Pharmacogenetic Testing: What It Can and Can't Tell You

If you’ve tried a medication or two that didn’t help, or that caused side effects you couldn’t live with, you may have heard about pharmacogenetic testing. It’s often described as a gene test that shows which medications will work for you. The real picture is more useful, and more limited, than that.

Pharmacogenetic testing looks at how your genes may affect how you process certain medications. This post covers what it measures, what the research does and doesn’t show, what it can’t tell you, and what to know about cost and coverage. For our full review of the research, see our pharmacogenetic testing page.

What Pharmacogenetic Testing Looks At

Many psychiatric medications are broken down by enzymes in the liver. Genes such as CYP2D6 and CYP2C19 carry the instructions for some of those enzymes, and variation in these genes influences how the body breaks down many antidepressants. Other genes on some panels, such as HLA-B, are linked to rare but serious reactions to specific medications.

The results usually sort you into groups for each gene, often called metabolizer types. In plain words, they describe how quickly your body may process medications that depend on that enzyme.

• Poor or intermediate metabolizer: your body may break some medications down more slowly, so levels can run higher.
• Normal metabolizer: your body processes them at the expected rate.
• Rapid or ultrarapid metabolizer: your body may break some medications down faster, so levels can run lower. Not every gene has both groups: CYP2C19 results can be rapid or ultrarapid, while CYP2D6 results have only an ultrarapid group.

For some medications, the direction flips. A few medicines have to be turned into their active form by these enzymes first, so slower processing can mean less of the active medicine, not more.

These groups come from the way the Clinical Pharmacogenetics Implementation Consortium (CPIC), a group that writes guidelines for using gene results in prescribing, describes them. The sample can be blood, saliva or cells swabbed from your cheek, and in practice it’s often a quick cheek swab. Results can take from a few days to a few weeks.

Do results last? The variants these tests look for are inherited, and inherited variants are present throughout a person’s life, so the genes behind your results don’t change over time. Two things can still change. How you actually process a medication can also depend on age, diet and other medications, and the guidance for reading results is updated as research grows; CPIC’s 2023 antidepressant guideline, for example, updated its 2015 version.

Where the Evidence Is Strong, and Where It’s Debated

Whether this testing helps people get better is debated among researchers. Part of the science is well supported, though, and it helps to keep the two apart.

Well supported: some genes clearly change how some medications are processed. The FDA keeps a table of pharmacogenetic associations that includes a number of psychiatric medications, and CPIC publishes gene-by-gene guidelines.

Mixed: large U.S. trials of test-guided depression care have found modest, mixed results.

• GUIDED (2019). The GUIDED trial enrolled people whose depression hadn’t responded well to at least one antidepressant. Its main measure, symptom improvement at week 8, showed no significant difference. Response and remission, which were secondary measures, favored testing. The trial was sponsored by the company that makes the test.
• PRIME Care (2022). In the VA’s PRIME Care trial, testing changed which medications were prescribed. Remission was modestly more common across the 24 weeks, but at week 24 itself the difference wasn’t statistically significant.
• ADOPT PGx (2026). This NIH-funded trial of children and adults used just two genes, CYP2D6 and CYP2C19, read with CPIC’s public guidelines rather than a company’s formula. It found no difference on its main measure at 3 months, but more people were in remission at 6 months, a secondary measure.

Debated: reviews that pool these trials tend to find a modest benefit for remission, with real caveats. A 2023 review from Canada rated its certainty in the size of the effect as very low, though it was more confident about the direction. A 2023 review from China found a benefit at 8 and 12 weeks but not at 4 or 24 weeks, and concluded that testing may help people reach remission more quickly. A 2026 review from Italy found a benefit in the trials of one commercial test, but results across other panels were inconsistent and imprecise.

Skeptical: a 2024 review by an American Psychiatric Association research workgroup, which is a review rather than a practice guideline, concluded that the evidence doesn’t support using currently available combinatorial tools, tests that combine results from several genes into one report, to choose treatment for major depression. The 2022 VA/DoD depression guideline found insufficient evidence to recommend for or against testing to help choose an antidepressant. In 2018, the FDA warned against many genetic tests with unapproved claims to predict a patient’s response to specific medications, noting that “the relationship between DNA variations and the effectiveness of antidepressant medication has never been established.” Our service page walks through the full research review, including how pooled studies have been read.

