A survey of the accuracy of applying the KC60 diagnostic code for genital herpes: first attack (C10A) was undertaken by non-consultant career-grade doctors working in genitourinary (GU) medicine in England, Wales and Northern Ireland. In total, 148 forms were returned (58% response). Overall, 88/148, 59.5% (95% confidence interval 51.5-67.4%) applied the correct codes for all the clinical scenarios presented. Of doctors who personally assigned codes to clinical episodes, 57/104 (54.8%), were correct compared with 31/44 (70.5%) who did not personally apply codes. There was no difference between these two groups (P=0.08). The main error was assigning code C10A on clinical grounds only when there was no laboratory confirmation of herpes simplex virus. The authors suggest that the KC60 C10A code for first attack genital herpes should be simplified to accept a clinical diagnosis rather than insisting on viral confirmation. Alternatively, consideration could also be given to adopting a more comprehensive system (such as the Scottish example) for first attack genital herpes. We believe that either option would help improve the accuracy of GU clinic data relating to genital herpes.