The latest edition of The New England Journal of Medicine carries an interesting discussion of the validity of the latest new health technology - whole genome testing. Based on multiple single-nucleotide polymorphisms (SNPs - pronounced "snips") - these commercially available tests purport to be able to tell you your risk of diseases such as type 2 diabetes, coronary artery disease and various cancers.
Of course they are expensive - in the USA you are probably looking at around $1000. But surely it's money well spent if the test tells you your risk of developing a particular disease? Not according to experts in genetic analysis, who have listed a series of reasons why these tests may not offer what they say they do.
Apparently, at the current state of the technology, these tests are neglecting certain critical, established aspects of the evaluation of the usefulness or otherwise of genetic testing. The first is whether or not the test is analytically valid. In other words, can it "accurately and reliably measure the genotype of interest". Most reports apparently suggest that the tests themselves are analytically valid, but that errors in handling the DNA sample may result in misclassification errors. However, even very small errors per SNP will be magnified across the genome and could result in hundreds of misclassified genetic varients for any individual. At the moment, it is unlikely that the "real-world" performance of these tests is up to standard.
The second important point is the clinical validity of these tests. This is the ability to detect or predict any associated disease or disorder. This is a particularly contentious area and apparently even the most enthusiastic proponents of genome testing admit that we are still in the very early stages of identifying the full list varients that are associated with susceptibility to disease. The problem is that most of the diseases that these tests claim to predict are not associated with a single gene, but with multiple genes - all acted on in different ways by lifestyle and other factors.
Last, but not least, there is the question of how clinically useful the test actually is. If the test does show increased risk of a disease what can be done about it? And what about the increased risk? In most cases, all a genetic susceptibility tells you is that you have an increased relative risk of a disease. Everyone has a certain baseline risk of any particular disease, such as type 2 diabetes. Generally, all finding factors that are associated with increased susceptibility to a disease tell you is that your relative risk of the disease may be increased. And if your baseline risk is very low, then this information does not tell you, in absolute terms, whether or not you will actually develop the diease.
What these tests are doing, in essence, is assuming that any intervention that is beneficial in the general population will be equally beneficial in an at-risk population. In fact many interventions - particularly the usual ones of stopping smoking, eating better, losing weight and getting regular exercise - will be beneficial in reducing the risk of many diseases, regardless of an individual's genetic risk of that disease.
So, what should you do as an individual, or indeed as a doctor faced with a patient who has spent all this money or is considering spending it? The conclusion reached in this article is that the money would be better spent on a personal trainer or gym membership.
See the full article here http://content.nejm.org/cgi/content/full/358/2/105
