Objective: Get insight into excess mortality in the Rotterdam homeless population according to cause of death and changes in this mortality after introduction of social policies in Rotterdam, the Netherlands, in 2006.
Design: Uncontrolled before-and-after study.
Method: We included homeless adults who visited the Rotterdam shelter in 2001 and followed them for 10 years (2001-2010). We then linked the data of this cohort to Statistics Netherlands mortality figures. We calculated the shares of specific mortality causes in total mortality for the entire study period. We used the 'standardised mortality ratio' (SMR) to compare mortality in the homeless cohort of this period with mortality figures of the general Rotterdam population. In order to be able to compare the homeless death rates according to cause of death in the period before (2001-2005) and after introduction of social policy measures (2006-2010), the hazard ratio (HR) was calculated.
Results: Our cohort consisted of 2130 homeless persons with a mean age of 40.3 years. The most important causes of death were unnatural death (26%; 95% CI: 21-32), cardiovascular diseases (22%; 17-27) and cancer (17%; 13-22). Suicide and murder together were responsible for 50% of the unnatural deaths. The largest differences in mortality in comparison with the Rotterdam population were those for unnatural death (SMR: 14.8; 95% CI: 11.5-18.7), infectious diseases (SMR: 10.0; 5.2-17.5) and psychiatric conditions (SMR: 7.7; 4.0-13.5). Mortality due to suicide or murder was significantly different for both study periods (HR: 0.45; 0.20-0.97).
Conclusion: Prevention of unnatural deaths among the homeless should be one of the priorities in homeless policy. Improvement of their living conditions may reduce the number of murders and suicides in this vulnerable group.