Methods: A semi-Markov model was developed to evaluate CE of BRCA1/2 testing. We assumed that women are tested at age 35 and undergo prophylactic oophorectomy if they have a mutation. The model accrues costs and QALYs from a societal perspective annually until the woman is age 105 or dead. The estimates of health care costs, life expectancy, and patient preferences (using time-trade-off approach) for treatment and certainty about their BRCA1/2 status were based on the published literature. Costs (2004 dollars) and patient preferences were discounted at an annual rate of 3%. We examined confidence ellipses to determine the uncertainty of the incremental CE ratio.
Results: Using current guidelines, the average life-long costs and QALYs were $104,500 (SD=$18,600) and 23.56 (SD=5.85) for testing and $105,700 (SD=$21,600) and 23.59 (SD=6.18) for not testing. The cost savings is due to fewer women requiring treatment for breast/ovarian cancer after oophorectomy. The loss in QALYs for testing is due to the decreased utility of living with knowledge of the mutation and associated breast/ovarian cancer risk, even when otherwise healthy. If the utility of living with the knowledge of a BRCA1/2 mutation is increased from 0.68 to 0.70 or the utility gain of a negative test result is increased from 0.01 to 0.05, testing dominates (lower costs and higher benefits). The genetic testing strategy dominates when the probability of mutation is 19%, but not testing dominates when the probability is 3%.
Conclusions: The costs and effects for testing or not testing are very similar and are sensitive to the underlying probability of mutation, the disutility of living with knowledge of BRCA1/2 mutation, and the utility gain of receiving a negative test result. Therefore, patient preferences should be examined and considered when determining if testing is appropriate.