Self-reported Physical Health of Women Who Did and Did Not Terminate Pregnancy After Seeking Abortion Services: A Cohort Study

It was incredibly disappointing that Annals would publish articles with "anonymous" funding. This would not be tolerated if any other author or funder sought to promote a given procedure or device, and represents incredible risks of bias. How can we be sure that the funder did not dictate terms of the research, choose the research team and influence this study. The usual assurances that funders had no role in design or implementation of the study is hardly sufficient if a funder remains anonymous. What if a study claiming the benefits from high dose opioid use was found to be anonymously funded by Pfizer? Would we take the authors’ assurances for granted?

This is unacceptable and indicates a political agenda that the Annals seems to be increasingly promoting as research.

I am appalled that Annals chose to publish such a flawed, and poorly designed study while hiding the source of its funding. Political agendas have no place on the pages of Annals' research study pages.

While I appreciate the challenges that this type of research presents and am impressed with the efforts the researchers made in this important topic, the Annals seems to have glossed over significant issues that limit their conclusions, including the low response rates (37%) and the fact that their results were actually based on 558 persons out of 3016 - 19% of the original cohort. This is far too low a number to try to make a) the conclusions and b) the political case in which the Annals seems to be invested.

The authors’ also comment that “One in 10 (9.4%) women who gave birth after being denied a wanted abortion reported a new diagnosis of gestational hypertension over 5 years; most of these cases (10 of 16) were reported for the index pregnancy. These women would have been able to avoid, or at least postpone, this outcome had they been able to obtain desired abortion care.” Gestational hypertension is something that occurs during pregnancy, usually after 20 weeks [1]. The other 6 could not have had gestational hypertension, or “postpone” it. The 6 valid cases results in about 5% incidence which is about the average, given the overall rate of GHTN of 6-8% [1].

Particularly with a very one-sided editorial accompanying the article, using these limited data to claim tongue-in-cheek that “health exceptions to abortion bans would apply, . . . well, always, ” suggests that Annals willingly allowed itself to cross the line from scientific journal to a “political science” blog (of note, this issue has a total of 2 research articles, the rest are reviews or opinion pieces).
1. American College of Obstetricians and Gynecologists. Task Force on Hypertension in Pregnancy, author. Hypertension in pregnancy / developed by the Task Force on Hypertension in Pregnancy. 2013. https://www.acog.org/~/media/Task%20Force%20and%20Work%20Group%20Reports/public/HypertensioninPregnancy.pdf

Dear Editor:

Any student of the social sciences should tell you that no reliable conclusions can be drawn from a self-selected minority of individuals (31%), especially when over half (55%) of that group drop out during the course of the study. Yet, this is just one of many problems with the Turnaway Study which, in its most recent incarnation, asserts that abortion has no measurable effects on women’s physical health.(1)

The self-selection bias inherent with the Turnaway Study’s design is further aggravated by the fact that women who expect to have the most trouble coping with an abortion are least likely to participate in follow-up interviews.(2) In short, since so many abortions are for women in emotionally charged circumstance, selection bias is even more pronounced than for other types of social science research.

Another problem with this case series data is that the Turnaway Study was specifically designed with a disproportionate focused on women seeking late terms abortions. As a result, its non-random results have literally nothing to tell us about the 90% of women seeking first trimester abortions.

Most important of all, Turnaway Study is methodologically weak in comparison to record linkage studies from Canada and the United Kingdom that abortion contributes to significant increase in subsequent on demands on medical care.(3,4) These findings are based on actual medical records, not a self-assessment of health. Moreover, another eleven record linkage studies, examining populations in Finland, Denmark, and the United States have shown increased risk of premature death (a good proxy for overall health) following pregnancy loss.(5) There is even evidence of a dose effect, increased risk of death upon exposure to multiple pregnancy losses, with mortality rates associated with abortion higher than those associated with natural losses, delivery, or no history of pregnancy.(5)

Obviously, these record linkage do not suffer any of the Turnaway Study’s problems with self-selection bias and self-assessments of physical health. Yet collectively, they strongly refute the Turnaway Study conclusions.

