Consumer blood testing: revision 1 to 2

Revision 1

Revision 2

categories: ["longevity", "society"]

tags: ["diagnostics", "screening", "overdiagnosis", "preventive medicine", "biomarkers", "direct-to-consumer"]

summary: "The sale of clinical laboratory panels directly to asymptomatic consumers as preventive medicine, where accurate assays meet a screening logic that guidelines reject."

summary: "The sale of clinical laboratory panels directly to asymptomatic consumers as preventive medicine, where accurate assays meet a screening logic that guidelines reject."

updated: "2026-07-27"

updated: "2026-10-29"

humanEvidence: "The individual assays are standard clinical chemistry validated in people; no randomized trial has shown that broad panels in asymptomatic adults reduce illness or death, and trials of periodic general health checks found no mortality benefit."

humanEvidence: "The individual assays are standard clinical chemistry validated in people; no randomized trial has shown that broad panels in asymptomatic adults reduce illness or death, and trials of periodic general health checks found no mortality benefit."

access: "Sold online in most US states and in several other countries without a physician referral; single panels cost tens to hundreds of dollars and annual memberships several hundred, almost never reimbursed."

access: "Sold online in most US states and in several other countries without a physician referral; single panels cost tens to hundreds of dollars and annual memberships several hundred, almost never reimbursed."

reversibility: "context"

**Consumer blood testing** is the sale of clinical laboratory panels directly to people who have no

symptoms and no physician's referral, marketed as preventive or proactive medicine. The assays

themselves are ordinary hospital chemistry, run in certified laboratories, and mostly accurate. What

themselves are ordinary hospital chemistry, run in certified laboratories, and mostly accurate. The

is contested is the inference drawn from them: whether measuring dozens or hundreds of analytes in a

dispute is over the inference drawn from them: whether measuring dozens or hundreds of analytes in a

person who feels well yields information that makes them healthier, or mainly yields

person who feels well yields information that makes them healthier, or mainly produces

abnormal-looking numbers.

```keyfacts

## What is sold

The transaction is uniform across sellers. A consumer orders a panel online; a contracted physician

The transaction works the same way across sellers. A consumer orders a panel online; a contracted

network signs the requisition where state law still requires a clinician's order; blood is drawn at a

physician network signs the requisition where state law still requires a clinician's order; blood is

commercial service centre, by a mobile phlebotomist, or through a mail-in capillary kit. Results

drawn at a commercial service centre, by a mobile phlebotomist, or through a mail-in capillary kit.

arrive in an application rather than a consultation, colour-coded against a range and accompanied by

Results arrive in an application rather than a consultation, colour-coded against a range and

generated advice on diet, [[dietary-supplements]], sleep and training.

accompanied by generated advice on diet, [[dietary-supplements]], sleep and training.

Quest and Labcorp, the dominant American reference laboratories, sell consumer-initiated versions of

the tests they run for hospitals. Everlywell built its business on mail-in kits aimed at single

> [!note] What certification covers

> A CLIA-certified laboratory has demonstrated that it measures what it claims to measure, to a

> stated precision, under audit. Certification says nothing about whether the test should have been

> ordered. Consumer marketing routinely answers the accuracy question while appearing to answer the

> ordered. Consumer marketing routinely answers the question of accuracy in a way that appears to

> utility one.

> answer the question of usefulness.

## Development history

```timeline

[

  { "year": "2013–2014", "title": "Theranos enters retail", "text": "Theranos opens wellness centres inside Walgreens stores, first in California and then across the Phoenix area, promising a full laboratory panel from a fingerstick sample." },

  { "year": "2013-2014", "title": "Theranos enters retail", "text": "Theranos opens wellness centres inside Walgreens stores, first in California and then across the Phoenix area, promising a full laboratory panel from a fingerstick sample." },

  { "year": "2015", "title": "Direct-access testing broadens", "text": "Arizona enacts a law, lobbied for by Theranos, letting residents order any laboratory test without a physician's order; a Wall Street Journal investigation later that year reports that the company runs only a small fraction of the tests it offers on its own device." },

  { "year": "2015", "title": "Direct-access testing broadens", "text": "Arizona enacts a law, lobbied for by Theranos, letting residents order any laboratory test without a physician's order; a Wall Street Journal investigation later that year reports that the company runs only a small fraction of the tests it offers on its own device." },

  { "year": "2016", "title": "Regulators act", "text": "The US agency that certifies clinical laboratories revokes the certificate of Theranos's California facility and bars Elizabeth Holmes from operating a lab for two years. The company voids or corrects large numbers of previously reported results." },

  { "year": "2016", "title": "Regulators act", "text": "The US agency that certifies clinical laboratories revokes the certificate of Theranos's California facility and bars Elizabeth Holmes from operating a lab for two years. The company voids or corrects large numbers of previously reported results." },