What FDA Labels and CPIC Guidelines Say

The strongest evidence sits with specific gene and medication pairs. Two psychiatric medications show what that looks like.

• Citalopram and CYP2C19. In people who are CYP2C19 poor metabolizers, citalopram levels can run higher, which raises the risk of a heart rhythm change called QT prolongation. The citalopram label sets a lower maximum dose for poor CYP2C19 processors, and also for people taking certain medicines that block the same enzyme. CPIC’s guideline suggests considering an antidepressant that isn’t mainly processed by CYP2C19, or adjusting the dose if citalopram is used.
• Carbamazepine and the HLA-B gene. Carbamazepine is a mood stabilizer and seizure medication. A variant called HLA-B*15:02 is found almost only in people with ancestry from broad areas of Asia. Carriers have a strongly increased risk of two rare, severe skin reactions, Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN). The carbamazepine label says people with ancestry in at-risk populations should be screened before starting it, and carriers shouldn’t take it unless the benefit clearly outweighs the risk. This variant doesn’t predict milder rashes. Most of these reactions happen in the first few months of treatment.
• Carbamazepine and the HLA-A gene. A second variant, HLA-A*31:01, is linked to a wider range of skin and allergic reactions to carbamazepine. It appears in both the label and the CPIC carbamazepine guideline.

Notice what these examples are about: how much medication builds up in your body and the risk of a specific reaction. Neither one says whether the medication will ease your symptoms.

Some test reports also include serotonin-related genes such as SLC6A4 and HTR2A. CPIC’s guideline says the data for these two don’t support using them in antidepressant prescribing.

What the Test Can’t Tell You

Knowing the limits matters as much as knowing the uses. Pharmacogenetic testing can’t tell you:

• Whether a medication will work for you. It can’t predict that. The NIH’s genome institute says “pharmacogenomic testing cannot identify the ‘perfect’ medication for a patient.” How fast you process a medication can still matter: for some people, very fast processing means lower drug levels and a weaker effect. That’s why CPIC suggests considering an antidepressant other than citalopram or escitalopram for CYP2C19 ultrarapid metabolizers.
• Your diagnosis. It doesn’t diagnose depression, anxiety or any other condition.
• Your risk of developing a condition. Pharmacogenetic testing isn’t designed to tell you your risk of developing a condition, even one that runs in your family. MedlinePlus puts it plainly: “Pharmacogenetic testing cannot diagnose any conditions or tell you about your risk for developing them.” Still, some panels include genes with other health implications, so ask what your panel covers.
• Side effects that aren’t gene-related. Many side effects have nothing to do with how fast you process a medication.

The FDA notes that your genetic makeup is only one of many factors that affect how a medication works. CPIC calls test results one of many pieces of clinical information.

A result with no expected gene interactions isn’t a guarantee either. Other medications, age and diet can shift how you process a drug, and the genome institute notes that pharmacogenomic tests “may miss important genomic variants that are more common in certain populations.”

Please don’t change a medication based on test results alone; talk with your provider first. Please don’t stop a medication on your own; some can cause withdrawal symptoms or a return of symptoms if stopped suddenly.

Who Might Consider It

Pharmacogenetic testing isn’t routinely needed before starting most medications. A few FDA labels are exceptions. The carbamazepine label recommends HLA-B*15:02 screening before starting it in people with ancestry from at-risk populations, and the pimozide label, for a medication used for tics in Tourette’s disorder, calls for CYP2D6 testing before higher doses. Otherwise, testing may be most worth discussing if:

• You’ve tried several medications without enough improvement.
• You’ve had strong side effects from medications.

Even then, the evidence is limited, and testing is one option among several. A careful look at the diagnosis, how long and at what dose earlier medications were tried, other medications, sleep and stress often matters just as much. Our posts on why finding the right medication can take time and what to do if medication doesn’t work go further.

