The Fragile Foundations of the American Psychological Association's Position on Abortion and Mental Health, Which Rests on a Single Study: A Critical Examination of the Methodology, Selection Bias, and Interpretations of Gilchrist et al. (1995)

Abstract A 1995 study by Gilchrist et al. (hereinafter Gilchrist) has been highly influential in concluding that abortion does not increase psychiatric morbidity compared to carrying an unplanned pregnancy to term. This finding was adopted as the central basis for the American Psychological Association’s (APA) official statements, adopted in 2008, regarding the safety of abortion from a mental health perspective. The APA has not revised that view or conducted a new literature review since 2008. Moreover, the APA has been clear that it has a political position on abortion as a basic civil right that is independent of any evidence regarding the risks or benefits of abortion to women’s mental health. Still, this political view may influence the APA’s willingness to recognize and acknowledge research demonstrating abortion’s contribution to any significant negative mental health effects. Indeed, the APA’s continued reliance on Gilchrist as evidence that abortion poses no significant risks to women is remarkable given (a) the large body of subsequent and better studies showing abortion is linked to higher rates of mental illnesses, and (b) the fact that Gilchrist was so fundamentally flawed in design and execution that it should never have been relied upon in the first place. This paper offers a complete examination of Gilchrist’s limitations, spanning fundamental design flaws, problematic measurement of outcomes, inadequate control for confounding variables, issues with attrition and generalizability and selective reporting. In short, this paper demonstrates that the methodological constraints of Gilchrist render its findings insufficient as evidence justifying any broad clinical or policy conclusions. Therefore, the position statements of the American Psychological Association, and other health organizations relying upon it, are not based on the best, or even reasonably good scientific evidence.