In short, research based on the Turnaway Study dataset is useless for drawing any conclusions regarding the general population of women having one or more first trimester abortions. Even if one attempted to narrow its conclusions to women seeking late term abortion, when those conclusions are then weighed in the context of more comprehensive reviews of the literature,(2,5) they still weigh very little.



REFERENCES

1. Ralph LJ, Schwarz EB, Grossman D, Foster DG. Self-reported Physical Health of Women Who Did and Did Not Terminate Pregnancy After Seeking Abortion Services: A Cohort Study. Ann Intern Med [Internet]. 2019; Available from: http://annals.org/article.aspx?doi=10.7326/M18-1666

2. Reardon DC. The abortion and mental health controversy: A comprehensive literature review of common ground agreements, disagreements, actionable recommendations, and research opportunities. SAGE Open Med [Internet]. 2018;6:205031211880762. Available from: http://journals.sagepub.com/doi/10.1177/2050312118807624

3. 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. South Med J [Internet]. 2002;95(8):834–41. Available from: http://www.scopus.com/inward/record.url?eid=2-s2.0-0036333844&partnerID=tZOtx3y1

4. Østbye T, Wenghofer EF, Woodward CA, Gold G, Craighead J. Health services utilization after induced abortions in Ontario: a comparison between community clinics and hospitals. Am J Med Qual [Internet]. 2001 Jan [cited 2015 Jul 17];16(3):99–106. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11392176

5. Reardon DC, Thorp JM. Pregnancy associated death in record linkage studies relative to delivery, termination of pregnancy, and natural losses: A systematic review with a narrative synthesis and meta-analysis. SAGE Open Med [Internet]. 2017 Dec 13 [cited 2018 Aug 15];5:205031211774049. Available from: http://journals.sagepub.com/doi/10.1177/2050312117740490

We appreciate this opportunity to respond to the comments raised about our article.

Two comments raise concern about the generalizability of our findings given the proportion of individuals approached that participated (38%). Enrolling participants in multi-year health studies is a universal challenge; just 24% agreed to participate in the widely-cited longitudinal Nurse’s Health Study II[1]. Reassuringly, the demographic profile of our study population closely mirrors U.S. women seeking abortion[2], ruling out the possibility of differential participation on the basis of measurable characteristics. And in terms of less often measured characteristics such as anticipated emotional reactions to abortion, our sample also closely resembles all people seeking abortion at one large clinic[3].

Two comments raise concern that our study results could be biased by loss to follow-up. This is true if the people lost over five years are systematically different from those who remained on factors related to our outcomes. However, there was no differential loss to follow-up by self-rated physical health, mental health status[4], or difficulty deciding about their abortion[5], minimizing concern about this type of bias.

The comment that our study reveals nothing about women obtaining first trimester abortions is incorrect. Over one-third (37%) of participants received a first trimester procedure. Of note, they reported better overall self-rated health over five years than those who gave birth after being denied the abortion care they sought.

One comment suggests that medical record studies are superior to ours. While medical records studies can be a rich source of data, they lack an appropriate group of women to compare to those having an abortion, and are therefore unable to provide an unbiased estimate of the effect of abortion on women’s physical health. Further, as not all illness individuals experience is reported to or diagnosed by a clinician, this study’s use of self-reported health measures represents a strength, rather than a weakness.

Finally, no funders played a role in the study’s design, conduct or interpretation of findings. Given the harassment and domestic terrorism that those working to ensure abortion access in the U.S. have been subjected to, our funders’ hesitation to be identified is understandable. As scientific evidence is increasingly needed to inform policy-making around abortion (as noted in the Supreme Court’s decision in Whole Woman’s Health vs. Hellerstedt), it is important that the National Institute of Health fund abortion-related research, rather than requiring researchers in this area to rely on private support.


