Neighborhood level features have been found to influence a person’s ability to connect to a HIV care provider and remain in care. The objective of this analysis was to: 1) identify clusters of geographic areas with poor outcomes along the HIV care continuum and 2) assess person-level factors associated with residing in clusters among persons living with HIV (PLWH) enrolled in the DC Cohort study, a city-wide clinical cohort in Washington, DC.
Among PLWH with ≥ 1 year of follow-up in the DC Cohort, retention in care (RIC), prescribed antiretroviral therapy (ART), and viral suppression (VS) were estimated by residential ZIP code from June 2014 to June 2015. RIC was defined as evidence of ≥ 2 HIV-related encounters ≥90 days apart in 12 months. ART prescription was among those RIC. VS was defined as VL <200 copies/mL at last visit among those RIC and on ART. Clusters of adjacent ZIPs with similar outcomes were identified using Moran’s I and Getis-Ord Gi*. χ2 statistics were used to assess differences in person-level attributes by cluster status.
Among 4,413 PLWH, they resided in 20 ZIP codes. Median RIC was 71% (range: 41-79%) with clustering of low RIC ZIPs in the West (p<0.05). Median percentage on ART was 97% (range: 85-100%) with no clustering. Median VS was 89% (range: 75-100%) with clustering of low VS ZIPs and high VS ZIPs in the SE and NW, respectively (p<0.05). RIC in PLWH living in clusters of low RIC was 10% lower than in other ZIPs, though the difference was not significant (64% vs74%; P>0.05). This group was less likely to be non-Hispanic (NH) black (18% vs 83%), more likely to be employed (47% vs 19%), permanently housed (90% vs 78%), and privately insured (55% vs 20%). VS in PLWH living in clusters of high VS was 19% higher (100% vs 84%; P<0.05, Figure 1); this group was less likely to be NH Black (40% vs 83%), more likely to be employed (29% vs 20%), permanently housed (83% vs 78%), privately insured (37% vs 20%) and older (median: 57 vs 50 years).
Consistent with prior research, person-level and neighborhood-level retention may relate to yet not fully predict VS – even in the setting of high ART coverage. Clinically stable PLWH may see their HIV providers less often thus, while they may be virally suppressed, they may not meet current definitions of retention in care. Spatial analyses may inform the development of geographically targeted interventions to reduce drop offs along the continuum.