When I started college in the early 1970's, prior to the Roe decision, some of my professors taught that motherhood was a barrier to women's success in life, and that abortion was necessary for women's empowerment. But this teaching was rooted in unproven theory, not research. In the forty-six years since then, research has not confirmed this theory.
A meta-analysis is a type of study design in which data from many studies are combined to provide a larger sample size with greater statistical power. This gives increased ability to identify an effect when the original studies were too small to detect an effect. It is considered a useful type of research for resolving controversies when smaller studies have given conflicting results (Rosner, 2011; Higgins & Green, 2011)
In a 2013 meta-analysis, Dr. David Fergusson and colleagues addressed the question: "Does abortion reduce the mental health risks of unwanted or unintended pregnancy?" The conclusion states: "There is no available evidence to suggest that abortion has therapeutic effects in reducing the mental health risks of unwanted or unintended pregnancy."
Echoing this, Dr. Paul Sullins (2016) wrote: "To date, although some studies have minimized the risk of distress following abortion, not a single study has documented mental health benefits for women from the practice of induced abortion."
Sullins' 2016 research was a 13-year longitudinal study - a study that involves repeated measurements of the same people over an extended period. His study attracted little media attention, but it is among the strongest and most important studies of abortion and mental health to date. The quality of the data set used is remarkable and is part of what makes this a truly important study.
The National Longitudinal Study of Adolescent to Adult Health (abbreviated "Add Health"), was started in 1994 with funding from The National Institute of Child Health and Human Development and 23 other U.S. government agencies and private foundations. (Harris, 2009) This project created a nationally representative sample of more than 20,000 U.S. adolescents in grades 7 through 12, and has continued to follow these students into adulthood.
Extensive data was obtained directly from students and also from their parents, siblings, school administrators, and romantic partners, using questionnaires, interviews, and medical tests (Harris, 2009). Existing data bases giving information about the community were also included. Thus, the Add Health data base provides the largest and most comprehensive study of the health-related behaviors of U.S. adolescents during the transition to adulthood. Add Health has become a resource for more than 30,000 researchers and has given rise to 7,500 publications to date. It is important to understand that this is an extremely high-quality data set which can be used to shed light on many different health issues.
Although the Add Health data set was not designed specifically for study of abortion and mental health, it does overcome some problems that have plagued research on abortion and mental health.
Enrollment in this study was school-based, with parents being told they could direct the school to keep their child out of the study - but if parents did not object, students were enrolled. Thus, there was a high initial participation rate and relatively few drop outs, with 80.5% of the available original sample completing follow up at seven years and at thirteen years (Sullins, 2016). This is in contrast to many abortion studies in which initial participation rates are often low, and dropout rates are high. That can be a problem because research with abortion patients and with more general samples has shown that people who are more distressed tend to decline to participate in research, and have a greater tendency to drop out (Broen et al., 2005; Weisaeth, 1989); if those who are most distressed are among the dropouts, these studies may give a false assurance that there is no problem.
It is also worth noting that Add Health used rigorous methods for assessing multiple mental health conditions repeatedly over time. For example, "several dozen specific questions" were asked about various types of substance abuse (Sullins, 2016), a 20-item rating scale was used for depression (the CES-D, by Radloff, 1977, cited by Sullins), and multiple ratings for anxiety were used including one 7-item scale. Most of the rating scales were consistent with DSM-IV diagnostic criteria for each disorder. (This is the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition of the American Psychiatric Association.
Dr. Sullins (2016) used the Add Health data to study reproductive events including live birth, abortion, involuntary pregnancy loss (miscarriage, stillbirth or ectopic pregnancy) or never pregnant. These reproductive events were studied in regard to risk of seven mental health conditions: depression, anxiety, suicidal ideation, alcohol abuse, drug abuse, cannabis abuse, and nicotine dependence. The data set provided complete mental health and reproductive health data for 8,005 women, who had been followed over a 13-year period of time. Because the Add Health data set had a wealth of information about each participant, it was possible to statistically control for 20 different potentially confounding factors, such as poverty, childhood abuse, intimate partner violence, and rape.
Sullins' (2016) results showed that having an abortion was consistently associated with increased risk of mental health disorders. Giving birth to a living child was consistently associated with decreased risk of mental health disorders. Involuntary pregnancy losses were not consistently associated with either increased or decreased risk of mental health problems.
Sullins gives several reasons for asserting that these results provide some of the strongest evidence to date that the association of abortion with psychological distress is not simply co-occurring, but is causal. The results were consistent and strong even when accounting for potential confounders. The data set and Sullins' methods assured that pregnancy events occurred before the measures of the mental health conditions by up to five years. The mental health risks increased with repeat abortions, indicating that the distress is likely due to the abortions and not some co- occurring or pre-existing condition. The fact that only abortion (and not other reproductive events) was associated with a consistent increased risk of mental health problems, also is consistent with causality. Although Sullins did not specifically discuss the Bradford Hill criteria for causality, his study does meet most of these established criteria (Hill, 1965; Shuping, 2016). Sullins' results were similar to longitudinal studies by Pederson in Norway and Fergusson in New Zealand.
Another study during 2016 attracted far more widespread media attention, one based on the Turnaway Study data. The Turnaway Study is a longitudinal study conducted by a team of researchers at the University of California at San Francisco. This study was designed to compare the various effects of having an abortion to the effects of being denied an abortion of an unwanted pregnancy - turned away due to waiting until the pregnancy had progressed beyond gestational limits of that clinic or that state. The Turnaway Study produced a data set that has been used to generate approximately three dozen publications (ANSIRH, 2015).
Patients were recruited at 30 abortion clinics in the U.S., by approaching and inviting 3,016 women. Of these, only 37.5% consented to participate, and after giving informed consent by telephone, 15% dropped out before the first interview (Biggs et al., 2017). The study involved telephone interviews every 6 months for 5 years, with each woman being offered a $15 department store gift certificate for participation in informed consent, and a $50 certificate for each interview (Dobkin et al., 2014). However, women continued to drop out over the five years of the study. With only 18% of the original 3,016 women remaining by the end of the study, when the authors make claims about the experience of the "majority" of women, we must remember that the 82% of the women were not present at the end of the study, and there is evidence from several sources that those who are more distressed are less likely to participate (Broen et al., 2005; Weisaeth, 1989).
It is also important to note that their assessments for mental health symptoms was extremely limited. To rate women's emotions at one week after either having the abortion or being turned away, the authors did not use validated rating scales using symptoms from the psychiatric diagnostic manual, but created two extremely simplistic rating scales measuring a total of six items (Rocca et al., 2013).
In a study to determine the incidence of PTSD after abortion, the four-item Primary Care PTSD Screen (Prins, Ouimette, & Kimerling 2003) was used (Biggs et al., 2013). The authors conceded the methods used did not allow them to pinpoint whether PTSD symptoms were due to the pregnancy, the abortion, or other factors (Biggs et al., 2016).
Many of the conclusions from the Turnaway data are opposite to other respected studies using good methodology. For example, the webpage for the Turnaway Study claims that women who are denied an abortion are "more likely to stay tethered to abusive partners," yet a 2014 meta-analysis of 74 studies by Hall and colleagues showed that intimate partner violence was associated with termination and with repeat termination. In one study "... women presenting for a third termination were more than two and a half times more likely to have a history of physical or sexual violence than women presenting for their first termination..... The researchers' findings support the concept that violence can lead to pregnancy and to subsequent termination of pregnancy, and that there may be a repetitive cycle of abuse and pregnancy."
The next time a publication from the Turnaway Study Data makes the headlines, read carefully.