What Genetic Testing for Medications Costs, and When Insurance Pays

Oak
pharmacogenomics cpic pharmcat sequencing insurance
A small feathered reptile on a steel stand in black studio light, beside ten glowing feather-filled vials and one continuous luminous strand on a spool.

A self-pay pharmacogenomic (PGx) panel, meaning a targeted genotyping test that reads a fixed list of variants in drug-metabolizing genes, costs between about $200 and $500 in the United States today. GeneSight lists $330 or less for cash payers, Mercy’s health-system panel lists $295, and most laboratory-developed panels sit in the same band. Insurance covers these tests inconsistently: Medicare Part B pays for a single-gene or panel PGx test when the beneficiary has been prescribed, or is a candidate for, a drug whose label or guideline supports the specific gene, and commercial plans vary from full coverage to categorical denial of “multi-gene panels” as investigational. The $5,000 bills that show up on Reddit are almost always out-of-network laboratory charges billed per gene before an insurer adjudicates them down, not what anyone ends up paying. Our view is that if you are going to spend money on this once, spend it on a 30x whole-genome sequence rather than a panel, because the genotype is fixed for life and the interpretation layer is the part that keeps changing.

What the tests cost and what the billing codes mean

PGx testing is billed per gene under CPT codes, which is why the sticker price and the reimbursed price diverge so widely. The common ones are 81225 (CYP2C19), 81226 (CYP2D6, including copy number), 81227 (CYP2C9), 81328 (SLCO1B1), 81355 (VKORC1), 81306 (NUDT15), 81350 (UGT1A1) and 81335 (TPMT). A ten-gene panel billed as ten separate codes can generate a five-figure gross charge. Many laboratories now bill a single proprietary laboratory analysis (PLA) code instead, a 0xxxU code specific to one vendor’s assay, which is cleaner but means coverage hinges on whether your plan has a policy for that exact code.

Cash prices are the more useful number because they are what you can plan around. Expect $200-$500 for a 10-25 gene panel from a clinical laboratory, $99-$250 for consumer-grade PGx add-ons built on genotyping arrays, and $600-$1,500 for a clinical-grade whole exome. Whole-genome sequencing at 30x coverage from a research or direct-to-consumer provider is now commonly $400-$1,000 for the sequencing alone, and clinical-grade WGS with interpretation runs several times that. The cost gap between panels and sequencing has collapsed over the last decade: the MedSeq Project measured the full cost of WGS including interpretation and reporting at a time when the sequencing itself dominated, and found that labor for interpretation was the persistent component while sequencing costs fell around it 1. The earlier framework for these analyses made the same point, that the informatics and clinical-interpretation steps, not the reagents, determine what genomic testing costs a health system 2.

When insurance pays

Coverage follows a narrow logic: payers reimburse a test when it changes a decision about a drug you are about to take, and they resist paying for a panel that anticipates decisions you have not made yet. Medicare’s national coverage determination for pharmacogenomic testing of CYP2C19 in combination with clopidogrel, and the MolDX local coverage policies that govern most PGx claims, both require documentation that the patient is a candidate for a specific drug with a supporting FDA label or CPIC guideline, and they generally allow one test per gene per lifetime. If you meet those criteria, your out-of-pocket cost under Part B is typically nothing, because clinical laboratory services are paid at 100% without coinsurance.

Commercial coverage is where the variance lives. A systematic review of payer policies for pharmacogenomic and multi-gene testing found substantial inconsistency across insurers in which gene-drug pairs were covered and how “medical necessity” was defined 3. The broader pattern is structural: preventive genetic testing in people who are not yet sick falls into a gap where the test is neither diagnostic nor an established preventive service, and coverage decisions in that gap have historically favored those who can pay out of pocket 4. Out-of-pocket exposure also shapes whether tests get ordered at all. In a study of next-generation sequencing panels for hereditary cancer risk, clinicians reported that anticipated patient cost-sharing directly influenced whether they ordered a test and which one 5.

