Abstract
BACKGROUND: People experiencing homelessness have an increased risk of mortality. The association between being at risk of homelessness and premature mortality is unclear. We aimed to determine all-cause and cause-specific mortality in patients who were homeless, at risk of homelessness (marginally housed), or housed. METHODS: This retrospective longitudinal cohort study compared mortality patterns in adult patients identified in 2003/04 by linking data from an Australian metropolitan emergency department to national mortality data. We used Cox proportional hazards models to estimate associations between housing status and mortality. To address competing risks, cause-specific hazards were modelled and transformed into stacked cumulative incidence functions. FINDINGS: Data from 6290 patients (homeless deceased = 382/1050, marginally housed deceased = 259/518, housed deceased = 1204/4722) found increased risk of mortality in homeless [hazard ratio (HR) = 4.0, 95% confidence interval (CI) = 2.0-3.3) and marginally housed (HR = 2.6, 95% CI = 3.4-4.8) patients. Homeless patients had an excess risk from external causes (HR = 6.1, 95% CI = 4.47-8.35), cardiovascular disease (HR = 4.9, 95% CI = 2.78-8.70) and cancer (HR = 1.5, 95% CI = 1.15-2.09). Marginally housed patients had increased risk from external causes (HR = 3.6, 95% CI = 2.36-5.40) and respiratory diseases (HR = 4.7, 95% CI = 1.82-12.05). Taking account of competing risk, marked inequality was observed, with homeless, marginally housed and housed patients having probabilities of death by 55 years of 0.2, 0.1 and 0.02, respectively. CONCLUSIONS: Mortality rates were elevated in patients who were homeless or at risk of homelessness. Increasing numbers of people are at risk of homelessness, and the effect of this on mortality is relatively unrecognized. Marginal housing may assuage some risk of premature mortality associated with homelessness; however, it is not equivalent to stable housing.