus, when Pinto refers readers to a study by Reardon and colleagues published in the Canadian Medical Association Journal10, describing it “more comprehensive and objective research [that] proves that all women need more support and alternatives to abortion and information about the possible harmful impact on their health rather than being falsely encouraged by biased samples such as this one that abortion is consequence free” – she could not be more mistaken.   The study referred to is one of two records-based studies using Medi-Cal data from the California health system that Reardon and colleagues have published, both with the same problems.10, 11 First, the studies are on low income women and can be generalized to other women only with great caution. Indeed, the results should not be generalized to women applying for abortions in the Medi-Cal system itself as the number of women who had abortions or deliveries and left the system are unknown, and it is easier to stay in the system when you have a new infant to support.   Second, the studies lack even the most basic statistical controls for race, marital status, or parity. Most importantly, there is no identification of whether the pregnancy was wanted or unwanted in either abortion or delivery groups. Multiple abortions are excluded only from the abortion group, and there is no way to ascertain previous psychiatric history. These are but a few of the problems with the approach of these ill-designed studies.   The only conclusion that can be drawn from such work is that low income women who have pregnancies terminated by abortion and stay in the Medi-Cal system over 4 years may at high risk for a variety of problems – but whether those women would be at higher risk for such problems had they chosen to deliver those pregnancies cannot be ascertained.   Similar problems are found in another analysis of the NLSY data by Cougle, Reardon, & Coleman,12 which basically repeats the analyses and assertions made in the previous article that was published in BMJ,2 except the intendedness of the first pregnancy is not identified. That study is based on miscoded data as well. Citation of the findings of such studies evidence that a pregnant women’s choice to terminate rather than deliver an unwanted pregnancy creates increased risk for depression is inappropriate.   Our previous findings underestimated the number of unwanted first pregnancies in the NLSY data set. As pointed out previously the NLSY data set is complex and difficult to code, so we based our coding on information provided by NLSY staff familiar with the data set. Although the first pregnancies were coded correctly in the iterative process used in our study, we stopped identifying pregnancy wantedness in 1983 when we should have stopped in 1982 (the year wantedness of first pregnancy was compiled for previous years).13   As a result, although the codes used identified all women included in the initial sample correctly as described (i.e., the data used were not miscoded), there was a large number of women in the delivery group with unwanted first pregnancies who were not included in the study, reducing the power of our statistical analyses to detect relationships. Thus, we reanalyzed the data to ensure that we had not failed to detect a significant relationship that might exist in the larger NLSY sample. Additional findings based on the expanded sample as described above are presented in Tables 1-5b   As can be seen from examination of the findings, including the additional women in the sample and increasing power does not change the significance of the results.  Whatever the comparison, whether based on our study’s design or the design of the Reardon & Cougle study which selected only those pregnancies occurring 1980 and later (and which we believe inappropriately uses differential exclusion in addition to miscoding the predictor variable), abortion is not found to be significantly related to higher depression risk.   Specifically, in Tables 1, findings again confirm that the exclusion of pre-1980 pregnancies in Reardon & Cougle’s previous design results in disproportionate exclusion of women in the delivery group who exceeded the CES-D cut-off score. Similarly, when comparing mean CES-D scores for the four groups, women who were in the pre-1980 delivery group had significantly higher depression scores than all other groups, who in turn did not differ significantly from each other.    