  { "year": "2019–2021", "title": "Reference labs go direct", "text": "Quest and Labcorp launch consumer-facing platforms selling much of their own catalogue without a physician visit." },

  { "year": "2019-2021", "title": "Reference labs go direct", "text": "Quest and Labcorp launch consumer-facing platforms selling much of their own catalogue without a physician visit." },

  { "year": "2020", "title": "The pandemic normalizes home collection", "text": "Emergency authorizations for at-home sample collection, including one granted to Everlywell, make mail-in specimen kits a familiar consumer product." },

  { "year": "2020", "title": "The pandemic normalizes home collection", "text": "Emergency authorizations for at-home sample collection, including one granted to Everlywell, make mail-in specimen kits a familiar consumer product." },

  { "year": "2022", "title": "Fraud conviction", "text": "Holmes is convicted in January on four counts of defrauding investors, acquitted on the counts relating to patients, and sentenced later that year to more than eleven years." },

  { "year": "2022", "title": "Fraud conviction", "text": "Holmes is convicted in January on four counts of defrauding investors, acquitted on the counts relating to patients, and sentenced later that year to more than eleven years." },

  { "year": "2023–2026", "title": "Longevity-branded panels", "text": "Membership services selling a hundred or more analytes a year, framed around biological aging rather than disease, become a prominent part of the market." }

  { "year": "2023-2026", "title": "Longevity-branded panels", "text": "Membership services selling a hundred or more analytes a year, framed around biological aging rather than disease, become a prominent part of the market." }

