BACKGROUND: A consulting method known as 'shared decision making' (SDM) has been described and operationalized in terms of several 'competences'. One of these competences concerns the discussion of the risks and benefits of treatment or care options-'risk communication'. Few data exist on clinicians' ability to acquire skills and implement the competences of SDM or risk communication in consultations with patients. OBJECTIVE: The aims of this study were to evaluate the effects of skill development workshops for SDM and the use of risk communication aids on the process of consultations. METHODS: A cluster randomized trial with crossover was carried out with the participation of 20 recently qualified GPs in urban and rural general practices in Gwent, South Wales. A total of 747 patients with known atrial fibrillation, prostatism, menorrhagia or menopausal symptoms were invited to a consultation to review their condition or treatments. Half the consultations were randomly selected for audio-taping, of which 352 patients attended and were audio-taped successfully. After baseline, participating doctors were randomized to receive training in (i) SDM skills or (ii) the use of simple risk communication aids, using simulated patients. The alternative training was then provided for the final study phase. Patients were allocated randomly to a consultation during baseline or intervention 1 (SDM or risk communication aids) or intervention 2 phases. A randomly selected half of the consultations were audio-taped from each phase. Raters (independent, trained and blinded to study phase) assessed the audio-tapes using a validated scale to assess levels of patient involvement (OPTION: observing patient involvement), and to analyse the nature of risk information discussed. Clinicians completed questionnaires after each consultation, assessing perceived clinician-patient agreement and level of patient involvement in decisions. Multilevel modelling was carried out with the OPTION score as the dependent variable, and rater, consultation and clinician levels of data, standardized by rater within clinician. RESULTS: Following each of the interventions, the clinicians significantly increased their involvement of patients in decision making (OPTION score increased by 10.6 following risk communication training [95% confidence interval (CI) 7.9 -13.3; P < 0.001] and by 12.9 after SDM skill development (95% CI 10 -15.8, P < 0.001), a moderate effect size. The level of involvement achieved by the risk communication aids was significantly increased by the subsequent introduction of the skill development workshops (7.7 increase in OPTION score, 95% CI 3.4-12; P < 0.001). The alternative sequence (skills followed by risk communication aids) did not achieve this effect. The use of most risk information formats increased after the provision of specific risk communication aids (P < 0.001). Clinicians using the risk communication tools perceived significantly higher patient and clinician agreement on treatment (P < 0.001), patient satisfaction with information (P < 0.01), clinician satisfaction with decision (P < 0.01) and general overall satisfaction with the consultation (P < 0.001) than those who were exposed to SDM skill development workshops. CONCLUSIONS: These clinicians were able to acquire the skills to implement SDM competences and to use risk communication aids. Each intervention provided independent effects. Further progress towards greater patient involvement in health care decision making is possible, and skill development in this area should be incorporated into postgraduate professional development programmes.