What Counselors Tell Low-Risk Clients About HIV Test Performance

In 1998, Gigerenzer et al. studied how heterosexual men with low-risk behavior were counseled about the accuracy of HIV test results. Most professional counselors conveyed the illusions that false positives do not occur and that a positive HIV test result means that the client is certainly infected. To help improve counseling quality, the authors provided feedback to all counseling centers in Germany. Sixteen years later we assessed whether HIV counseling in Germany has improved by replicating the original study with an expanded sample of 32 randomly selected counseling centers across the country. Since the original research, the positive predictive value (PPV) of HIV testing for a low-risk client has improved from about 50% to 96%. Hence, among every 26 low-risk clients who test positive, we can expect that one is actually not infected. Whereas test performance has improved over the last 16 years, counseling has not. About half of professional counselors communicated the illusion of certainty for sensitivity (15 of 30), specificity (16 of 30), and the PPV (18 of 30). Only one of the 30 counselors could correctly state the PPV. In what follows, we explain how to improve counselors' and clients' understanding of the PPV by representing the information in terms of natural frequencies rather than conditional probabilities. Doing so has been shown to improve the quality of counseling in different medical settings and may enhance future HIV counseling as well.