Consequently, findings based on the post-1980 sample in either our sample or Reardon & Cougle’s sample should not be generalized to all first unwanted pregnancies. We provide the information below only to satisfy questions about how our findings compared to those of Reardon & Cougle.2   Table 1:  Proportion of participants in high-risk category for depression stratified by pregnancies before 1980 and those occurring from 1981 onwards     Group                                               No (%) exceeding                                                                                  CES-D cut-off                Odds Ratio (CI), p-value Delivery Group, pre 19801                    202/603 (33.5)             Delivery Group, post-1980                   161/679  (23.7)             .62 (.48 to   .79), P < .0001                  Abortion Group, pre-1980                       57/215 (26.5)              .72 (.51 to 1.01), P = .06                          Abortion Group, post-1980                    53/224  (23.7)              .62 (.43 to  .88), P =.007                    Group                                                     Mean (SD)                                                                                     CES-D Score              Mean difference (CI), p-value Delivery Group, pre-19801                    13.1  (10.8)                       Delivery Group, post-1980                    10.5   (9.0)               2.63 (1.54 to 3.71), P < .0001 Abortion Group, pre-1980                     10.96 (9.6)                 2.16 (0.52 to 3.80), P = .006                      Abortion Group, post-1980                   10.9   (10.6)               2.20 (0.54 to 3.85), P = .009 * n = 23 cases missing due to no data on the year the first pregnancy began   1. Referent group from which the other three groups are compared.  For each comparison, the pre-1980 delivery group was coded as “0” and each other group coded as “1” so that odds ratios less than 1.0 indicate greater depression in the pre-1980 delivery group.     As Table 2a shows, although a larger proportion of women in the delivery group than the abortion group exceeded the CES-D cut off score, this finding was not statistically significant.  The lack of significant differences between groups is a consistent pattern across the tables.   Table 2a: Odds ratios (95% confidence intervals) predicting depression cut-off scores from pregnancy outcome, stratified by marital status in 1992, unadjusted and adjusted for explanatory variables.                                         Logistic Regression predicting CES-D cut-off scores                                                   No. (%) exceeding               Unadjusted               Adjusted                                                      CES-D cut-off                    OR (95% CI)               OR (95% CI)       Full Sample All Women: Delivery Group                         363/1283 (28.3)                .84 (.66 to 1.08)            1.15 (.85 to 1.56)*                    Abortion Group                           115/461 (25.0)                       P = 0.17                           P = 0.36   Unmarried Women: Delivery Group                         222/630 (35.2)                .89 (.64 to 1.23)              1.31 (.85 to 2.0)#                    Abortion Group                           70/215  (32.6)                     P = 0.48                           P = 0.22   Married Women: Delivery Group                        137/393  (21.2)                .80 (.55 to 1.17)               .99  (.4 to 1.54)#                    Abortion Group                         43/242   (17.8)                      P = 0.25                           P = 0.79   Post 1979 Pregnancies only All Women: Delivery Group                        161/679 (23.7)                 .997 (.70 to 1.42)             1.24 (.80 to 1.92)†                   Abortion Group                          53/224 (23.7)                       P = 0.99                            P = 0.33              Unmarried Women: Delivery Group                        100/307 (32.6)                 .98 (.61 to 1.57)              1.27 (.68 to 2.36)&                    Abortion Group                          34/106 (32.1)                         P = 0.92                           P = 0.45            Married Women: Delivery Group                         60/370  (16.2)                 .95 (.54 to 1.68)              1.25 (.66 to 2.35)&                    Abortion Group                          18/116  (15.5)                        P = 0.99                           P = 0.49            Note. Pregnancy outcome coded "0" for delivery and "1" for abortion; higher CES-D scores indicate greater levels of depression. Odds ratios greater than 1.0 would thus indicate a higher risk for depression in the abortion group *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                              # Adjusted for race, age at first pregnancy, education, and family income          & Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, education, and family income  † Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, and 1992 marital status, education, and family income    Table 2b: Means, standard deviations, and unadjusted and adjusted betas of CES-D depression scores, stratified by marital status and year of pregnancy outcome (pre-1980/1980 and beyond)           OLS Regression predicting CES-D continuous scores                                                         Mean (SD)                        Unadjusted                 Adjusted                                                                                                         Beta (S.E)                     Beta (S. E)                                                                                    Full Sample All Women: Delivery Group                               11.7 (9.97)                           -.84 (.54)                      .51 (.63)*                  Abortion Group                              10.9 (9.96)                            P = 0.12                       P = 0.42   Unmarried Women: Delivery Group                             13.4 (10.5)                             -.15 (.84)                     1.97 (1.06)#                    Abortion Group                              13.2 (10.8)                            P = 0.86                       P = 0.06              Married Women: Delivery Group                              10.0 (9.1)                              -1.26 (.67)                   -.61 (.76)#                     Abortion Group                                 8.8 (8.6)                            P = 0.06                       P = 0.43                Post 1979 Pregnancies only   All Women: Delivery Group                               10.5 (9.0)                               .43 (.72)                   .84 (.82)† Abortion Group                            10.9 (10.6)                             P = 0.55                     P = 0.30   Unmarried Women: Delivery Group                            12.7 (9.9)                              1.04 (1.18)                 1.98 (1.50)&                    Abortion Group                           13.7 (12.0)                                P = 0.38                    P = 0.19    Married Women: Delivery Group                              8.7 (7.7)                                -.37 (.84)                    .43 (.93)&                    Abortion Group                               8.3 (8.4)                              P = 0.66                    P = 0.64     Note. Pregnancy outcome coded "0" for delivery and "1" for abortion; higher CES-D scores indicate greater levels of depression.  Positive beta coefficients would thus indicate a higher risk for depression in the abortion group. *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                              # Adjusted for race, age at first pregnancy, education, and family income          & Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, education, and family income  † Adjusted for 1979 measure of Rotter internal locus of control, race, age at first pregnancy, and 1992 marital status, education, and family income      Table 3: Relationship between depression and pregnancy outcome within sub-populations known to vary in amount of underreporting of abortion.                                                                                                                                                 Test                                                                        Delivery Group            Abortion Group               Statistic No (%) exceeding                                                                                                           (chi-square) CES-D cut-off    Married White Women (n = 624)                82/422 (19.4)               30/194 (15.5)               1.41(p = .24)   Married Black Women (n = 213)                 45/180 (25.0)                 8/31   (25.8)              .009 (p = .92)   Unmarried White Women (n = 134)           78/219 (35.6)                45/140 (32.1)               .46 (p = .50)   Unmarried Black Women (n = 288)         128/373 (34.3)                20/62  (32.3)                .10 (p = .75)   non-Catholics (n = 1193)                           266/890 (29.9)               78/292 (26.7)              1.07 (p = .30)   Catholics (n = 563)                                    95/388 (24.5)                37/169 (21.9)                .44 (p = .51)   Mean (SD) Continuous                                                                                                     (t-test) CES-D score   Married White Women (n = 624)                    9.2  (8.8)                     8.0 (7.9)                 1.71 (p = .09)   Married Black Women (n = 213)                   11.7 (9.4)                     11.5 (10.4)                .12 (p = .91)   Unmarried White Women (n = 134)              13.3 (11.2)                   13.3 (11.6)                .04 (p = .97)   Unmarried Black Women (n = 288)              13.2 (9.2)                     12.8 (9.4)                  .26 (p = .79)   non-Catholics (n = 1193)                                12.0  (10.0)                 11.5 (10.2)                .74 (p = .46)   Catholics (n = 563)                                        11.1  (9.8)                     9.8 (9.4)                1.38 (p = .17) Note. Pregnancy outcome coded "0" for delivery and "1" for abortion; higher CES-D scores indicate greater levels of depression.  Chi-square statistic compares CES-D cut-off by pregnancy outcome; t-tests compare CES-D continuous scores by pregnancy outcome.     Table 4: Other outcome variables                             Delivery              Abortion        t-test comparing             OR                     Beta                            Group                  Group                  groups                   (95% CI)            (S.E.)      