Although the efficacy of mass screening for breast cancer has been established in Western countries, this strategy may be too costly for other countries with low incidence rates of breast cancer. We propose an alternative approach to screen female relatives of breast-cancer-index cases from hospitals, as part of the Taiwan multicentre cancer screening (TAMCAS) project. In order to assess the efficacy of this programme, and to estimate how often this high-risk group should be screened, we firstly elucidated the disease natural history from the pre-clinical screen-detectable phase (PCDP) by estimating the relevant parameters based on Markov chain models. We further predicted the proportion of interval cancers, advanced breast tumours and deaths from breast cancers by different screening frequencies. Results showed that the estimate of mean sojourn time (MST) for this high-risk group (1.9 years; 95% CI.1.18-4.86) is shorter than that for females from the general population. Analysis of a surrogate endpoint based on regional lymph-node spread and tumour size shows that annual screening for this high-risk group is likely to confer a significant 33% reduction in breast-cancer mortality compared with a non-significant 25 and 20% reduction for 2 yearly and 3-yearly screening regimes respectively. The above results suggest that a 1-year interval might be appropriate for this high-risk group. A simple cost-effectiveness analysis indicates a cost per year of life saved for mass screening ($72,480) 15 times that for the high-risk group ($4,851).