Hoptel equalizes length of stay for homeless and domiciled inpatients

Abstract

Background: There is growing evidence that homeless individuals have longer inpatient lengths of stay with significantly higher medical costs than domiciled individuals. We compared adjusted mean lengths of stay among domiciled patients and homeless patients discharged to a hospital hotel (hoptel). Because the hoptel allowed homeless patients to be discharged when medically indicated, we hypothesized no significant differences in lengths of stay between the 2 groups after adjustment for severity of illness.

Methods: Demographic, diagnosis, and length of stay data were obtained for all patients discharged from all inpatient wards at a large, urban Veterans Affairs Medical Center during the first 32 months of hoptel operation (1996 through 1998). Adjusted mean lengths of stay for domiciled patients and homeless patients discharged to the hoptel were compared. ANCOVA was used to adjust for demographic characteristics (age, income, race/ethnicity), diagnosis-related group (DRG), principal diagnosis, and substance abuse comorbidity. Three types of hospitalizations were analyzed on the basis of major DRG categories: general medical, psychiatric, and all hospitalizations.

Results: No significant differences were found in adjusted mean lengths of stay between domiciled patients and homeless patients discharged to the hoptel. Homeless hoptel patients stayed an average 0.7, 0.3, and 0.4 days longer than domiciled patients for all, general medical, and psychiatric hospitalizations (P = 0.504, 0.627, and 0.839), respectively.

Conclusions: These data suggest that hoptels may help equalize inpatient lengths of stay among domiciled and homeless patients discharged to a hoptel. Public health care systems seeking innovative ways of reducing inpatient lengths of stay among homeless patients ought to consider establishing hoptels.