Estimating Incidence of HIV Infection Among Men Who Have Sex with Men, San Francisco, 2004–2014

After 30 years of the HIV epidemic in San Francisco there is hope that the number of new infections among men who have sex with men (MSM) is decreasing and that current novel interventions such as treatment as prevention and pre-exposure prophylaxis will hasten the year that the city sees the last of new HIV infections. In addition, new HIV cases/ incidence is the key indicator to measure the trajectory of the HIV epidemic. In this analysis we present an alternate age-cohort approach to estimating HIV incidence and compare our results to other indicators of incidence. Data for the present analysis were collected through National HIV Behavioral Surveillance conducted among MSM in San Francisco using time location sampling from 2004 to 2014. We estimated HIV incidence using a model where a closed population of 100 was divided into number infected and uninfected according to the HIV prevalence of the 21-25 year group and then estimated what incidence over 30 years would result in the HIV prevalence at age 50?. Incidence estimates were 7 per 1000 person years (PY) (338 cases), 7 per 1000 PY (312), 6 per 1000 PY (285) and 6 per 1000 PY (271) for 2004, 2008, 2011 and 2014, respectively. Conclusion: Our data suggest that recent declines in new HIV diagnoses among MSM in San Francisco maybe due to a reduction in a "back log" of undiagnosed cases and not as large of a decline in new cases or HIV incidence. We hypothesize that the decline in new HIV infections among MSM in San Francisco is much slower than suggested by the decline in new HIV diagnoses.