Two practical consequences. First, ask the laboratory for its cash price before you involve insurance, because the cash price is often lower than your deductible-phase responsibility. Second, if you are already on a drug with PGx information in its label, that is the strongest coverage argument you have. Utilization of testing in exactly that population remains low: among patients prescribed drugs carrying pharmacogenomic information in FDA-approved labeling, documented rates of relevant genetic testing were in the low single-digit percentages 6.

Panel versus sequence

A panel is a fixed list of positions. That is its advantage (CYP2D6 copy number and hybrid alleles are hard, and good clinical panels use targeted assays such as long-range PCR or MLPA to resolve them) and its limitation (a variant that is not on the list is reported as absent, which for star-allele calling means silently assigned to the reference allele). The clinically important example is DPYD: the HapB3 variant rs75017182, an intronic splice variant, is absent from several older panels and from most genotyping arrays, and missing it produces a normal-metabolizer call in someone who is not one. Arrays have the same problem at scale, since imputation cannot recover structural variation in CYP2D6.

Sequencing inverts the tradeoff. A 30x whole genome gives you every single-nucleotide variant and small indel in every pharmacogene at once, including ones that have not been assigned clinical significance yet, and you can reanalyze it against a new CPIC guideline in an afternoon without drawing blood again. Formal cost-effectiveness analyses of genomic tests turn on exactly this kind of reuse, because the value of a sequence accrues across many downstream decisions rather than one 7. Sequencing is also the substrate for questions a PGx panel cannot address at all, including eligibility for variant-specific gene-targeted therapies, where the genome sequence is what establishes whether a given mechanism applies to a given person 8.

The weakness of short-read WGS is CYP2D6. The gene sits next to the CYP2D7 pseudogene, shares roughly 94% sequence identity with it, and generates hybrid alleles and whole-gene duplications that 150 bp reads resolve imperfectly. Dedicated callers handle most of it, and the remainder you should treat as uncertain rather than negative.

Running the interpretation yourself

If you have a VCF, the tool we use is PharmCAT. It takes a GRCh38-aligned VCF, normalizes it, matches diplotypes, and emits phenotype calls plus CPIC and DPWG recommendations as JSON and HTML.

# 1. Preprocess: normalize, split multiallelics, restrict to PharmCAT positions
pharmcat_vcf_preprocessor -vcf sample.g.vcf.gz \
  -refFna GRCh38_full_analysis_set_plus_decoy_hla.fa \
  -o ./pgx/

# 2. CYP2D6 from WGS via Cyrius (star-allele + CNV aware)
star_caller.py --manifest bam_list.txt --genome 38 \
  --prefix sample --outDir ./cyp2d6/ --threads 4

# 3. Run PharmCAT, feeding in the outside CYP2D6 call
pharmcat -vcf ./pgx/sample.preprocessed.vcf.bgz \
  -po ./cyp2d6/sample.cyp2d6.tsv \
  -reporterJson -o ./report/

Three details determine whether the output is trustworthy. Use a gVCF or a joint-called VCF that represents reference-confident regions explicitly, so PharmCAT can distinguish “reference at this position” from “no data at this position” and flag the latter as a no-call instead of assuming wild type. Check the phasing status in the match JSON, because unphased heterozygous calls at multiple positions in CYP2C19 or CYP2B6 yield ambiguous diplotypes and PharmCAT will report all consistent possibilities. And handle HLA separately: HLA-B*57:01 and HLA-A*31:01 are not part of standard star-allele calling and need a dedicated typer such as HLA*LA or OptiType, or a clinical serology-grade assay.

PharmVar is the reference for allele definitions and PharmGKB for the evidence levels behind each gene-drug pair. Read the CPIC guideline for the specific pair before you form any opinion about what a phenotype call means.

Where this stops being a data problem

A phenotype call is a measurement, and measurements do not prescribe. Whether a CYP2C19 intermediate-metabolizer result should change anything about a drug you take is a clinical decision that belongs to your prescriber or a pharmacist, who has to weigh the drug, the indication, your other medications, kidney and liver function, and what has already worked or not worked for you. Bring the PharmCAT report and the relevant CPIC guideline to that conversation rather than a conclusion. Be aware, too, that genomic results carry non-medical consequences that differ by country and product line: in the UK, for instance, the interaction between genomic data and insurance underwriting is governed by a voluntary code whose boundaries are still being negotiated 9.

