Selection for genetic testing for pathogenic mutations in BRCA1 and BRCA2 is an important area of healthcare. While testing costs for mutational analysis are falling, costs of tests in North America remain in excess of $3,000. Most countries state that there should be at least a 10-20% likelihood of detecting a mutation in BRCA1 or BRCA2 within a family before mutational analysis is performed. A number of computer-based models have been developed to assess this likelihood, and these continue to be improved to incorporate mutation frequencies, breast cancer incidence and tumour histology. However, these can be time-consuming and difficult to use in a busy clinic. The Manchester scoring system was developed in 2003, and we have continued to validate its use in Western populations. The scoring system discriminates well at both the 10 and 20% threshold for testing and compares very well with more complex computer-based models. However, it should not be used in its current form in founder populations or populations with low incidence of breast cancer, although a lower points threshold could be used to determine an appropriate cut off. The development of the Manchester score and its comparison with other models will be described in this chapter.