Previous studies have suggested a positive effect of retirement on depressive symptoms. The present study took advantage of the large-scale, prospective Gaz et Electricité (GAZEL) cohort to examine whether personality could influence this effect.
Depressive symptoms were assessed in 1993, 1996, 1999, 2002, 2005, and 2008 with the Center for Epidemiologic Studies Depression Scale (CES-D). Among the participants for which changes in depressive symptoms after retirement could be computed, 9,755 had completed the Buss and Durkee Hostility Inventory and the Bortner Type A Rating Scale in 1993. Covariates included age, gender, occupational grade, history of sickness absences for depression, and alcohol consumption. The effect of hostility and type A personality on changes in depressive symptoms after retirement were assessed with general linear models.
Adjusting for all covariates, higher scores of total (p <0.001; η(2) = 0.017), cognitive (p <0.001; η(2) = 0.021), and behavioral hostility (p <0.001; η(2) = 0.004) as well as type A personality (p <0.001; η(2) = 0.002) were each associated with a smaller improvement of depressive symptoms after retirement. Regarding hostility subscales, only the association with cognitive hostility remained significant (p <0.001; η(2) = 0.018) when both were simultaneously entered in the model. Among participants meeting the CES-D threshold of clinical depression before retirement, those in the lowest quartile of cognitive hostility were two times more likely than those in the highest to fall short of this threshold after retirement (odds ratio: 1.99; 95% confidence interval: 1.54-2.58).
Individuals with high levels of cognitive hostility display less improvement of depressive symptoms after retirement.