Insights into the rise in HIV infections, 2001 to 2008: A Bayesian synthesis of prevalence evidence

Objective: To estimate trends in prevalence of HIV infection, undiagnosed and total, among adults aged 15-44 years in England and Wales since 2001. Design: Multiple surveillance systems and survey data are available to inform different aspects of the HIV epidemic in England and Wales. To coherently and consistently combine this information to estimate trends in HIV prevalence, we apply a multiparameter evidence synthesis in a Bayesian statistical framework. Methods: The study population is stratified by exposure group and region of residence. We synthesize data from behavioural and community surveys, unlinked anonymous seroprevalence surveys, and an annual survey of individuals with diagnosed HIV infection. Prevalence estimates are given with 95% credible intervals. Results: The estimated number of prevalent HIV infections in 15-44-year-olds has increased from 32 400 (29 600-35 900) in 2001 to 54 500 (50 500-59 100) in 2008, corresponding to an estimated prevalence of 1.5 per 1000 (1.4-1.7) rising to 2.4 per 1000 (2.3-2.6) in 2008. A rise in prevalence of diagnosed infection contributes substantially to the increase. There is no evidence of a statistically significant decrease in the prevalence of undiagnosed infection. The proportion of infections that are diagnosed has therefore also increased. Conclusion: Although the increase in the proportion of infections that are diagnosed is encouraging, the rise in HIV prevalence and lack of evidence of a decrease in prevalence of undiagnosed infection suggest that diagnosis rates are not high enough to reduce the pool of individuals unaware of their infection and that new infections must be occurring.