The short answer is that the tests are as accurate as they practically can be at present.
The accuracy of genetic tests is mainly determined by two factors: analytical validity and clinical validity.
Analytical validity refers to how well the test predicts the presence or absence of a particular gene or genetic change. This is all about the test itself – if the result says you have a particular gene, for example, can you trust it?
Analytical validity is very high: you don’t get many genes disguising themselves as things that they’re not.
Clinical validity is the next stage – if you do have a certain gene, say, how strong is the evidence that the presence (or absence) of that gene is related to the factor (be that to do with diet, nutrition, fitness, or health) it’s thought to be related to.
Clinical validity varies according to the genetic marker in question. All links between genes and particular traits are probabilities. Researchers are able to see how frequently a genetic code is associated with a particular trait, and so express this frequency as a probability, e.g. 80% of people with this genetic code show this particular trait.
This is complicated by the facts that some genes are only ‘expressed’ or ‘turned on’ when coupled with certain lifestyle factors, and some traits seem to rely on a combination of multiple genetic markers, together with multiple lifestyle factors.
This is why it’s so important to combine genetic testing with additional information about your lifestyle and psychology to come up with personalised recommendations.