Why are these numbers more accurate?

Why are these numbers more accurate? The credibility of the new HIV prevalence figures is very high because they are derived from not one but three authoritative sources. a) An expanded and upgraded Sentinel Surveillance, spread over 1,122 sentinel surveillance sites and covering all districts in the country. This represents an increase of 400 sites since the last time, and an eight-fold increase from the 180 surveillance sites NACO began with in 1998. b) The National Family Health Survey or NFHS-3, a population- based survey conducted in 2005-06, with a sample size of over 100,000 people for HIV testing. c) An integrated behavioural and biologic assessment – the National Behavioural Surveillance Survey and the Integrated Biological Behavioural Assessments Survey. This is a targeted surveillance system focusing on high-risk groups in high-prevalence states. It may be pointed out that NFHS-3 is a household survey among the general population, and excludes high-risk groups such as sex workers, MSM, IDUs, truckers and so on. Seen in isolation, NFHS-3 suggests a prevalence figure of 0.28 percent. If the high-risk groups are also taken into account – as they have been by NACO – then the prevalence estimate rises. The new, composite methodology represents the most modern and accurate system in the world and, indeed, the Indian survey could now be the template for estimating HIV and AIDS figures in other countries. It was the result of protracted discussion and consultation with national and international agencies and experts. It incorporated the WHO/UNAIDS Workbook Approach, which consisted of two panels: one for high-risk groups and the other for the low-risk general population. The new estimation process was overseen by the Technical Resource Group on Surveillance, jointly chaired by the Director-Generals of NACO and ICMR. The new methodology allows for a “back calculation” of prevalence figures since 2002, based on the new assumptions and measures. The previous data for India offered a wide range – 3.4 million to 9.4 million – within which AIDS prevalence figures actually lay. The new estimation process has severely contracted that range, to between 2 and 3.1 million people. International validation of both the numbers and the process has been swift. This would indicate NACO’s estimation methodology would probably be replicated in other countries to gauge the spread of the HIV and AIDS epidemic there. As Paul Delay, Director of Evidence Monitoring and Policy at UNAIDS, Geneva, pointed out at a background briefing to contextualise the new Indian numbers, “India now really represents the state-of-the-art methodology, as far as combining different data sources is concerned. And now the analysis that is taking place will hopefully provide us with a most accurate picture that we have ever had on India.”