Housing as Healthcare
Abstract Format : The U.S. Housing Opportunities for Persons with AIDS (HOPWA) program provides housing assistance and related supportive services for low-income persons living with HIV and their families. The New York City (NYC) Department of Health and Mental Hygiene directly oversees 37 HOPWA contracts in NYC across three service categories: housing placement assistance, supportive permanent housing, and rental assistance. We evaluated CD4 count and viral load improvements after NYC HOPWA enrollment compared with matched controls.
Methods : During July 2014-December 2015, 574 HIV-positive adults newly enrolled in NYC HOPWA. To investigate the effect of housing enrollment on HIV health outcomes, we conducted a 1:2 matched cohort study. The control group consisted of 1,148 NYC persons living with HIV who never enrolled in HOPWA based on the NYC HIV surveillance registry and housing administrative databases, including eCOMPAS, the data system for NYC HOPWA. To avoid confounding and make groups as comparable as possible, to be able to attribute different outcomes to HOPWA enrollment alone, the control group was matched with HOPWA enrollees on as many of these factors as possible: age or age category, gender, race/ethnicity, area of birth, other housing program use, HIV transmission risk, area-level poverty, clinical status, and HIV diagnosis year or HIV diagnosis year category; Chi-square test was used to determine if baseline characteristics were balanced between groups after matching. Length of HOPWA enrollment was classified as short-term (<1 year) or long-term (≥1 year). All persons included in the analysis were alive through 2016, lived in NYC in 2014-2016, and were diagnosed HIV before 2013.
Last CD4 count (grouped as <200, 200-499, ≥500 cells/µL, or missing) and viral load were measured at 1 year pre- and post-enrollment per laboratory tests electronically reported to the registry. Conditional logistic regression measured if HOPWA consumers were more likely than matched controls to improve (e.g., from <200 or missing to 200-499) or maintain optimal (≥500) CD4 count. McNemar's test analyzed if the proportion virally suppressed (VL≤200 copies/mL) increased for each group.
Results : Baseline characteristics were balanced between groups after matching, indicating the matching was effective. Proportions in care post-enrollment for long-term and short-term HOPWA consumers were 99% and 98%, respectively. Compared to their respective non-HOPWA controls, the 284 long-term HOPWA consumers were 87% more likely (66% HOPWA vs. 51% non-HOPWA; 95% CI: 1.38-2.56), and the 300 short-term HOPWA consumers were 43% more likely (58% HOPWA vs. 50% non-HOPWA; 95% CI: 1.06-1.92), to improve or maintain an optimal CD4 count. Proportion virally suppressed among long-term HOPWA consumers increased from 79% pre-enrollment to 87% post-enrollment (p=0.003), while it increased 2 percentage points in their controls, from 79% to 81%. Short-term HOPWA consumers did not show significant improvement in proportion virally suppressed (75% pre-enrollment vs. 77% post-enrollment; p=0.54), while there was significant improvement among their controls (72% vs. 79%; p=0.02). By service category, enrollment length affected supportive permanent housing consumers most: proportion virally suppressed increased 12 percentage points for long-term supportive permanent housing consumers (71% vs. 83%; p=0.07) but decreased 7 percentage points for short-term consumers (77% vs. 70%; p=0.41).
Conclusions : Providing HOPWA housing services to persons living with HIV resulted in improved CD4 count and VL within 1 year relative to matched controls. Compared to short-term consumers, improvements in viral suppression and CD4 count were greater among long-term consumers for all service categories, especially for supportive permanent housing consumers who need more supportive services.