This is a pretty important topic.

I've had one of the tests done. I've read about what's available, but I've also read a lot of the research these tests are based on.

There is one test, which I've had, which reveals which genetic variants you have for liver enzymes which break down your meds. For example, I've had the test, and I'm a CYP2D6 poor metabolizer. About 10% of people of European descent have this variant. A LOT of psych meds are broken down by CYP2D6; as you might expect, I often end up on very low doses of psych meds, because my liver clears them slowly and they build up in my system.

The implications of liver function in prescribing are pretty obvious. Knowing the relative speed at which you will clear drugs will inform decisions on dosage, drug interactions, and so forth. So that particular test is very useful - IF your prescriber understands what is actually being tested (my pdoc doesn't like how the results are presented) and if they are fully capable of using that information.

There are other tests, for example, the 5-HTTLPR variant. The research these are based on is preliminary at best. Little if any is backed by solid purpose-built studies of real world outcomes for patients with the various genetic variants. It's mostly smaller and exploratory research at this point, in the form of "small study suggests that related meds A and B don't work very well if you have the variant." The research to replicate these results and explore them further hasn't been completed yet in many cases.

My pdoc, whom I respect a great deal, does not like how the tests are marketed and how the information is presented. She says essentially that the tests are framed as being far more definitive than they should be based on the quality of the research, and also that they're framing the results as recommending for or against various meds, when often the research is often much more nuanced.

This is the future of medicine, but it's not fully baked yet.