Own Your Labs: Ordering Your Own Blood Work and Keeping the Data

Oak
labs biomarkers direct-access-testing data-ownership loinc
A kneeling figure in a glowing forest tends a glass cabinet of luminous colored vials wired by tendrils to nearby trees.

“Own your labs” describes two different things, and the companies that rank for the phrase have solved only one of them. The first is ordering blood work yourself, without a physician visit. That is legal for most tests in most US states through direct-access testing. A marketplace routes your order to a physician network that signs the requisition, you draw at a Quest or LabCorp patient service center, and results come back to you. The second is owning the result as data, meaning a structured record with a LOINC code, a numeric value, units, the reference interval the lab used, the assay method, and a collection timestamp, stored somewhere you control and queryable years later. The first costs $30 to $300. The second costs an afternoon of setup, and it is the part almost nobody does.

How self-ordered lab testing works

It helps to understand what these services are and are not before you compare them on price. Direct-access testing marketplaces such as Own Your Labs, Walk-In Lab, Marek, Quest’s own Quest Health, and Labcorp OnDemand are not laboratories. They are order-entry and payment layers on top of the two national reference labs. You pay cash, a licensed physician in the network signs the requisition without seeing you, and you receive an order number and a patient service center address. The specimen then runs on the same Roche cobas, Abbott Architect, or Siemens Atellica analyzers that process physician-ordered samples from the same draw station. The analytical pipeline is identical; what differs is who authorized the order and who receives the report.

A small number of states restrict this arrangement. New York has been the most restrictive jurisdiction for consumer-initiated testing, and a few others limit specific analyte categories. Some tests are gated everywhere, for good reasons: anything where the result drives immediate clinical action, and anything with reportable-disease implications.

The prices look low against a hospital bill because they are cash prices on a bundled panel with no insurance adjudication. A hospital outpatient draw generates a chargemaster line item, a payer negotiation, a patient responsibility calculation, and a collections workflow, and the facility fee is often larger than the assay cost. By contrast, the marginal cost of running a comprehensive metabolic panel on a high-throughput chemistry analyzer is a few dollars of reagent. So when a company sells you a CMP for $15, nothing suspicious is happening. You are seeing the reagent cost plus a phlebotomy slot plus a margin, without the billing apparatus.

Get the data out in a structured form, not as a PDF

Having ordered the test, the next question is what form the result takes, and this is where most people lose the value of the draw. The default deliverable is a PDF, which is the wrong artifact for anything you intend to track over time. Under the HIPAA right of access and the information-blocking provisions of the 21st Century Cures Act, you can request an electronic copy in the form and format you request if it is readily producible. Both national labs expose FHIR R4 patient APIs, and their portals will export results as structured Observation resources. That is what you want, because each Observation carries a LOINC code (the standard identifier for a lab test, so that “glucose, serum, fasting” resolves to 1558-6 rather than to whatever string your lab prints), a valueQuantity with UCUM units, a referenceRange, and an effectiveDateTime.

Our recommendation is a small Git repository holding three things, so that the record is both complete and reproducible. Those are the original PDF and any raw instrument report, a single labs.csv in long format, and a parser that regenerates the CSV from the source files. The schema we use has one row per analyte per draw:

collected_at, lab, panel, loinc, analyte, value, unit,
ref_low, ref_high, flag, method, specimen, fasting_hours, source_file

Of these fields, method is the one people most often leave empty, and it is the one that saves you later. Total testosterone by competitive immunoassay and by liquid chromatography–tandem mass spectrometry are different measurements with different biases, especially at low concentrations. 25-hydroxyvitamin D results differ across platforms. Folate measurement has a long history of assay change, and comparing values across eras without accounting for the method is a known source of error in even carefully run national surveys 1. If you change labs and your ferritin appears to jump 20%, the first hypothesis should be the platform, not your iron stores.

Interpreting a change: reference change value, not the reference interval

Once you have a series rather than a single result, the interpretive question changes, and the number printed on the report stops being the right guide. The reference interval is a population statement, usually the central 95% of a reference sample. Being inside it tells you little about whether you have moved. For a longitudinal series on yourself, the quantity that matters is the reference change value:

RCV = 2**0.5 * Z * sqrt(CVa**2 + CVi**2)

Here CVa is the analytical coefficient of variation for the assay on that platform, which the lab will supply and many publish in their test directory. CVi is the within-subject biological variation for that analyte, for which the EFLM Biological Variation Database is the reference source. Z is 1.96 for a two-sided 95% criterion and 2.33 for a one-sided 99%. Analytes with tight biological variation, such as serum sodium or albumin, need only a small move to be real. Analytes with wide within-subject variation can swing substantially between draws with nothing underlying it. Triglycerides, ferritin, and CRP are examples. Compute RCV once per analyte, store it in a lookup table alongside your CSV, and let your plotting code draw the band. It will stop you from chasing noise.

Before refining the analysis, though, tighten the collection. Preanalytical control does more for your data quality than any analytical refinement. Standardize the draw. Use the same patient service center where possible and the same time of day, with mornings for cortisol and testosterone, which have strong diurnal structure. Fast 10 to 12 hours for lipids and glucose if you are trending them, and sit for five minutes before the draw, because posture shifts plasma volume and therefore every concentration measured. Avoid heavy exercise in the preceding 48 hours if you are tracking creatine kinase, creatinine, or ALT. Note high-dose biotin supplementation, which interferes with streptavidin-biotin immunoassays and can distort thyroid and troponin results. Record hemolysis and other specimen flags in your CSV rather than discarding them, because a hemolyzed sample inflates potassium and LDH in a way that looks like signal.

