Evaluating AIDS Prevention Programs, Expanded Edition

The National Research Council Panel on the Evaluation of AIDS Interventions follows its 1st edition of recommended strategies for evaluating 3 major US Center for Disease Control (CDC) AIDS prevention programs with more recommendations. The panel emphasizes that prevention requires life long behavioral changes so program design implementation and evaluation should not be considered in the short term. It advocates research strategies (formative process and outcome evaluation) that entail major investments of talent time planning and money which even though are costly will make future evaluations easier and less costly. It is important the CDC clearly define program objectives in terms of measurable outcomes to test the effectiveness of programs. The panel stresses behavioral outcomes as the primary outcome variables for most AIDS intervention programs. Before launching a national AIDS media campaign or a new community bases organization (CBO) approach CDC should test a proposed campaign or CBO program in several randomly selected test market areas and compare these results with results from areas where the campaign was not aired or the CBO program was not operational. CDC should evaluate a selected sample of CBO projects and have all CBOs submit administrative reports to CDC periodically. The panel believes it is unethical to use no treatment control groups to evaluate the effects of HIV testing and counseling. Researchers should instead conduct randomized experiments with alternative treatment controls. Indeed randomized experiments for evaluating any AIDS prevention activity avoids selection bias and should be the backbone of any evaluation strategy. To avoid the difficulties of comparing data from any studies when the definitions specifications of variables and other factors differ the panel recommends subsets of common data elements be repeated across studies.