Education      12.28 (2.02)        13.34 (2.34)              9.31**                         .83**                    -.83**                                                                                                                     (.79 to .88)              (.11)    Income             3.21 (2.11)         4.19 (2.44)               7.60**                         .79**                  -1.01**                                                                                                            (.74 to .84)             (.11)   Family Size     2.44 (1.14)         1.23 (1.17)             -19.56**                       1.11*                     .77**                                                                                                          (1.02 to 1.20)           (.19)    Note. *P < .05, ** P< .0001 Pregnancy outcome was coded "0" for delivery and "1" for abortion.  For the income variable, the mean of 3.15 corresponds to an income in the range of 20,001 to 30,000, and the mean of 4.15 corresponds to an income in the range of 30,001 to 40,000.  Positive t-values comparing education, income, and family size across groups indicate higher scores in the abortion group; negative t-values indicate lower scores in the abortion group.  Odds ratios less than 1 and negative beta coefficients indicate that greater income and education are associated with decreased depression; larger family size is associated with increased depression, as evidenced by the odds ratio greater than 1 and positive beta coefficient.   Table 5a: Logistic and OLS regression results if all women with subsequent abortions are excluded:                                               No. (%) exceeding               Unadjusted                     Adjusted                                                                       CES-D cut-off                    OR (95% CI)                 OR (95% CI)  Dichotomous CES-D: Delivery Group                        309/1100 (28.1)              .67 (.49 to .91)            1.06 (.74 to 1.53)*                       Abortion Group                          64/09 (20.7)                       P = 0.01                      P = 0.75                                                                                                                                                                                                                                                                                  Mean (SD)                  Unadjusted                       Adjusted                                                                                        Beta (S.E)                        Beta (S. E)      Continuous CES-D: Delivery Group                            11.6 (9.9)                     -1.81 (.62)                       .37 (.73)*              Abortion Group                             9.8 (9.1)                      P = 0.004                       P = 0.62        *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                                Table 5b: Logistic and OLS regression results if only women in the delivery group with subsequent abortions are excluded:                                                  No. (%) exceeding                 Unadjusted                   Adjusted                                                                           CES-D cut-off                  OR (95% CI)                 OR (95% CI)     Dichotomous CES-D: Delivery Group                      309/1100 (28.1)                .85 (.66 to 1.09)          1.13 (.83 to 1.55)*                    Abortion Group                      115/461   (25.0)               P = 0.22                       P = 0.44                                                                                                                                                                                                                                                                            Mean (SD)                     Unadjusted                Adjusted                                                                                         Beta (S.E)                  Beta (S. E)           Continuous CES-D: Delivery Group                            11.6 (9.9)                   -.74 (.55)                       .45 (.64)*                  Abortion Group                           10.9 (9.96)                 P = 0.18                        P = 0.49            *Adjusted for race, age at first pregnancy, and 1992 marital status, education, and family income                                                   In summary, even with a larger sample, based on coding of unwanted first pregnancy that is verified as accurate by NLSY staff,14 there is still no evidence that abortion of a first unwanted pregnancy increases risk for depression, and this lack of evidence is a consistent finding across groups that are known to vary in underreporting of abortion.   References [1] Schmiege SJ, Russo NF. Depression and unwanted first pregnancy: longitudinal cohort study. BMJ  2005;331:1303 (3 December), doi:10.1136/bmj.38623.532384.55 (first published 28 October 2005). [2] Reardon DC, Cougle JR. Depression and unintended pregnancy in the National Longitudinal Survey of Youth: a cohort study. British Medical Journal. 