There has been a lot of news lately about the accuracy and reliability of direct to consumer (DTC) genetic testing kits. These are relatively inexpensive testing kits that use a sample of your DNA, obtained by scraping the inside of your cheek, to analyze for gene alterations and to determine your genetic predisposition for a disease or other trait like hair color or if you smell that strange smell that asparagus turns your pee! Basically, you send your DNA sample to a company and they tell you your “risk” of developing different diseases over your lifetime. The issue is that these tests are not very reliable. Unfortunately, while the kits themselves are cheap, the profiling can be quite pricey. So, you are paying for data that may be worthless.
A few months ago, the FDA tested these kits. They bought 10 kits from 4 companies and sent 2 samples from each person. One sample had factual data about age, race and gender but the second sample had fictional data supplied to test the accuracy of the findings.
The results of the test were surprising. First, some of the companies failed to disclose before the test was submitted that the African American or Asian samples could not be accurately analyzed. Secondly, the different samples yielded different results. Some samples did not align with current medical conditions, i.e. existence of heart disease. One sample scored below average using one kit for risk of prostate cancer and hypertension but above average on a different kit for the same disease. Why are they different? They should be the same.
These results raise important scientific questions about genetic testing and personalized medicine. Several genes that predict how a patient will respond to a certain therapy have been identified. For example, estrogen receptor alpha and progesterone receptor alpha expression for administration of tamoxifen in breast cancer, expression of certain mutations of BRCA1 or 2 increases the risk for developing breast cancer, expression of K-Ras in colon cancer identifying those likely to respond to cetuximab or panitumumab, expression of EGFR protein predicts use of Iressa, gefitinib or erltinib as a lung cancer therapy. The list goes on. Personalized medicine and the genes known to help predict disease or response to therapy is powerful stuff. It can, and will, really change how patients are tested and treated for disease. It is a young field and changing rapidly as more is being understood, but it isn’t ready for companies like these that provide information on your risk for diseases. The data they provide is incomplete or worse, inconclusive and wrong. The data isn’t reviewed by a healthcare provider and therefore has no medical oversight. Providing such services just muddies the water and prevents real progress in understanding the science of diseases.
There is active debate about regulating these companies with some proposing a complete ban on them for now. Whatever happens, I just hope that this doesn’t keep people from believing in science and the power of the new technologies and insights being realized today.