[1] Brigham and Women's Hospital, Harvard Medical School, and the Harvard T.H. Chan School of Public Health. History of the Nurses Health Studies. Available at: https://www.nurseshealthstudy.org/about-nhs/history.
[2] Jerman J, Jones RK, Onda T. Characteristics of U.S. Abortion Patients in 2014 and Changes Since 2008. New York, NY. Guttmacher Institute; 2016. Available at: https://www.guttmacher.org/report/characteristics-us-abortion-patients-2014.
[3] Foster DG, Gould H, Taylor J, Weitz TA. Attitudes and decision making among women seeking abortions at one U.S. clinic. Perspect Sex Reprod Health. 2012;44:117-24.
[4] Biggs M, Upadhyay UD, McCulloch CE, Foster DG. Women’s mental health and well-being 5 years after receiving or being denied an abortion: A prospective, longitudinal cohort study. JAMA Psychiatry. 2017;74:169-78.
[5] Rocca CH, Kimport K, Roberts SCM, Gould H, Neuhaus J, Foster DG. Decision Rightness and Emotional Responses to Abortion in the United States: A Longitudinal Study. PLoS One. 2015;10:e0128832.

I am disappointed by a glaring oversight by the authors of this study.(1) In the discussion the authors state, “The results offer no evidence that having an abortion is detrimental to women's physical health. Instead, they indicate that for several dimensions of physical health, including overall self-rated health, women denied access to a wanted abortion reported worse long-term physical health than those who received abortion.” This statement ignores the fact that every pregnancy involves two patients, the woman and the baby, and half of babies are females. The authors completely ignore the fact that every abortion destroys the life of the baby. The destruction of the life of one of the patients is totally ignored in this study of the health affects of women having abortions.



1. Ralph LJ, Schwarz EB, Grossman D, et al. Self-reported physical health of women who did and did not terminate pregnancy after seeking abortion services. A cohort study. Ann Intern Med. 2019;171:238-47.

With great interest we read the article by Ralph and colleagues (1), and found their data insufficient to support the conclusion that physical health is no worse in women who sought and underwent abortion. We suggest that the interpretation of this study be qualified by taking into account greater, more robust data sets indicating that physical and non-physical health in women is adversely affected by abortion.

The Turnaway Study data are flawed by the low participation rate, with 37% originally consenting, but only 31% completing the first interview; the five-year dropout rate of 41%; and that from the original sample of eligible women, only 18.3% (588) remained at the end of the study (1).

Moreover, Ralph et al. state “women giving birth experienced a 14-fold higher risk for death” than those having abortions (1) but every record linkage study, eleven studies from three different countries, show the risk of death after abortion is higher (2). Increased risk is reported within 180 days, and remains increased for years. A dose effect is also reported with each abortion increasing the risk (2).

To understand the health effects of abortion, one must consider that seeking abortion is associated with intimate partner violence (IPV) (3). In one study, 25.7% of women seeking abortion experienced IPV during past 12 months, compared to 9.3% of women continuing pregnancy; this rate is nearly three times higher for women seeking abortion (4). Although Ralph et al. reports 12-month IPV data, it is not reported how women were screened, whether there were disproportionate numbers of IPV survivors among dropouts, or whether there was any differential in severity of abuse that may have impacted health and served as a confounder.

Further, a nationally representative cohort of 8005 young women was followed for 13 years with 81% completion, revealing there were significantly higher rates of suicidal ideation, alcohol abuse and dependence, illicit drug abuse and dependence, and cannabis use and dependence, in the abortion group compared to those who gave birth (5). These results are consistent with previous studies showing increased of substance abuse and suicide after abortion (2)

The evidence for the health of women seeking abortion appears contrary to the conclusion reached by Ralph and colleagues in their study (1); this appears to be, in part, due to use of a flawed data set.

