Explaining excess morbidity amongst homeless shelter users: A multivariate analysis for the Danish adult population

Abstract

Aims: This article analyses excess morbidity amongst homeless shelter users compared to the general Danish population. The study provides an extensive control for confounding and investigates to what extent excess morbidity is explained by homelessness or other risk factors. Methods: Data set includes administrative micro-data for 4,068,926 Danes who were 23 years or older on 1 January 2007. Nationwide data on shelter use identified 14,730 individuals as shelter users from 2002 to 2006. Somatic diseases were measured from 2007 to 2011 through diagnosis data from hospital discharges. The risk of somatic diseases amongst shelter users was analysed through a multivariate model that decomposed the total effect into a direct effect and indirect effects mediated by other risk factors. Results: The excess morbidity associated with shelter use is substantially lower than in studies that did not include an extensive control. Approximately 80% of excess morbidity amongst shelter users is attributed to other risk factors. A large part of the excess morbidity is explained by substance abuse problems and lack of employment, whilst mental illness, low income, low education, civil status and ethnic minority background explain only a limited part. However, when conducting an extensive control for confounding, a significantly higher morbidity was identified amongst shelter users for infectious diseases, lung, skin, blood and digestive diseases, injuries, and poisoning. Conclusions: Ill health amongst homeless shelter users is widely explained by substance abuse problems and other risk factors. Nonetheless, for many diseases homelessness poses an additional risk to the health.