O1-S11.03 How many infections are averted by behaviour change after early HIV diagnosis & counselling of MSM? Estimates from a stochastic individual-based model

Background A recent paper (Fox et al HIV Medicine 2009) reported that MSM in the UK significantly reduced their transmission-risk behaviour following HIV diagnosis and suggested that this could be effective in averting transmission during the highly-infectious primary infection stage. However, cost-effectiveness analysis is required to inform policy-making. To assess the effectiveness of early HIV diagnosis in MSM as a prevention strategy we quantified its potential impact in terms of transmission HIV events averted. Methods We developed an individual-based stochastic transmission model to calculate the number of HIV-transmission events expected to occur from a cohort of recently-diagnosed MSM with and without the changes in behaviour that occurred post-diagnosis and counselling. The model incorporates different types of sex-act, patterns of condom use, and distinguishes between regular and casual partners. Results In the 12 weeks post-diagnosis, for a large majority of respondents there was a reduction in the expected number of casual partners who would be infected: 76% of participants eliminated risk of onward transmission entirely. However, a small proportion still presented a transmission risk. Overall, reductions in HIV transmission risk behaviour post-diagnosis would have reduced estimated secondary transmission during primary HIV infection (PHI) from been 33 (23–37) to 12 (6–14)—a reduction of 62% (32%–83%). Diagnosis after PHI produces a more modest reduction in transmission by missing the high-infectivity period. Conclusions Diagnosis of PHI can produce a large proportionate reduction in HIV-transmission events by reducing transmission-risk behaviour. Due to the high infectivity but short duration of primary infection, even short-term behaviour change can significantly reduce transmission. Later diagnosis is less effective, whilst early diagnosis requires frequent or highly-targeted testing. Whilst further work is required to determine the costs of different testing strategies, our quantification of the number of infections averted is an essential component of an assessment of the cost-effectiveness of strategies to increase early diagnoses of HIV infection.