• Amy R Wilson and James G Kahn.
• Division of Health Services Research and Policy, School of Public Health, University of Minnesota, Twin Cities, MN, USA. arwilson@umn.edu
• J Urban Health. 2003 Sep 1;80(3):465-81.

AbstractInjection drug users (IDUs) transmit the human immunodeficiency virus (HIV) via both needle sharing and sex. This analysis explores the effects of population risk behaviors, intervention effectiveness, intervention costs, and budget and capacity constraints when allocating funds between two prevention programs to maximize effectiveness. The two interventions, methadone maintenance and street outreach, address different types of risk behavior. We developed a model of the spread of HIV and divided IDUs into susceptible (uninfected) persons and infective persons and separately portrayed sex and injection risk. We simulated the epidemic in San Francisco, California, and New York City for periods from the mid-1980s to the mid-1990s and incorporated the behavioral effects of the two interventions. We used the simulation to find the allocation of a fixed budget to the two interventions that averted the greatest number of infections in the IDUs and their noninjecting sex partners. We assumed that interventions have increasing marginal costs. In the epidemic scenarios, our analysis found that the best allocation nearly always involves spending as much as possible on street outreach. This result is largely insensitive to variations in epidemic scenario, intervention efficacy, and cost. However, the absolute and relative benefits of the best allocation varied. In mid-1990s San Francisco, maximizing spending on outreach averted 3.5% of total HIV infections expected and 10 times the 0.3% from maximizing spending on treatment. In late 1980s New York City, the difference is five-fold (2.6% vs. 0.44%, respectively). Our analyses suggest that, even though prevention works better in higher risk scenarios, the choice of intervention mix is more important in the lower risk scenarios. Models and analyses such as those presented here may help decision makers adapt individual prevention programs to their own communities and to reallocate resources among programs to reflect the evolution of their own epidemics.

28 keywords...

hide…