1National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, 376 Victoria Street, Darlinghurst, NSW 2010, Australia 2School of Statistical Science
Estimates of recent trends in HIV incidence are crucial for formulating sensible strategies aimed at controlling HIV infection. Precise estimates of recent trends in HIV infection are unlikely if we use only AIDS incidence data, because there is little statistical information in AIDS incidence data about the number of individuals infected recently. The objective of this paper is to investigate the extent to which AIDS incidence data can reveal the recent trend in HIV infection, and to assess how well data on HIV diagnoses can improve the ability to detect such trends. An additive hazards model is proposed to utilise data on both HIV diagnoses and AIDS diagnoses. The study shows that, if only AIDS incidence data are used, we cannot confidently detect a trend in HIV infection in the recent five years for an epidemic like that in the USA. When HIV diagnosis information is incorporated, we may detect a trend of 5% annual decline in the most recent 3 years. Both HIV and AIDS diagnosis information should be used to detect recent trends in HIV incidence.