We report an approach for BRCA1/2 testing whereby genetic testing can be offered to families at high risk of hereditary breast and ovarian cancer but where no DNA from affected relatives is available. By testing two or more unaffected relatives at 50% risk of being heterozygous for a potential BRCA1/2 mutation, there is a chance of up to 99% of finding a mutation that would have been detectable in an affected individual from the same family. The overall likelihood of identifying a mutation is dependent on the family history, and therefore 'indirect' testing would be most applicable for families with a very high risk of carrying a BRCA1/2 mutation. Using this approach also requires balancing issues of testing resource limitations, family dynamics and adequate preparation of unaffected persons for a positive test, with the advantages of targeting screening and prophylactic surgery.