Allocating HIV drugs to cities prevents the greatest number of infections

The most effective way to control the AIDS pandemic in hard-hit South Africa would be to concentrate the allocation of scarce antiretroviral drugs in urban areas.

This, however, would not be the most ethical approach, according to an innovative new study from the UCLA AIDS Institute.

The article is scheduled to be published in the Proceedings of the National Academy of Sciences online Early Edition during the week of Sept. 11-15.

Using data from the KwaZulu-Natal province for their parameters, researchers from UCLA and the University of California, San Francisco, devised a mathematical model to predict the impact of drug allocation strategies that the South African government is implementing to treat 500,000 people by 2008. These data included birth rates, natural death rates and death rates stemming from AIDS.

They looked at three drug allocation strategies: one that would allocate antiretroviral drugs only to the city of Durban and two making them available in both urban and rural areas.

Of those, the Durban-only strategy would be the most effective in preventing new infections, reducing them by up to 46 percent -- amounting to preventing an additional 15,000 infections by 2008 -- compared with the two strategies that would include both urban and rural areas. The strategy also would avert the greatest number of deaths from AIDS and generate the least amount of drug resistance.

But major problems would emerge with that approach, said Sally Blower, professor at the Semel Institute for Neuroscience and Human Behavior at UCLA and senior author of the study. Most important, this approach is against basic ethical principles guiding treatment equity and would lead to more urban/rural healthcare disparities than already exist.

"If there was rational planning, you could determine drug allocation strategies by balancing ethical objectives with epidemiological objectives," said Blower, a member of the UCLA AIDS Institute. "But it's obviously unlikely that this type of rational planning would or could occur. So it's much more likely that the actual drug allocation strategy will be determined by a mix of politics and feasibility."

She added: "Unfortunately, you can't have the maximum impact on the epidemic and be ethical."

The methodology and results in the paper can also be very easily applied to other regions with scarce drug availability, said Dr. David Wilson, who served as the study's lead author as a postdoctoral fellow in Blower's lab and is now at the University of New South Wales in Australia.

"If policymakers in KwaZulu-Natal -- as well as other resource-constrained regions -- can rationally plan drug allocation, then modeling like we have done can inform these authorities of likely consequences of different allocation strategies," he said.

The researchers found that under the Durban-only strategy:

• Transmission would fall 25 percent to 46 percent in Durban but by less than 5 percent in rural areas.
• Transmitted resistance to antiretroviral drugs would increase 0.4 percent to 5.5 percent in Durban but would not emerge in rural areas.
• Death rates from AIDS would fall by a median 42 percent in Durban but by only 0.1 percent in rural areas.

Under the rural/urban drug sharing strategy, the results on transmission, drug resistance and death rates would be similar. For instance:

• Transmission would decrease by 11 percent to 28 percent in Durban and by 17 percent to 37 percent in rural areas.
• Resistance would hover between 0.1 percent and 3 percent in the city and 0.2 percent to 4.5 percent in rural areas.
• AIDS-related deaths would fall a median 26 percent in Durban and 34 percent in rural communities.

This was the first time that city and village models had been used to evaluate the epidemiological impact of treatment strategies and of the consequences resulting from various drug allocation strategies, as well as to predict the evolution of drug resistance and of the possible directions that the HIV epidemic can take in KwaZulu-Natal, the researchers wrote. Also, they said they were not aware of any other model combining both a dynamic epidemiological model and an allocation model for scarce resources to evaluate ethical decision-making.

Dr. James Kahn, professor of medicine at UCSF and co-author of the study, noted that clinical decision-making to help limit the epidemic involves hard choices.

"The best outcomes would be immediate and widespread use of anti-HIV medications and the systems to implement the bold vision, but that is not possible," Kahn said. "Thus the modeling can be used to help guide the immediate roll-out plans. But all of these models require us to focus on bringing effective medications and the needed infrastructure to help stem the epidemic as quickly as possible."