PREVALENCE AND DETECTION OF HIV RISK BEHAVIOR IN PRIMARY-CARE - IMPLICATIONS FOR CLINICAL PREVENTIVE SERVICES

Despite their potential to reduce the incidence of HIV infection throu gh primary prevention, family physicians report low levels of routine identification of patients at risk and counseling. This may reflect pe rceptions that few of their patients are at risk, that patients at ris k will self-disclose during consultations, or that a physician-initiat ed approach is unacceptable to patients presenting for non-HIV-related problems. Our aim was to determine the prevalence of risk factors for HIV infection and HIV testing among patients in general practice, the acceptability to patients of opportunistic identification of risk dur ing routine consultations and the accuracy of general practitioners' a ssessment of HIV risk. Our setting included randomly selected general practitioners' surgeries in metropolitan Sydney, Australia. We conduct ed a self-administered survey about risk factors in a consecutive samp le of patients 18-50 years of age and compared it to a checklist about patient's risk factors completed by general practitioners unaware of the patients' answers. Of 1,030 patients, 43 (4%) were at risk of HIV infection having received blood transfusions between 1980 and 1985 and seven (1%) had injected intravenous drugs in the previous 12 months. In the previous 12 months, 133 (21%) female and 110 (28%) male patient s had been in nonmutually monogamous heterosexual relationships. Only 26% always used condoms. While the majority of patients indicated they were heterosexual, 42 (4%) were homosexual and 23 (2%) bisexual. Of t hose men who had had sex with other men, 39% always used condoms. Of a ll patients, 303 (29%) had had a previous HIV test: 2 (< 1%) tests wer e positive although 10 (3%) of those who had been tested did not know their results. Questions about specific HIV risk factors were acceptab le although 12% would mind if their doctor asked about number of sexua l partners. Using 1,021 matching checklists, we found that the specifi city of physicians' judgment of sexual preference was high but sensiti vity was poor. Physicians detected 38% at most of those at risk. The p revalence of risk factors for HIV infection is not insubstantial. Phys icians do not judge HIV risk accurately. Their detection of risk needs to be reconsidered in the light of these findings. A physician-initia ted approach to risk assessment appears acceptable to patients althoug h some questions were less acceptable than others. We recommend a rigo rous trial of risk detection and counseling in family practice to eval uate its cost-effectiveness in changing risk behavior and to monitor a ny potential for negative impact on patient satisfaction.