Cost and Insurance

Coverage for this testing varies a lot. In a 2025 review of public policies from 14 U.S. insurers, 7 didn’t cover pharmacogenomic testing for psychiatric care, and the other 7 covered it only under certain conditions or for specific tests. Some plans also require prior authorization. Coverage is more common for some specific gene and medication pairs: a 2025 U.S. review reported that in 12 policies from large insurers and lab benefit managers, HLA-B testing for carbamazepine was covered in all 12, while CYP2C19 testing for citalopram was covered in 4. Costs vary by test and by insurance plan. Before testing, check with your plan whether it’s covered, and ask for a cost estimate.

Medicare covers this testing only in limited situations, and its rules vary by region. We don’t accept Medicare (including Medicare Advantage) or Medicaid (NJ FamilyCare). Because of Medicare and Medicaid rules, we also can’t see patients with these plans on a private-pay basis.

How We Use It at ANK Behavioral Health

At ANK, pharmacogenetic testing is ordered when your provider thinks it will help, after a careful evaluation. If they recommend it, they’ll explain why, what it can show and how they’ll use the results. When results come back, we review them together. The test is a tool, not the whole answer: your symptoms, history and goals still guide the plan.

Whether pharmacogenetic testing makes sense, which test is used, and who orders and interprets it are decisions your provider makes with you. Not every provider orders it. Within our practice, your provider may order and review it with you directly, or may have you meet with a colleague who focuses on pharmacogenetic testing for that part of your care.

If you’ve already had a pharmacogenetic test elsewhere, bring your results to your visit. If you’d like us to, and with your permission, we can share information with your primary care provider so your care fits together.

Most visits are by secure video, anywhere in New Jersey. Our office is in Mount Laurel, NJ, and in-person visits there are limited and by appointment only; ask our office. Looking for a psychiatrist in New Jersey? Our providers are board-certified psychiatric nurse practitioners (PMHNPs), who provide psychiatric evaluation and medication management.

Questions to Ask Your Provider

If testing comes up, a few questions can help you decide. The most important one is simple: would the results change anything in my plan? Our post on evidence-based psychiatry explains how research guides these decisions.

Frequently Asked Questions

Is pharmacogenetic testing worth it?

It depends on your situation, and researchers disagree about how much it helps. It may be worth discussing if several medications haven’t helped enough or caused strong side effects, as long as you know it’s a guide, not a guarantee.

Does insurance cover pharmacogenetic testing?

Sometimes. Coverage varies by plan, and many insurers cover it only under certain conditions or not at all for psychiatric care. Check with your plan before testing.

Will my provider order this test?

Not necessarily. Whether testing makes sense, which test is used, and who orders and interprets it are decisions your provider makes with you. Within our practice, your provider may order and review it with you directly, or may have you meet with a colleague who focuses on pharmacogenetic testing for that part of your care.

Can pharmacogenetic testing tell me if I’ll develop depression or another condition?

No. Pharmacogenetic testing isn’t designed to tell you your risk of developing a condition. It looks at how your body may process certain medications. Some panels include genes with other health implications, so ask what your panel covers.

Do I need to repeat the test?

Usually not for the same genes, since inherited variants don’t change over your life. Still, older or more limited panels may have missed some variants, since different tests look at different sets of variants, and how results are read is updated as research grows. After a liver or stem cell (bone marrow) transplant, the sample tested may not reflect the enzymes in your liver that process many medicines, and CPIC advises caution in reading results. Ask your provider whether older results still apply.

Is a psychiatric nurse practitioner the same as a psychiatrist?

No. Psychiatrists are medical doctors (MD or DO) who complete a residency in psychiatry. Psychiatric nurse practitioners are advanced practice registered nurses with graduate training in psychiatric care, and ours are board-certified (PMHNP-BC). In New Jersey, both can evaluate, diagnose and prescribe psychiatric medication. Our providers focus on careful evaluation and medication management.

If several medications haven’t worked out, a psychiatric evaluation can take a fresh look at the whole picture, including whether pharmacogenetic testing makes sense, and medication management can help you stay on track once you have a plan.

This article is for general educational purposes only. It does not constitute medical advice, does not establish a provider-patient relationship, and is not a substitute for individual evaluation, diagnosis, or treatment by a qualified clinician. If you’re in crisis or thinking about harming yourself, call or text 988 (Suicide & Crisis Lifeline), any time, day or night. In an emergency, call 911 or go to the nearest emergency room.