1. Ralph LJ, Schwarz EB, Grossman D, Foster DG. Self-reported physical health of women who did and did not terminate pregnancy after seeking abortion services: A cohort study. Annals of Internal Medicine [Internet]. 2019; Available from: http://annals.org/article.aspx?dio_10.7326/M18-1666.

2. Reardon, DC, Thorp, JM. Pregnancy associated death in record linkage studies relative to delivery, termination of pregnancy, and natural losses: A systematic review with a narrative synthesis and meta-analysis. SAGE Open Medicine [Internet]. 2017 Nov 13;5:2050312117740490. doi: 10.1177/2050312117740490. Available from: https://www.ncbi.nlm.nih.gov/pubmed/29163945

3. American College of Obstetricians and Gynecologists. Reproductive and sexual coercion. Committee Opinion No. 554. Obstet Gynceol [Internet]. 2013:121:411-5. Available from: https://www.acog.org/Clinical-Guidance-and-Publications/Committee-Opinions/Committee-on-Health-Care-for-Underserved-Women/Reproductive-and-Sexual-Coercion

4. Bourassa D, Berube J. The prevalence of intimate partner violence among women and teenagers seeking abortion compared with those continuing pregnancy. Journal of Obstetrics & Gynaecology Canada. 2007;29:415–23. [Internet]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/17493373

5. Sullins, DP. Abortion, substance abuse and mental health in early adulthood. Thirteen-year longitudinal evidence from the United States. 2016. Sage Open Medicine [Internet]. Available from: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2813546

The original research article from the August 20 Issue by Ralph et al. reveals more about the political views of the Annals editors than it adds to the body of knowledge crucial to internists. Rather than “promoting and conducting research to enhance the quality of practice” and “unifying the many voices of internal medicine and its subspecialties,” the editors’ promotion of this study assumes that internists should endorse the availability of safe, legal abortions. In fact, the published supplemental editorial [Harris LH, Dalton V. Abortion and Women's Physical Health: An Issue for All Physicians. Ann Intern Med. [Epub ahead of print 11 June 2019]171:287–288. doi: 10.7326/M19-1740] asserts that “if we care about women's health, we must not only preserve abortion rights but also shift the conditions that lead to seeking abortion care.” That statement’s converse, that one does not care about women’s health if one does not actively promote abortion rights, is a highly offensive non sequitur: logically invalid, empirically unsupported, and morally inflammatory.
Never mind the selection bias inherent to the inconsistent observational measurements of their studied cohort (i.e., 841 women with 41% loss to follow up at five years) and the authors’ (three of the four) connections to The Bixby Center, whose mission statement states that they “work to ensure that all people have access to birth control, abortion” etc. That such a flimsy product of advocacy research could emerge without serious critique or counterweight demonstrates a blatant disregard of the diverse medical, ethical, and moral opinions of internists regarding reproductive health and abortion. Returning to the article’s thesis, however, does yield one particularly interesting logical outcome. Ralph et al.’s conclusions, reduced to the absurd, could reasonably lead one to promote abortions over childbirth for all women, given the reduced incidence of the study’s endpoints.
Beyond the low quality evidence and poor (not merely “uncertain”) generalizability of the study for internal medicine, the most disappointing aspect of this article is that the ACP would go to such lengths to assert their political beliefs. I consider Annals of Internal Medicine to be a primary source of educational material for my practice. This publication, however, fell well short of “advocating responsible positions on individual health and on public policy relating to health care for the benefit of the public.” In the future, I hope the editors can return to their own explicit ACP goals to serve the diverse body of internists in a truly responsible way.

Although Annals is less polluted by political,and leftish articles than NEJM, this article is unfortunately a clearly sponsored staff. The article attempts to view pregnancy as a disease. No mention of psyche, society, culture, religion, ethics. Shame Annals.