2002; 324:151-2. http://bmj.bmjjournals.com/cgi/reprint/324/7330/151 [3] Russo, NF, Schmiege, SJ Debates about our design are beside the point: The Reardon and Cougle findings are invalid and cannot be reproduced with properly coded data. BMJ.com 18 November 2005. [4] Pinto, FK., Biased sample, BMJ.com, 29 Oct 2005. [5] Leahy, P. Not very significant, BMJ.com 31 October 2005. [6] Rey, A. Re: Study Fails to Address Our Previous Findings and Subject to Misleading Interpretations, BMJ.com, 2 December 2005. [7] Gilchrist AC. et al, Termination of pregnancy and psychiatric morbidity, British Journal of Psychiatry 1995; 167:243-248. [8] Gerrard, J. Signing the blue form. 19 November 2005. [9] Major, B. Reardon Response to Russo and Schmiege Misleads by Omission and Commission, BMJ.com, 14 January 2006. [10] Reardon, DC, Cougle, JR, Rue, VM, Shuping, MW, Coleman, PK, & Ney, PG. Psychiatric admissions of low-income women following abortion and childbirth Canadian Medical Association Journal, 2003: 68 (10) http://www.cmaj.ca/cgi/content/full/168/10/1253 [11] Reardon DC, Ney PG., Scheuren F., Cougle J, Coleman PK, Strahan TW. Deaths associated with pregnancy outcome: a record linkage study of low income women. Southern Medical Journal, 2002: 95(8), 834-41. [12] Cougle JR, Reardon DC, &   PK. Depression associated with abortion and childbirth: a long-term analysis of the NLSY cohort, Medical Science Monitor, 2003: 9(4), 105-12. [13] We want to thank David Reardon for drawing the 1982 data to our attention. [14] We want to thank Canada Keck of the Center for Human Resources Research at Ohio State University for reviewing our coding and verifying its accuracy. Our coding used for these analyses follows:   Revised SAS codes: libname in 'g:\british'; OPTIONS SYSPARM='.'; OPTIONS NOCENTER; OPTIONS LS=80;   DATA a; SET in.nlsy_1110; rename r0769500=wantp182; rename R1017500=wantp183; rename R1325400=wantp184; rename R1522056=outp1_84; rename R1702800=wantp185; rename R1892758=outp1_85; rename R2002900=wantp186; rename R2259858=outp1_86; rename R2579200=wantp188; rename R2879700=outp1_88; rename R3188900=wantp190; rename R3409800=outp1_90; rename R3792500=wantp192; rename R4009468=outp1_92;   data b; set a; if R0214800=2; proc freq; tables wantp182 wantp183 wantp184 outp1_84 wantp185 outp1_85 wantp186 outp1_86 wantp188 outp1_88 wantp190 outp1_90 wantp192 outp1_92;   data c; set b; p1_out=.; if outp1_92>-4 then p1_out=outp1_92; else if outp1_90>-4 then p1_out=outp1_90; else if outp1_88>-4 then p1_out=outp1_88; else if outp1_86>-4 then p1_out=outp1_86; else if outp1_85>-4 then p1_out=outp1_85; else if outp1_84>-4 then p1_out=outp1_84;   wanted=.; if p1_out>-4 and p1_out<6 then do; if outp1_84>-4 and wantp182>=-3 then wanted=wantp182; else if outp1_84>-4 and wantp183>=-3 then wanted=wantp183; else if outp1_84>-4 and wantp184>=-3 then wanted=wantp184; else if outp1_85>-4 and wantp185>=-3 then wanted=wantp185; else if outp1_86>-4 and wantp186>=-3 then wanted=wantp186; else if outp1_88>-4 and wantp188>=-3 then wanted=wantp188; else if outp1_90>-4 and wantp190>=-3 then wanted=wantp190; else if outp1_92>-4 and wantp192>=-3 then wanted=wantp192; else if wanted=. and outp1_92 in (2 3 4 5) then wanted=999; end;   data d; set c; if wanted=3 or wanted=4 then do; if p1_out = 1 then pregout = 0; end; if wanted>2 then do; if p1_out=4 then pregout = 1; end;   proc format; value p1_out 1='live birth' 2='miscarriage' 3='stillborn' 4='abortion' 5='still pregnant';   value pregout 0='live birth' 1='abortion';   value wanted 1='wanted to be preg' 2="didn't matte" 3='no, not then' 4='no, never' 999='not asked, 1992'; run;   Competing interests: None declared
Obscuring the Suffering of Women and Compromising Science 25 March 2006
Priscilla K. Coleman, Associate Professor of Human Development and Family Studies Bowling Green State University, Bowling Green, OH 43402 Send response to journal: Re: Obscuring the Suffering of Women and Compromising Science	As an academic researcher with a great deal of experience studying and publishing on the psychology of abortion, I feel compelled to comment on a few critical and overlooked shortcomings of the Schmiege and Russo original presentation of findings and the supplemental results offered in the Rapid Response forum. The analyses presented in Table 3 of the article do not incorporate controls for variables identified as significant predictors of abortion (higher education and income and smaller family size). These associations between pregnancy outcome and depression are troubling since lower education and income and larger family size predicted depression (see Table 4). Without the controls, the delivery group, which is associated with lower education and income and larger families, will have more depression variance erroneously attributed to pregnancy resolution. Among the unmarried, white women, 30% of those in the abortion group had scores exceeding the clinical cut-off for depression, compared to 16% of the delivery group. Statistical significance is likely to have been achieved with the controls instituted. This group is important to focus on as unmarried, white women represent the