Comment www.thelancet.com Vol 372 August 9, 2008 423 “Know your epidemic, know your response”: a useful approach, if we get it right Led by UNAIDS, Know your epidemic, know your response has become a rallying cry for an intensified focus on HIV prevention, spurred by the sobering realisation that for every person enrolled on antiretroviral treatment, many more become newly infected. 1 The quest to better understand epidemics reflects growing recognition that there is no single global HIV epidemic, but rather a multitude of diverse epidemics. No single prescription can apply to countries as diverse as South Africa, Egypt, Russia, Thailand, or Papua New Guinea. The era of standard global prevention guidance is over. However, there is a globally useful distinction between concentrated and generalised epidemics, which are fundamentally different—not because of arbitrary pre- valence thresholds, but about who gets infected and how. Epidemics are concentrated if transmission occurs largely in defined vulnerable groups—typically sex workers, men who have sex with men, and injecting drug users, and their sexual partners—and if protecting them would protect wider society. Conversely, epidemics are generalised if transmission is sustained by sexual behaviour in the general population and would persist despite effective programmes for vulnerable groups. For too long, the global HIV-prevention community has pursued generalised responses in concentrated epidemics, concentrated approaches in generalised epidemics, or hedged their bets and done a bit of everything. 2–7 At the extremes, the differences between concen- trated and generalised epidemics are stark. Those in Latin America, the Middle East, Europe, and Asia—ie, most of the world—are and undoubtedly will remain concentrated, while most of southern and parts of eastern Africa are generalised. 2–9 Between these extremes, it is less clear whether some epidemics of the Caribbean, central and west Africa, and parts of the Pacific are concentrated, low-grade generalised, or mixed (table). While the global quest to know your epidemic is welcome, there are pitfalls to avoid. First, we must understand, but not overcomplicate. Broad rapid brush- strokes are sufficient for action. We can build our ships as we sail, guided by an overarching question: are our epidemics essentially concentrated, generalised, or (in some cases) substantially mixed? And where, in broad categories, are most new infections occurring: in sex workers, men who have sex with men, or injecting drug users (and their sexual partners), or through multiple concurrent partnerships in the general population? Second, although mathematical modelling of incident infections may be helpful, such models are in their infancy, make several major assumptions, and require better data than are generally available. Therefore the seductively precise graphs produced by such models must be carefully triangulated against other sources, including empirical risk-factor and incidence studies and rigorously grounded epidemiological syntheses. Published Online August 6, 2008 DOI:10.1016/S0140- 6736(08)60883-1 See Series page 475 Factor Concentrated epidemics Generalised epidemics Potentially mixed epidemics Geographic areas North, Central and South America, Europe, Middle East, Asia, Australasia Most of southern Africa and parts of east Africa Parts of the Caribbean, west Africa, horn of Africa, and the Pacific region Priorities needed for surveillance, monitoring, and evaluation Far greater emphasis on biological and behavioural surveillance of vulnerable groups (sex workers, men who have sex with men, injecting drug users) Antenatal and episodic population-based surveillance Both vulnerable group and antenatal/general population surveillance Analysis HIV prevalence, mapping, population-size estimation, behavioural interactions within vulnerable groups and between vulnerable groups, and their sexual or injecting partners Greater focus on understanding how to fundamentally change societal norms of sexual behaviour Greater focus on understanding transmission dynamics, including behavioural interactions between vulnerable groups and general population Investments Invest in surveillance, targeted interventions for vulnerable groups, and stigma-reduction campaigns for general population Investments should focus on promoting normative and social change to reduce multiple and concurrent partnerships, and to greatly increase availability of safe and affordable male circumcision services Investments should be matched to sources of transmission, which may vary across subnational regions and over time Interventions Goal is saturation coverage of vulnerable groups Goal is to help change community norms, values, and sexual behaviour at population level, and to frame male circumcision within broader rubric of male reproductive health and HIV prevention/behaviour change Goal is to relate interventions to transmission sources and maintain objective balance between targeted and general population activities Key research questions How to reach vulnerable groups with high coverage of high-quality targeted interventions How to change fundamental community norms and to de-norm multiple and concurrent partnerships at population level How to more accurately estimate relative proportion of infections from different transmission sources, and how to combine vulnerable group and general population interventions to reflect transmission patterns Table: HIV epidemic characteristics and priorities