Background:   

People experiencing homelessness (PEH) stay in a range of shelter modalities, including tents, tiny homes, motels, and traditional communal shelters, which differ in their eligibility criteria, physical infrastructure, duration of stay, and support for residents. Results from studies that have explored disease prevalence in different shelter modalities have been mixed with some suggesting higher rates of conditions such as diabetes and heart disease among unsheltered PEH while others have found no differences or higher rates among those in shelters. Understanding disease prevalence in different shelter modalities has important implications for prioritizing healthcare resources.

Objective:

In partnership with the Homeless Healthcare Collaborative (HHC), UCLA Health’s mobile healthcare program for PEH in Los Angeles County, we aimed to study the relationship between shelter type and the prevalence of hypertension and diabetes. 

Methods:   

We conducted a cross-sectional study using electronic health record data. We included patients of all ages who were seen by HHC between January 2022 and April 2024. We grouped patients based on their living situation derived from the location of the HHC clinic encounter: unsheltered, communal shelter, individual shelter, or indeterminate shelter. These categorizations were made based on shelter descriptions from the Los Angeles County Homeless Authority Housing Inventory Count (HIC), internet search, or verification with HHC leadership. Communal shelters were characterized by communal living quarters and identified by the HIC housing type label “Site-based – single site.” Individual shelters were characterized by private living quarters and identified by the HIC housing type label “Tenant-based – scattered site.” The outcomes of interest were hypertension and diabetes. Hypertension was defined as an ICD-10 hypertension diagnosis, receipt of an anti-hypertensive medication, and/or blood pressure readings >140/90 mmHg on two distinct ambulatory encounters. Diabetes was defined as an ICD-10 diabetes diagnosis, receipt of a diabetes medication, Hgb A1c ≥ 6.5%, and/or fasting glucose ≥ 126. We fit multivariable logistic regression models adjusting for self-reported age, sex, race, and ethnicity to estimate the odds of hypertension and diabetes among unsheltered patients compared to communally and individually sheltered patients.  

Results:   

The study included 5192 unique patients. 2794 (53.8%) were unsheltered, 346 (6.7%) resided in communal shelters, 613 (11.8%) resided in individual shelters, and 1439 (27.7%) were in an indeterminate shelter. The mean age of participants was 51.3 years in communal shelter patients, and 48.4 years among unsheltered and individual shelter patients. Of the 5192 unique patients, 1104 (21.3%) were Hispanic, 952 (18.3%) were non-Hispanic Black and 869 (16.7%) were non-Hispanic White; 1980 (38.1%) had missing race and ethnicity data. 1310 (25.2%) patients had a diagnosis of hypertension and 473 (9.1%) had a diagnosis of diabetes. The highest prevalence of hypertension (37.9%) was observed among patients in communal shelters. The highest prevalence of diabetes (9.3%) was observed among patients in individual shelters. Based on multivariable logistic regression models, the odds of having hypertension were higher among patients living within communal shelters (AOR: 1.74, 95% CI: 1.27–2.37) and individual shelters (AOR: 3.02, 95% CI: 1.62–5.60) compared to unsheltered patients. Similarly, the odds of diabetes were higher among patients living in individual shelters (AOR: 1.28, 95% CI: 1.01–1.51) and communal shelters (AOR: 1.13, 95% CI: 1.04–1.23) compared to unsheltered patients.  

Conclusions:   

Patients residing within individual and communal shelters had significantly higher odds of having hypertension compared to unsheltered patients. To a lesser extent, the odds of having diabetes were higher among patients experiencing sheltered homelessness compared to unsheltered homelessness.