segment of the U.S. population obtaining the majority of abortions [1]. Failure to convey the most scientifically defensible information is inexcusable when the data set contains the necessary variables. I strongly urge the authors to run these analyses. Curiously, in all the comparisons throughout the article, the authors neglect to control for family size without any explanation. I further recommend including women who aborted pregnancies deemed wanted at some point in the controlled analyses of white, unmarried women. Pregnancy wantedness is a nebulously defined variable, open to multiple subjective interpretations. Obviously the women who fell into the “wanted” category had mixed feelings with not wanting the pregnancy over-riding any wantedness, as they did in fact terminate their pregnancies. There is insufficient justification provided by the authors for excluding these cases. The authors state in one of their Rapid Response postings that removal of the 17 cases (29.4% exceeded the cut-off for clinical depression) from a sample of over 1200 did not change the results. Which results are they referring to exactly? Many different analyses were conducted. Why not provide the numerical evidence? The number of abortion cases in the various analyses ranged from a low of 33 to a high of 479, rendering the above statement focusing on the number in the full sample (abortion and delivery cases) misleading. Many of the 17 cases would likely fall into the unmarried, white sub-sample and the addition of even 8 cases to the sub-sample of 74 represents a nearly 11% increase. Finally, Major wrote a Rapid Response echoing Schmiege and Russo’s contention that the body of evidence does not support an association between abortion and mental health, a view that the American Psychological Association has promulgated despite the publication of numerous studies in top peer-reviewed journals indicating otherwise. Major, Russo and others have sternly criticized published findings demonstrating associations between abortion and negative psychological health for insufficient control of third variables (e.g., prior psychological health, substance use, exposure to violence, etc.) and for use of data likely to be compromised by high rates of concealment. Yet they are unwilling to apply the same methodological standards to their own work and the work of colleagues who publish politically acceptable findings. A group of New Zealand authors of a recently published, well- controlled prospective investigation revealing strong associations between abortion history and anxiety, depression, suicide ideation, and substance use boldly challenged the APA’s recent conclusion that “well-designed studies of psychological responses following abortion have consistently shown that risk of psychological harm is low” [2]. Fergusson and his colleagues noted that this conclusion was based on a small number of studies, which suffer from significant methodological deficiencies and a general disregard for studies showing negative effects. The authors concluded: “the present research raises the possibility that for some young women, exposure to abortion is a traumatic life event which increases longer-term susceptibility to common mental disorders.” This paper and several responses to it provide a disconcerting microcosm of the scholarly study of abortion and women’s mental health in the United States over the past several decades. Pursuit of a socio- political agenda seems to have taken precedence over genuine concern for women’s health, upholding principles of scientific integrity, and the advancement of knowledge. The best available empirical evidence indicates than a minimum of 10-20% of women experience severe, prolonged psychological problems associated with an abortion experience [3-7]. The emotional suffering of countless women for whom abortion was clearly not an adaptive choice can not be ethically ignored. [1] Jones, R.K., Darroch, J.E., & Henshaw, S. K. (2002). Patterns in the scoieconomic characteristics of women obtaining abortions in 2000- 2001. Perspectives on Sexual and Reproductive Health, 34 (5) September/October. [2] Fergusson, D. M., Horwood, J., & Ridder, E. M. (2006). Abortion in young women and subsequent mental health. Journal of Child Psychology and Psychiatry, 47, 16-24. [3] Bradshaw, Z., & Slade P. (2003). The effects of induced abortion on emotional experiences and relationships: A critical review of the literature. Clinical Psychology Review, 23, 929-958. [4] Coleman, P. K., Reardon, D.C., Strahan, T., & Cougle, J. (2005). The psychology of abortion: A review and suggestions for future research. Psychology & Health, 20, 237-271. [5] Lewis, W. J. (1997). Factors associated with post-abortion adjustment problems: Implications for triage. The Canadian Journal of Human Sexuality, 6, 9-17. [6] Major, B., & Cozzarelli, C. (1992). Psychological predictors of adjustment to abortion. Journal of Social Issues, 48, 121-142. [7] Zolese, G., & Blacker C. V. R. (1992). The psychological complications of therapeutic abortion. British Journal of Psychiatry, 160, 742-749. Competing interests: None declared