Spousal loss is a common experience associated with old age, and a major life-event that entails a period of intense suffering for most individuals. In addition to emotional shock and grief, the period after spousal loss is often characterized by vast changes in many areas of the surviving spouse’s life, from everyday habits to financial security. Accordingly, several studies have shown that those who have lost a spouse suffer from a greater risk of dying themselves, compared to their married counterparts. This thesis explores different aspects of the association between widowhood and mortality (often referred to as the widowhood effect) using data from Swedish population registers. The overarching aim is to outline how the strength of the association has changed over time and how it varies between different social and demographic groups, and thereby contribute to a more detailed understanding of the pathways between widowhood and mortality and the importance of different mediating factors. The widowhood effect may be linked to several psychological, social and material mechanisms and individuals may have different capacities to deal with stressors such as grief, social isolation and financial strain depending on factors like sex, age, socioeconomic status and their social environment. In addition, how a specific factor influences the association between widowhood and mortality may in turn often depend on a combination of other factors.

The empirical part of the thesis consists of three separate studies, focusing on i) the changing demography of widowhood, ii) the widowhood effect as it differs for men and women according to socioeconomic status and iii) period trends in the widowhood effect, with focus on the importance of education and duration of widowhood. The results show that the widowhood effect increased over the last five decades, especially among women and in younger age groups. During the same period, the relative number of women and younger individuals in the widowed population decreased, which partly offset the total observed effect of this trend. Higher socioeconomic status amplifies the widowhood effect for men, while the association for women remains ambiguous. Between the 1990s and the 2010s, the difference in excess mortality between those who had lost their spouse recently and those who had been widowed a longer period became somewhat larger. No systematic change in the association between socioeconomic status and excess mortality was evident during the same period.