Fingerstick and at-home collection are a real step down

Collection method matters as much as collection conditions, and the convenient option carries a measurable cost. Low-volume self-collected samples are easier to obtain and measurably worse in quality. A controlled evaluation of direct-to-consumer low-volume testing in healthy adults compared self-collected results against standard venipuncture and found discordance across multiple analytes, with variability large enough to shift results across clinical decision thresholds 2. If a marker matters enough to track, draw it venous. Reserve at-home collection for analytes you are sampling frequently for trend shape rather than absolute level.

Know which tests are ready and which are not

The maturity of the assay should shape how much weight you put on the number. Self-ordering is straightforward for chemistry, hematology, and lipids. It is equally straightforward for thyroid function and common hormone and vitamin assays, all of which are well standardized. The situation is different for biomarkers at the research or early-clinical frontier. Plasma phosphorylated tau assays for Alzheimer pathology have moved quickly, with the first FDA clearance for a blood-based Alzheimer biomarker test arriving in 2025 3, and the available assays and their clinical performance data are reviewed thoroughly in the recent literature 4. Reimbursement and appropriate-use questions for these tests remain unsettled 5. We would not order one outside a clinical relationship, because the result changes what you should do next and the interpretation depends on a pretest probability that you cannot assess yourself.

The general caution here is older than any of these assays. The path from a candidate protein marker to a validated clinical test is long, and most candidates fail at analytical validation or in independent cohorts 6. The same holds for metabolomic panels, where quantitative standardization is still being worked out 7. It also holds for single-study diagnostics such as the nanoelectronic impedance assay proposed for ME/CFS, which was a promising proof of concept in a small cohort rather than a test you can act on 8. When you see a consumer panel offering a proprietary score derived from markers you cannot find in a standardization program, treat the number as uncalibrated.

One boundary remains in place regardless of how you obtain the test. An abnormal result, a symptom, or any question about what to do next belongs with a clinician. Ordering the test yourself is an access question, and deciding what it means for your care is a separate one.

Questions people also ask

What does “own your labs” mean? In common usage it refers to consumer-initiated lab testing, where you order and pay for blood work directly instead of going through a physician visit and insurance. We would extend it to include keeping the results as structured, portable data with codes, units, methods, and timestamps, because a PDF in a portal you may lose access to is not ownership.

Can I request my own blood work without a doctor? In most US states, yes, for most routine tests. A physician in the marketplace’s network signs the requisition on your behalf, which satisfies the legal requirement without a clinical encounter. New York is the notable exception, and certain test categories are restricted more broadly.

Which lab is cheaper, Quest or LabCorp? Neither is consistently cheaper. Cash prices vary by panel and by which marketplace you order through, and the same assay can differ by more than twofold across ordering channels. Price the specific panel across two or three marketplaces before ordering, and weigh proximity of the draw station, because consistency of collection site matters more for a longitudinal series than a few dollars.

Why is my bloodwork so expensive through my doctor? Because you are paying an insurance-negotiated rate plus a facility or draw fee, often with a deductible applied, rather than a cash price. The reagent and instrument cost of a standard chemistry panel is small. The billing chain is where the money goes.

Why is self-ordered lab testing so cheap? Bundled panels run on high-throughput analyzers at very low marginal cost, and skipping insurance billing removes a large administrative expense. The specimens are processed in the same CLIA-certified laboratories as physician-ordered samples, so a low price does not imply lower analytical quality.

What labs can you order for yourself? Broadly: complete blood count, comprehensive metabolic panel, lipids including ApoB and Lp(a), HbA1c and fasting insulin, thyroid panel with antibodies, iron studies and ferritin, vitamin D and B12, hs-CRP, sex hormones, and most infectious disease serologies. What you generally cannot self-order are tests tied to immediate clinical decisions and some genetic and oncology assays that require a treating clinician.

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. Barry Shane. Folate status assessment history: implications for measurement of biomarkers in NHANES. The American Journal of Clinical Nutrition, 2011. https://doi.org/10.3945/ajcn.111.013367 ↩

  2. Brian A. Kidd, Gabriel Hoffman, Noah Zimmerman, et al. Evaluation of direct-to-consumer low-volume lab tests in healthy adults. Journal of Clinical Investigation, 2016. https://doi.org/10.1172/jci86318 ↩

  3. Rita Rubin. What to Know About the First FDA-Cleared Blood Test for Alzheimer Biomarkers. JAMA, 2025. https://doi.org/10.1001/jama.2025.9013 ↩

  4. Araya Dimtsu Assfaw, Suzanne E Schindler, John C Morris. Advances in blood biomarkers for Alzheimer disease ( AD ): A review. The Kaohsiung Journal of Medical Sciences, 2024. https://doi.org/10.1002/kjm2.12870 ↩

  5. Eric Peterson. Who’s Paying for Those Blood-based Biomarker Tests for Alzheimer’s Disease?. Neurology Today, 2024. https://doi.org/10.1097/01.wnt.0001095436.22662.2c ↩

  6. Nader Rifai, Michael A Gillette, Steven A Carr. Protein biomarker discovery and validation: the long and uncertain path to clinical utility. Nature Biotechnology, 2006. https://doi.org/10.1038/nbt1235 ↩

  7. Natalie J. Serkova, Theodore J. Standiford, Kathleen A. Stringer. The Emerging Field of Quantitative Blood Metabolomics for Biomarker Discovery in Critical Illnesses. American Journal of Respiratory and Critical Care Medicine, 2011. https://doi.org/10.1164/rccm.201103-0474ci ↩

  8. R. Esfandyarpour, A. Kashi, M. Nemat-Gorgani, et al. A nanoelectronics-blood-based diagnostic biomarker for myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Proceedings of the National Academy of Sciences, 2019. https://doi.org/10.1073/pnas.1901274116 ↩