Use of observational data, including surveillance studies, for evaluating AIDS therapies

This article discusses some of the limitations of observational data for assessing therapeutic effects and categorizes observational studies according to the strength of evidence. Ecologic studies based on AIDS surveillance registries have demonstrated improvements over time in median survival following the onset of AIDS, especially for patients with an initial diagnosis of Pneumocystis carinii pneumonia. Some of this improvement may be the result of secular changes in the stage of disease at diagnosis or in the completeness of detection of deaths in these cohorts, however. Other ecologic studies of the use of AZT in various subgroups strongly suggest that treatment had a favorable impact on US AIDS incidence rates in gay men in the year beginning July 1987. This article is reprinted with minor changes from Chapter 24 of AIDS Clinical Trials.