Objective: To evaluate the feasibility, fidelity, and effect of a human immunodeficiency virus (HIV) risk reduction intervention delivered to HIV-infected patients by lay counsellors during routine HIV counselling and testing (HCT) public service in Mpumalanga, South Africa. Methods: A total of 488 HIV-infected patients, aged 18 years and older, receiving HCT service at clinics in Mpumalanga, received an "Options for Health" intervention delivered by clinic lay counsellors. Intervention lay counsellors implemented a brief risk reduction intervention at on average 2.3 sessions to help patients reduce their unprotected sexual behaviour. Questionnaires were administered at baseline and 4 months following the intervention to assess HIV risk behaviour in previous 3 months. Results: A significant increase in AIDS knowledge, behavioural intentions and risk reduction efficacy among participants at follow-up was found. Participants at 4-month follow-up reported a significant reduction of multiple partners, unprotected sex, alcohol or drug use in a sexual context, and transactional sex. In addition, sexual abstinence increased and alcohol use decreased, in the past 3 months. Conclusions: A lay counsellor-delivered HIV prevention intervention targeting HIV-infected patients seems to be feasible to implement with fidelity in the South African HCT setting and promising at reducing risk behaviour. (Contains 3 tables.)