Pattern and levels of spending allocated to HIV prevention programs in low- and middle-income countries.

Background: AIDS continues to spread at an estimated 2.6 new million infections per year, making the prevention of HIV transmission a critical public health issue. The dramatic growth in global resources for AIDS has produced a steady scale-up in treatment and care that has not been equally matched by preventive services. This paper is a detailed analysis of how countries are choosing to spend these more limited prevention funds.

Methods: We analyzed prevention spending in 69 low- and middle-income countries with a variety of epidemic types, using data from national domestic spending reports. Spending information was from public and international sources and was analyzed based on the National AIDS Spending Assessment (NASA) methods and classifications.

Results: Overall, prevention received 21% of HIV resources compared to 53% of funding allocated to treatment and care. Prevention relies primarily on international donors, who accounted for 65% of all prevention resources and 93% of funding in low-income countries. For the subset of 53 countries that provided detailed spending information, we found that 60% of prevention resources were spent in five areas: communication for social and behavioral change (16%), voluntary counselling and testing (14%), prevention of mother-to-child transmission (13%), blood safety (10%) and condom programs (7%). Only 7% of funding was spent on most-at-risk populations and less than 1% on male circumcision. Spending patterns did not consistently reflect current evidence and the HIV specific transmission context of each country.

Conclusions: Despite recognition of its importance, countries are not allocating resources in ways that are likely to achieve the greatest impact on prevention across all epidemic types. Within prevention spending itself, a greater share of resources need to be matched with interventions that approximate the specific needs and drivers of each country's epidemic.