Examining the association of abortion history and current mental health: A rejoinder to and extension of Coleman, Coyle, Shuping, and Rue

Using the National Comorbidity Survey (NCS), Coleman, Coyle, Shuping and Rue (Journal of Psychiatric Research, 2009) found that compared to women who had never had an abortion, women who had had an abortion were at an increased risk of several anxiety, mood, and substance use disorders. Our paper has two purposes. First, we demonstrate that these results are not replicable. That is, using the same data, codes, and sample as indicated by the authors, we were not able to replicate the bivariate statistics (with no factors controlled) reported by Coleman et al. Furthermore, among women with prior pregnancies, we investigated whether having zero, one, or multiple abortions was associated with having mood, anxiety, or substance use disorders at the time of the interview. Abortion history was not significantly related to having a mood disorder, but in bivariate analyses, it was related to having an anxiety or substance use disorder, p < .05. However, when we controlled for prior mental health, violence experience, sociodemographic factors, and other pregnancy events, no significant relationship was found between abortion history and anxiety disorders. A significant relationship between multiple versus no abortions and substance use disorders remained (OR = 2.95, p < .05), likely because we were unable to control for other risk factors. These findings do not support policies and clinical practices based on the notion that abortion causes poor mental health. Instead, policies, practice, and research should focus on the impact that violence and prior mental health problems can have on health outcomes.