Forum: On the limited utility of KAP-style survey data in the practical epidemiology of AIDS.

Smith’s (1993) paper is very welcome in that it stimulates a much needed debate on raising research standards in the study of sexual networks and the transmission of HIV/AIDS. Nevertheless, his paper is difficult to grapple with in that he has several targets, including perhaps population-based epidemiology itself. At least one would derive that impression from his insistence that the focus should be on the person truly at risk, defined as the one whose partner is seropositive, in contrast to advocating estimates of what fraction of persons are likely to become infected if their partners practise a certain sexual lifestyle. In contrast, we are told that what should be employed is ‘practical epidemiology’, which seems little more than traditional (and praiseworthy) public health and medical practice devoted to patients, those they might infect, and other who might be infected by those from whom they received the disease. This comment will focus on three issues: the possibility of undertaking better population-based epidemiology in order to understand the epidemic; the extent to which the approach advocated by Smith can be said to lead to an understanding of the epidemic; and the public-health value of population-based research.