Risk factors associated with HIV in a population-based study in Andhra Pradesh state of India

Background: Population-based data on risk factors associated with HIV are not readily available from India. This understanding and an estimate of the impact of addressing behavioural factors on reducing HIV would be useful. Methods: We interviewed a population-based sample of 12 617 persons 15-49 years old from 66 rural and urban clusters in Guntur district in the south Indian state of Andhra Pradesh and tested their dried blood spots for HIV. We used multiple logistic regression to assess the association of risk factors with HIV and calculated population impact numbers for HIV reduction if behavioural factors were addressed. Results: Among men there was significant association between HIV and history of sex with men blood transfusion having ever visited sex worker or multiple lifetime women sex partners consuming alcohol before sex recreational drug use male non-circumcision and tattooing (odds ratios 5.74-1.97 P<0.03 R2=0.11). Among women the only identified behavioural factor associated with HIV was multiple lifetime men sex partners (P=0.001 R2=0.10). Taking into account the relative risk and prevalence of risk factors the highest impact on reducing the HIV number per unit population was for male circumcision. Conclusions: Among the identified factors male circumcision was estimated to have the highest relative impact on reducing HIV per unit population but the feasibility of this intervention in India needs further investigation. The low explanatory power in the regression models of the usually considered risk factors for HIV suggests that better understanding of HIV dynamics at the population level in India is needed.