Objective: To assess whether routinely collected data from partner notification for HIV infection could be used to study HIV epidemiology. The issues addressed were measures of contact patterns and behaviour change, variables influencing transmission risks, and indications of HIV incidence.

Design: We collected anonymous questionnaire data from all partner notifications performed from seropositive patients diagnosed in Sweden between 1 January 1989 and 30 June 1990.

Method: A structured questionnaire was completed by the physician or counsellor interviewing newly diagnosed seropositive patients and counselling their reported partners. The questions focused on temporal and behavioural aspects of all contacts between index patients and partners.

Results: Questionnaires were completed for 365 of the 403 (91%) index patients diagnosed during the study period, for 350 of the 390 (90%) located partners, and for 274 of the 297 (92%) relationships where results of HIV testing were known for index patient and partner. Seropositive individuals diagnosed in 1989 or later reported less risk behaviour than those diagnosed earlier. Risk of transmission in sexual contact increased when the infectious partner developed symptoms of HIV infection. Anal intercourse was found to be approximately twice as infectious as vaginal, and transmission risk from a seropositive insertive partner approximately twice as high as from a receptive. The total HIV incidence in Sweden appears to be declining, as does the number of newly diagnosed infected homosexual men.

Conclusion: Carefully collected data acquired from a partner notification programme are well suited to describe and follow the epidemiology of HIV infection.