Background External and alcohol-related mortality is elevated postseparation, but the role of poor mental health in explaining this excess is unclear. We assess postseparation excess mortality by union type and over time since separation and examine how psychiatric morbidity present already before separation, during the separation process and after separation attenuates this excess.

Methods Using individual-level register data from 1995 to 2012, we followed 311 751 Finns in long-term unions. Psychiatric morbidity was identified from dates of prescription medication purchases and hospital admissions, separations from dates of moving out of joint households and mortality from the Death Register. Cox regression was used to analyse postseparation mortality controlling for psychiatric morbidity before, during and after separation.

Results External and alcohol-related excess mortality is most pronounced immediately after separation, particularly among men, and is much larger following marital than nonmarital separation. After sociodemographic factors are adjusted for, further adjustment for psychiatric morbidity attenuates the excess by about 25%. Psychiatric morbidity poorly explains alcohol-related postseparation excess mortality, but for suicide mortality, adjustment for psychiatric morbidity reduces the excess by about 40% among men and 50% among women. Among women, this is largely due to psychiatric morbidity present already before separation, whereas amongmen the attenuation is also due to psychiatric morbidity during the separation process and after it.

Conclusion Separation may exacerbate the problems of people already in poor mental health, and relationship dynamics should thus be considered during treatment. Particularly among men separation is a risk factor for suicide even without pre-existing mental health problems.