|
HIV incidence in sub-Saharan Africa: a systematic review with implications for surveillance and prevention planning
S.L. Braunstein1, J.H.H.M. van de Wijgert2, D. Nash3
1Columbia University - Mailman School of Public Health, Epidemiology, New York, United States, 2Academic Medical Center of the University of Amsterdam, Center for Poverty-related Communicable Diseases, Amsterdam, Netherlands, 3Columbia University - Mailman School of Public Health, International Center for AIDS Care and Treatment Programs, New York, United States
Background: Population-based HIV incidence estimation and risk factor identification is increasingly being incorporated into routine HIV/AIDS surveillance activities. The US recently produced a national HIV incidence estimate of 0.022 per 100 person-years [PY] using the serologic testing algorithm for recent HIV seroconversion (STARHS) method via BED capture ELISA. We conducted a systematic review of estimates and risk factors for HIV incidence from sub-Saharan Africa, including the methods by which they were obtained.
Methods: We examined peer-reviewed articles, conference proceedings and technical reports published between 1987-2008. Incidence estimates were classified by country, year, population group, and estimation method (prospective study or STARHS).
Results: Our search yielded HIV incidence estimates for 15 of 44 sub-Saharan African countries, including 57 studies, for a total of 264 unique HIV incidence estimates. Of these, 239 (91%) were obtained via prospective studies (range: 0.6-16.5 infections/100PY), and 25 (9%) via the STARHS method (24 using BED) (range:1.8-17.0/100PY). Only 5 countries (Kenya, South Africa, Tanzania, Uganda, and Zimbabwe) reported national population-based estimates (range: 0.6-9.0/100 PY). STARHS use increased over time, comprising 20% of estimates since 2006. However, African studies that compared STARHS estimates with prospectively observed or modeled estimates often found concerning levels of disagreement. There was a high degree of correspondence between risk factors for HIV incidence versus prevalence in studies where both were measured.
Conclusions: Population-based HIV incidence estimates and associated risk factor information in sub-Saharan Africa remain scant but increasingly available, and are 1-2 orders of magnitude higher than the most recent national estimate in the US. However, data from some studies using the STARHS method in sub-Saharan Africa suggest a need for further validation prior to widespread scale-up. In the meantime, prevalence and behavioral risk factor data remain an important mainstay for HIV prevention planning.
|