Tracing and treating contacts (partner notification, PN) is a crucial part of the control of sexually transmitted infections but few studies have quantified its outcome. A retrospective national study obtained information from 155 clinics on 1308 patients with gonorrhoea, 59% of all UK patients attending in January to March 1995. These gave 1887 contacts (1:1.5) of whom 621 attended, 75% being found to be infected. The ratio of contacts attending to index patients was 0.5 (621/1260). The ratio was not influenced by gender or ethnicity, but was affected by orientation, being 0.5 for contacts of male heterosexuals, 0.3 for contacts of male homosexuals (P <0.001). PN outcome was less good in metropolitan (0.3) than provincial clinics (0.5), suggesting that further research is necessary on the reasons for this difference. Reliable ongoing outcome audit will require good communication between clinics to verify that contacts have attended. The contact:index ratio is suitable for evaluation of PN as it relates to actual outcome rather than to the patient's account of contact numbers. Local work can be assessed against these national figures.