This longitudinal study aimed to document (i) the information-giving and patient-communication styles of clinical geneticists and genetic counsellors (consultants) in familial breast cancer clinics and (ii) assess the effect of these styles on women's knowledge, whether their expectations were met, satisfaction, risk perception and psychological status. A total of 158 women from high-risk breast cancer families completed self-report questionnaires at 2 weeks preconsultation and 4 weeks postconsultation. The consultations were audiotaped, transcribed and coded. Multivariate logistic regressions showed that discussing prophylactic mastectomy (P ¼ 0.00) and oophorectomy (P ¼ 0.01) led to women having significantly more expectations met; discussing genetic testing significantly decreased anxiety (P ¼ 0.03) and facilitating understanding significantly decreased depression (P ¼ 0.05). Receiving a summary letter of the consultation significantly lowered anxiety (P ¼ 0.01) and significantly increased the accuracy of perceived risk (P ¼ 0.02). Women whose consultant used more supportive communications experienced significantly more anxiety about breast cancer at the 4 weeks follow-up (P ¼ 0.00). These women were not significantly more anxious before genetic counselling. In conclusion, this study found that consultants vary in the amount of information they give and the way they communicate; and this variation can result in better or worse psychosocial outcomes. Greater use of supportive and counselling communications appeared to increase anxiety about breast cancer. Identifying methods to assist consultants to address emotional issues effectively may be helpful. Outcome studies in genetic counselling have largely relied on women's subjective assessment of the content and process of counselling, with few including an objective measure of what actually happened during the consultation. Furthermore, outcomes are often restricted to measures of satisfaction and understanding of risk. Such studies create a rather simplistic view of the goals of genetic counselling (Emery et al, 1972;Kessler, 1981). Furthermore, many of these studies did not evaluate the role of patient's sociodemographic factors (Evers Kiebooms and Van den Berghe, 1979;Marteau and Richards, 1996). The samples comprised women being counselled for a wide array of conditions, including chromosomal, autosomal recessive and a small proportion of autosomal dominant conditions, did not adequately control for or measure counsellor variability and used ad hoc measures of patient's input and outcome variables (Michie, 1993).The current study describes the process of genetic counselling for women from familial breast cancer families as shown by the analysis of audiotapes of counselling consultations. It examines the effect of different consultant communication styles on a variety of outcomes.It was hypothesised that the more the consultant facilitated communication and the more information was provided, the better would be patient outcomes; (a) higher knowledge; (b) mor...