Invited commentary: How can we reconcile the findings of Keyes et al.'s study with the experience of our patients in clinical practice?

Although the accompanying study by Keyes et al. (Am J Epidemiol. 2013;178(9):1378-1388) shows us that women currently using hormonal contraception (HC) have better scores on the Center for Epidemiologic Studies Depression Scale and report fewer suicide attempts, it does not show us that HC protects women from mood disorders or that HC is free of the mood-related side effects which cause high rates of discontinuation. The groups compared in the Keyes et al. study were different in many ways; the women using HC were younger, were more likely to engage in positive health behaviors, and had lower depression scores at each prior interview. Women with mood disorders are more likely to avoid or discontinue HC and more likely to experience worsening mood while on HC. The negative mood-related side effects experienced by women using HC (irritability and lability) are not captured by a screening tool for clinical depression, such as the depression scale used in this study. The database used in this study was longitudinal and multiwave, so the authors could have compared changes in depressive symptoms among women who switched from hormonal to nonhormonal contraceptive methods (and vice versa) across different waves. Only if the same women experienced greater levels of depressive symptoms after discontinuing HC and fewer symptoms when they restarted HC could we conclude that HC may protect women from mood disorders.