Objective: To assess the potential of antenatal surveillance data on HIV prevalence in young women as an indicator of trends in HIV incidence.
Design: Review of empirical data and discussion of problems encountered with surveillance systems, illustrated using cohort-component projection models.
Methods: Simple descriptive analyses are presented of prevalence and incidence data, with projection models used to explore aspects of the dynamic relationships between changes in HIV incidence and prevalence in young pregnant women for which empirical data are not yet available. Incidence changes due to change in risk among sexually active, and change in pattern of sexual debut are explored separately, and the resulting prevalence trends in pregnant women under age 25 years, and those expecting their first two births are described.
Results: HIV prevalence levels in young pregnant women categorized by age and by parity have different relationships to recent incidence levels. Age categorized prevalence data provide a reasonable indication of incidence under stable conditions, but may be very misleading if the age pattern of sexual debut changes. Prevalence levels categorized by parity are a reliable guide to incidence in the sexually active, but not necessarily to incidence in the population as a whole.
Conclusions: Ante-natal surveillance systems should categorize prevalence data by both age and parity to aid in the interpretation of underlying incidence levels.