In an article published in the Annals of Internal Medicine, researchers applied seven prediction models to 3,342 families to determine an individual family member’s probability of having a BRCA1/BRCA2 mutation. They then compared the results with actual genetic tests for the mutations.

They found that the seven models were similarly good at discriminating between people with a BRCA1/BRCA2 mutation and people without such mutations, but the models did make mistakes. The researchers found that predictions varied widely when the seven models were applied to one person, so genetic counselors may want to consider using several predictive models before recommending genetic testing. The genetic mutation tests are expensive, so clinicians use screening tests to try to determine if a person is likely to have the mutation before proceeding with the testing.

An accompanying editorial noted that all the prediction models require information from family medical history. It urged families to keep track of family members’ medical history and clinicians to ask about it.

SOURCES:
Annals of Internal Medicine, October 2, 2007
American College of Physicians (http://www.acponline.org)