In breast cancer, mutations of predisposition genes such as BRCA-1/2 and other genes as yet uncharacterised are manifest in up to 10% of cases. Although the prior probability of the presence of a breast cancer predisposing gene can be calculated for individual women, there is no published evidence to justify predicted risk as a selection criteria for screening. This study aims to define which patient groups with a significant family history should be screened, and whether clinical examination is necessary in addition to mammography. The Claus model was used to predict breast cancer risk in women with a family history. Women were divided into two groups according to their predicted risk: group I consisted of women at standard risk (lifetime risk less than 1:6) and group II with moderate/high risk (lifetime risk greater than or equal to 1:6). Women were cancer-free at the point of entry, and screening consisted of annual clinical examination and mammography from the age of 35 years. This study consisted of 1500 women in group I and 1078 in group II. The period of observation was 5902.0 and 4327.8 women years, respectively. A total of 31 cancers were detected, 12 in group I and 19 in group II. The median age at diagnosis in group II was 45 years (range 26-66 years) compared with 54.5 years (range 38-63 years) in group I (P=0.03). The relative risk of developing breast cancer in group II was 2.6 (95% confidence interval (CI) 1.2-5.8). When compared with breast cancer incidence in the normal population, the standardised incidence ratio in group II was significantly higher at 2.8 (95% CI: 1.7-4.2). The standardised incidence ratio of women in group I was similar to that of the general population (1.1 (95% CI: 0.6-1.8)). A total of 26/31 (84%) cancers detected were palpable, of which 14 (54%) were not visible on mammography. Approximately one-third of all palpable cancers were detected at routine follow-up. Mammography correctly identified 17/31 cancers (55%), but 29% of these were not palpable. Family history screening programmes are effective and women should be selected for screening according to predicted risk. The younger age of diagnosis in group II justifies screening from an earlier age using both annual clinical examination and mammography.