No one can dispute the scientific promise of genomic technologies. However, it is important that discoveries leading to specific testing applications in clinical or population health scenarios should be subjected to rigorous scientific evaluation, like any other scientific breakthroughs (Khoury et al. 2007). In a recent editorial Evans and Khoury (2007) discussed several reasons why genomic medicine has been slow to embrace the principles of evidence-based medicine. First, genetics has by and large focused on rare genetic diseases for which there are an inadequate number of individuals and families to study using randomized clinical trials or large observational studies. Second, genetics has focused on nondirective approaches to communicating information about genetic risks, mostly for highly penetrant conditions for which there may or may not be effective interventions. Third, the rapid advances in genomics makes difficult the conduct and update of evidence-based guidelines, and challenges traditional systematic review methods. Fourth, the concept of “clinical utility” in genetics has been variably defined and measured (Grosse et al. 2006). Overall clinical utility reflects the balance between benefits and harms, whether using the traditional focus on improved health outcomes for individual tested or considering other potential benefits for family members or information for the sake of information (“knowledge is power”).
Although many applications look biologically promising, recent systematic reviews of the available evidence have been rather disappointing. For example, as part of an evidence-based review, researchers from Duke University reviewed the cumulated evidence regarding whether testing for CYP450 polymorphisms in patients with depression treated with selective serotonin reuptake inhibitors (SSRIs) leads to improved outcomes, and whether test results are useful in decision making (AHRQ 2007). The evidence showed that data on the association between CYP450 genotypes and the metabolism, effectiveness, and side effects of SSRIs in the treatment of depression were mostly derived from heterogeneous studies with small samples. They did not find data on whether CYP450 testing in adults entering SSRI treatment for depression leads to improved clinical outcomes. They also found limitations in the quality of evidence that need to be considered in designing future studies of the validity and utility of CYP450 testing in the treatment of depression with SSRIs (AHRQ 2007). These findings prompted the independent Evaluation of Genomic Applications in Practice and Prevention (EGAPP) Working Group to issue its recommendation of “insufficient evidence” for the integration of this test into routine practice, discouraging its use until further research can close the evidence gap (EGAPP 2007).
The recent availability of testing for one million genetic variants is an extension of some previously promoted genomic profiles [e.g., cardiogenomic or osteogenomic profiles (Haga et al. 2003)]. In 2006, a Government Accountability Office (GAO 2006) investigation of such practices in the United States found major errors, discrepancies and misleading information provided to consumers on web sites offering genomic profiles directly to consumers. Genetic variants with weak or modest effects (odds ratios from 1 to 1.5 that genome-wide association studies are finding for genetic variants associated with common diseases) have limited added value in the prediction and prevention of common diseases unless specific effective interventions can be offered based on such information (Haga et al. 2003). For example, Janssens et al. assessed genetic profiles for risk of type 2 diabetes and showed that weak genetic effects have probably little added value in predicting future disease compared to more conventional tools such as body mass index, family history and other factors (2005, 2006).
In the United States, the US Preventive Services Task Force (USPSTF 2007), a well established independent US body that develops evidence based practice guidelines for primary care, has examined only two genetic topics between 2001 and 2006. The first is BRCA1 testing in breast and ovarian cancer (USPSTF 2005) and the second is screening for HFE mutations to identify individuals at risk for hereditary hemochromatosis in the general population (Whitlock et al. 2006). These two topics were chosen about 10 years after the genes for BRCA1 and hemochromatosis were discovered in 1994 and 1996, respectively. For BRCA1, many years after such tests made their way into practice, the task force found sufficient evidence for a subset of women with the appropriate family history for referral to genetic counseling for decision-making about the possible use of the tests (USPSTF 2005). For HFE testing, the task force found sufficient evidence (among others, the uncertain natural history and low penetrance of HFE mutations) to recommend against screening in the general population (Whitlock et al. 2006). A major obstacle to the USPSTF decision-making for both BRCA1 and HFE testing was impeded by the slow accumulation of scientific evidence on clinical utility for testing for these conditions.