Contrarian HIV Estimates in New Survey



A study of People Living with HIV (PLHIV) has proposed a contrarian view that the number of those living with this infection may be a third lower than current estimates of 5.2 million but warned that a nation-wide survey is required for a more accurate estimate. The National Reproductive and Child Health Survey’s population study of Guntur district in Andhra Pradesh looked at antenatal care patterns to extrapolate findings into indicative estimates of the HIV infected population numbers.

The researchers say that even a downward revision of numbers puts the numbers at a staggering 3.5 million and hotspots of infection with the imminent danger of explosive spread continue abated. They called for massive financial, technical, and training investments to scale services that will help prevent new infections and provide care and treatment to the infected millions.

India's HIV infection numbers are presently projected based on data collected each year by various State AIDS Control Societies and through HIV Sentinel Surveillance (HSS). Over a few months of each year, blood samples are collected from incoming patients at deemed ‘sentinel’ sites which could also be co-located in Ante-Natal Centers (ANC) treating pregnant women and departments treating those with sexually transmitted diseases or infections (STD or STI). This data is then aggregated and processed by the National AIDS Control Organization (NACO) to project prevalence estimates for the general population. Using HSS data in 2005, NACO’s model projected that Tamil Nadu, Andhra Pradesh, Karnataka, and Maharashtra together accounted for 3.7 million infections or about three fourths of present estimate of 5.2 million infections.

Differing from numbers projected by the NACO, Administrative Staff College Professor of Health Studies Lalit Dandona said that “population-based surveys like the Guntur study probably give us a picture closer to reality.” Dandona asserts that estimates of PLHIV in India to be “a gross overestimation.” Dandona says that the NACO model assumes that 5-6% of the general population are infected every year and “When the high HIV rates among STI patients attending large government hospitals are used for this assumption, the HIV estimates for the general population gets inflated.” He derides the reliance of STD data to project HIV infections as a “major distortion factor.” While the NACO used a self-selecting model, meaning sampled those who visited an ANC, the new study has adjusted downward the HIV burden based on deduction that ANC sites were picking up a higher number of HIV positive women since private practitioners do not treat HIV-positive patients. Since the only recourse for these women is to go to government hospitals the higher numbers that were wrongly projected for national prevalence using this skewed the data. Dandona says another aspect of this new study is that while the NACO model left out high-risk groups such as hostel inmates, prisoners, and military recruits, the new study has accommodated.

Hence, the Guntur study weighed numbers downward for ANC data and upwards for these other groups to extrapolate patterns and lowered estimated number of HIV infections in the four Southern States to 2 million and national numbers to 3.5 million. Acknowledging that the Guntur study was “well researched,” Director-General NACO Sujatha Rao declined further comment. Efforts to ascertain further comment from NACO was

Newspaper reports quoted John Hopkins School of Medicine Professor Robert Bollinger saying that “Comparing estimates from varying methods is like comparing apples and oranges.” Bollinger said that the only way India can get “more accurate estimates of where the epidemic is heading” is if it applies “similar estimation methods in a consistent way over time.”

More than the number of infected individuals, the Guntur study has exposed serious gaps in the nation’s estimation methodology for AIDS infections. It has also highlighted that current methods of using self-selecting model to project national numbers is incorrect and that a more proportional sampling of data from all risk groups is urgently required. More importantly, NACO and other non-government organizations in the area must create a monitoring system that has its finger on the direction of this epidemic. As repeatedly pointed out by Dandona, policy makers must not take this study as the final word nor should the nation be complacent in addressing this epidemic in a wholesome manner