Marcel Slockers

66 Chapter 4 as compared with the general population for the period 2002–10. 17 A large register-based Danish nationwide study 11 showed more than 5–6 times higher mortality rates among homeless people than among the general population. Three other recent studies found rate ratios of 3.1 for homeless men and 2.5 for homeless women, 3 4.4 for both sexes combined 10 and 9.6 for homeless persons ages 25–44 years, 4.5 for ages 45–64 years and 1.1 for ages 65–84 years. 18 These studies confirm prior studies focusing on the 20th century. 1,2,4–9,12,13,17,19 showing that the high burden of mortality among homeless men and women is persistent and present in different regions, countries and subpopulations. The lack of reduction of the excess mortality was also reported in a recent study by Baggett, 18 who found that in 2003–08 the mortality rate of 18–64-year-old homeless in Boston had not changed as compared with 1988–93. Our study adds that mortality among homeless remains high, notwithstanding the presence of substantial policy efforts to improve their living conditions. In our study, short-term effects (i.e. within 5 years after policy changes) of local policies could not be demonstrated among the population who was already homeless in 2001. We obtained our results in a very comprehensive sample of homeless adults. We have not limited our sample to homeless people in particular circumstances, such as users of convalescence care, but included a wide range of services with partly different users and including those sleeping rough. We also have not limited our sample to services provided by the local authorities, but we also included services provided by the church. Our study is also comprehensive by including both men and women with a long follow-up period of about 10 years (average follow-up of 8.4 years due to mortality). But our study also has some limitations. First, homeless people who did not use any local service in 2001 were not included, and not all institutes could provide data for 2001. We cannot rule out that specific groups were more likely to be missed (e.g. homeless immigrants without a legal residence status could not be matched and may have faced worse living circumstances than their legal homeless peers). We had no data on whether homeless people in 2001 were still homeless during follow-up. But to assess whether there are effects of policy measures on mortality in a homeless cohort, it is appropriate to include these persons, even when they are no longer homeless, as this could have been the success of the policy efforts. Despite an ambitious programme to guide all homeless persons into temporary care and social rehabilitation and from there to clustered or individual supported housing projects, with social education training and guidance into jobs and psychiatric support, we found no significant differences in mortality before and after the year 2006. One possible explanation is that it takes some time for policy measures to have their effect and, for that purpose, we did an additional analysis with three periods: one before policy changes (<2003), one in a