Predicting breast cancer risk: implications of a “weak” family history