Questions people also ask

Does insurance cover gene testing for medication? Sometimes, and narrowly. Coverage generally requires that you are already prescribed or being considered for a specific drug with FDA label or CPIC guideline support for the gene in question, and many plans exclude broad multi-gene panels as investigational. Policies differ substantially between insurers for the same gene-drug pair 3.

Does Medicare pay for pharmacogenomic testing? Yes, under conditions. Medicare pays for PGx testing when documentation shows the beneficiary is a candidate for a drug with recognized gene-drug evidence, typically limited to one test per gene per lifetime. Covered clinical laboratory services under Part B carry no coinsurance.

How much does genetic testing for meds cost without insurance? Roughly $200-$500 for a clinical 10-25 gene panel, less for consumer array-based add-ons, and $400-$1,000 for 30x whole-genome sequencing without clinical interpretation. HSA and FSA funds generally apply.

Is genetic testing for medication worth it? It is worth it if you are facing a specific decision about a drug with a strong gene-drug pair, and it is worth it as infrastructure if you buy a sequence rather than a panel, since the data outlives any single guideline. Panels bought speculatively often return normal-metabolizer calls across the board and change nothing.

Where can I get a pharmacogenomic test? Through a prescriber ordering from a clinical laboratory, through health-system pharmacy programs that increasingly run their own panels, or by sequencing your genome and running PharmCAT yourself. The third route gives you the data permanently but still requires a clinician to act on it.

Oak builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.

Footnotes

  1. Kurt D. Christensen, Kathryn A. Phillips, Robert C. Green, et al. Cost Analyses of Genomic Sequencing: Lessons Learned from the MedSeq Project. Value in Health, 2018. https://doi.org/10.1016/j.jval.2018.06.013 ↩

  2. Kurt Christensen, Dmitry Dukhovny, Uwe Siebert, et al. Assessing the Costs and Cost-Effectiveness of Genomic Sequencing. Journal of Personalized Medicine, 2015. https://doi.org/10.3390/jpm5040470 ↩

  3. Christine Lu, Stephanie Loomer, Rachel Ceccarelli, et al. Insurance Coverage Policies for Pharmacogenomic and Multi-Gene Testing for Cancer. Journal of Personalized Medicine, 2018. https://doi.org/10.3390/jpm8020019 ↩ ↩2

  4. Anya E.R. Prince. Prevention for those who can pay: insurance reimbursement of genetic-based preventive interventions in the liminal state between health and disease. Journal of Law and the Biosciences, 2015. https://doi.org/10.1093/jlb/lsv008 ↩

  5. Grace A. Lin, Julia R. Trosman, Michael P. Douglas, et al. Influence of payer coverage and out‐of‐pocket costs on ordering of NGS panel tests for hereditary cancer in diverse settings. Journal of Genetic Counseling, 2021. https://doi.org/10.1002/jgc4.1459 ↩

  6. John Young, Kaustuv Bhattacharya, Sujith Ramachandran, et al. Rates of genetic testing in patients prescribed drugs with pharmacogenomic information in FDA-approved labeling. The Pharmacogenomics Journal, 2021. https://doi.org/10.1038/s41397-021-00211-1 ↩

  7. Katherine Payne, Sean P. Gavan, Stuart J. Wright, et al. Cost-effectiveness analyses of genetic and genomic diagnostic tests. Nature Reviews Genetics, 2018. https://doi.org/10.1038/nrg.2017.108 ↩

  8. Jerry Vockley, Annemieke Aartsma‐Rus, Jennifer L. Cohen, et al. Whole‐genome sequencing holds the key to the success of gene‐targeted therapies. American Journal of Medical Genetics Part C: Seminars in Medical Genetics, 2022. https://doi.org/10.1002/ajmg.c.32017 ↩

  9. Padraig Dixon, Rachel H. Horton, William G. Newman, et al. Genomics and insurance in the United Kingdom: increasing complexity and emerging challenges. Health Economics, Policy and Law, 2024. https://doi.org/10.1017/s1744133124000070 ↩