Background:  At the individual level, HIV plasma viral load is linearly associated with HIV transmission. A recent cohort-level analysis demonstrated that viral load among injection drug users (IDU) predicts HIV incidence. In San Francisco, California, we characterized the overall mean community viral load and its spatial distribution in an effort to target community-level HIV prevention and access to care interventions to neighborhoods at greatest risk.

Methods:  We used San Francisco’s mature, mandatory, and accurate (>90% complete) laboratory reporting of HIV viral loads to calculate the overall San Francisco mean community viral load. We used Geographic Information Systems software ArcGIS v 9.1 to map mean community viral load by neighborhood to visually explore spatial differences in mean community viral load. We examined differences in the mean community viral load by various characteristics using the Kruskal-Wallis test.

Results:  The overall San Francisco mean community viral load was 20,563 copies ±81,793. As shown in the map, mean community viral load varied by neighborhood. Of the 5 neighborhoods, 4 with the highest mCVL (Tenderloin, South of Market, Bayview, and Visitacion Valley) have the lowest median incomes in San Francisco. Homeless mean community viral load was twice the San Francisco mean community viral load (43,818±103,492). The mean community viral load varied significantly (p <0.001) by demographic characteristics including race/ethnicity (African American, 31,849 copies ±123,239), and risk group—IDU 35,651±144,475; men who have sex with men (MSM)/IDU 29,634±77,700, transgender 31,298±77,030. There were also significant differences (p <0.001) by insurance status (public 30,617±129,150; private 15,129±73,753), engagement in care (seen every 6 months in the past year 16,988±69,894), and health status factors including log mean CD4 (114,513±151,536), and hepatitis C virus (HCV) co-infection (25,896±80,358.)

Conclusions:  Even in richly resourced San Francisco, the differences in mean community viral load are consistent with disparities in socioeconomic status and access to health care. Using mean community viral load to target and monitor structural and community-level interventions to improve health status and reduce HIV incidence merits exploration.