]

```

Theranos is the field's cautionary anchor, and it is routinely misremembered. The fraud lay not in

Theranos is the field's standard cautionary tale, and it is routinely misremembered. The fraud did

selling consumers their own laboratory data but in the specific, false claim that a full panel could

not lie in selling consumers their own laboratory data. It lay in the specific, false claim that a

be run from a few drops of capillary blood on a proprietary device.[^carreyrou2018] Most samples the

full panel could be run from a few drops of capillary blood on a proprietary device.[^carreyrou2018]

company processed went to conventional commercial analysers. The demand it identified outlived it,

Most samples the company processed went to conventional commercial analysers. The demand it

which is why the same retail model returned within a few years under companies making no device

identified outlived it, which is why the same retail model returned within a few years under

claims at all.

companies making no device claims at all.

## The arithmetic of reference ranges

The central scientific objection to broad panels is not about assay quality. A reference interval

The central scientific objection to broad panels has nothing to do with assay quality. A reference

describes a population; it is not a boundary between health and disease. By convention a laboratory

interval describes a population, and it does not mark a boundary between health and disease. By

sets it to cover the central 95 percent of a defined healthy reference group,[^clsi] so roughly one

convention a laboratory sets it to cover the central 95 percent of a defined healthy reference

result in twenty from a person with nothing wrong falls outside it by construction.

group,[^clsi] so roughly one result in twenty from a person with nothing wrong falls outside it by

construction.

Test many analytes at once and the flags accumulate. On twenty statistically independent tests the

When many analytes are tested at once, the flags accumulate. On twenty statistically independent

probability that a completely healthy person has at least one out-of-range value is about 64

tests the probability that a completely healthy person has at least one out-of-range value is about

percent; on a hundred, an all-normal report is the unusual outcome. Real analytes are correlated, so

64 percent; on a hundred, an all-normal report is the unusual outcome. Real analytes are correlated,

the true figure is lower than the naive calculation, but the direction is not in doubt. A panel wide

so the true figure is lower than the naive calculation, but the direction of the effect is clear. A

enough to be marketed as comprehensive is wide enough that most healthy customers will see something

panel wide enough to be marketed as comprehensive is wide enough that most healthy customers will

flagged.

see something flagged.

Intervals carry a second problem: each is drawn from whichever population a laboratory sampled on

Intervals carry a second problem. Each is drawn from whichever population a laboratory sampled, on

whichever platform it ran, so the same blood can be flagged by one laboratory and normal in another.

## What screening evidence supports

The closest direct test of the strategy is the trial literature on periodic general health checks,

which bundled bloodwork with examination. A Cochrane review of those randomized trials found no

effect on all-cause, cardiovascular or cancer mortality while documenting increases in diagnoses and

in treatment.[^krogsboll] That is a result about a programme rather than about any single analyte,

in treatment.[^krogsboll] The result concerns a programme, not any single analyte, but it remains

and it remains the nearest evidence to the claim consumer testing makes.

the nearest evidence to the claim consumer testing makes.

> [!debate] Population range or personal baseline

> The strongest case for repeat testing is not the cross-sectional flag but the within-person trend:

> The strongest case for repeat testing rests on the within-person trend, not the cross-sectional

> an individual varies around their own set point within a band narrower than the population

> flag. An individual varies around their own set point within a band narrower than the population

> interval, so a real drift can be visible while every value stays technically normal. Personal

> reference intervals are a long-standing idea in laboratory medicine, and the same argument underpins

> [[continuous-glucose-monitoring]] and [[wearable-health-sensors]]. Missing is any trial showing

> [[continuous-glucose-monitoring]] and [[wearable-health-sensors]]. What is missing is any trial

> that acting on such trends improves outcomes in healthy people.

> showing that acting on such trends improves outcomes in healthy people.

## Multi-cancer early detection

laboratory-developed tests they have fallen outside routine premarket review; a 2024 US rule that

would have changed that was vacated by a federal district court the following year.[^fdaldt2024]

The performance data are genuinely interesting and genuinely insufficient. In a published prospective

The performance data are interesting but insufficient. In a published prospective

study of Galleri in adults aged 50 and older, fewer than half of the participants with a positive

signal proved to have cancer, and resolution took months of imaging and procedures for many of

them.[^schrag2023] A large randomized trial in England enrolled on the order of 140,000 participants

## Harms and the diagnostic cascade

Overdiagnosis is the organising concept. Screening a healthy population finds indolent disease that

The harms are organised around overdiagnosis. Screening a healthy population finds indolent disease

would never have surfaced, and because the person is then treated, every such case enters the

that would never have surfaced, and because the person is then treated, every such case enters the

statistics as a life saved. Thyroid cancer in South Korea is the clearest documented case: incidence

rose roughly fifteenfold as ultrasound screening spread, with essentially no change in mortality

from the disease.[^ahn2014] That is an argument for grading each test against outcomes rather than

from the disease.[^ahn2014] The case argues for grading each test against outcomes instead of

against detection, not against screening as such.

detection; it is not an argument against screening as such.

The proximate harm from a flagged panel is the cascade. An out-of-range value invites a confirmatory

The most immediate harm from a flagged panel is the diagnostic cascade. An out-of-range value

test, then imaging, then an incidental finding on that imaging, then a procedure — a sequence in

invites a confirmatory test, then imaging, then an incidental finding on that imaging, then a

which each step is locally reasonable and the endpoint was not warranted by the starting point.

procedure. Each step is locally reasonable, yet the starting point did not warrant the endpoint.

Diagnostic labels also carry an effect independent of treatment, documented since a 1978 study found

absenteeism rose among steelworkers after they were told they were hypertensive.[^haynes1978]

The reversibility of the technology is unusual here. The draw leaves nothing behind and testing can

Reversibility is unusual in this case. The draw leaves nothing behind and testing can simply be

simply be stopped; a result cannot be un-known, and the workup it sets off can end in a biopsy, a

stopped, but a result cannot be un-known, and the workup it sets off can end in a biopsy, a

diagnosis, or a permanent notation that no later normal test removes.

Two structural harms round out the list. Companies that do not bill insurers may sit outside the

There are also two structural harms. Companies that do not bill insurers may sit outside the

health-privacy rules governing clinical records, accumulating longitudinal biological data under

consumer-software terms, which connects the practice to [[genetic-discrimination]] and

[[mental-privacy]]. And because the panels are paid out of pocket, they route clinician attention

[[glp-1-receptor-agonists]], and to composite [[biological-age|biological age]] scores whose

underlying [[aging-biomarkers|biomarker validity]] is unresolved.

Three developments would change the assessment. The first is randomized evaluation of the strategy

Three developments would change the assessment. The first is randomized evaluation of the strategy,

rather than the assay: whether people given a broad panel end up healthier than people not given one.

as opposed to the assay: whether people given a broad panel end up healthier than people not given one.

The second is a shift from population intervals to individual baselines with published reliability,

which would make repeat testing statistically defensible and is where

[[human-digital-twins|integrated personal modelling]] points. The third is triage that reduces

[[nanoparticle-diagnostics|nanoscale assays]].

Until then the field faces an uncomfortable comparison. The actions most reliably recommended after

a consumer panel — [[exercise-and-aging|structured exercise]], [[sleep-and-longevity|sleep]],

a consumer panel ([[exercise-and-aging|structured exercise]], [[sleep-and-longevity|sleep]], weight,

weight, alcohol, smoking —

alcohol, smoking) are the ones recommended without it, and the [[healthspan]] case built on the

are the ones recommended without it, and the [[healthspan]] case built on the

[[geroscience-hypothesis]] does not obviously require a hundred analytes to make. A test is worth

[[geroscience-hypothesis]] does not obviously require a hundred analytes to make. A test that

its cost if it changes what someone does, and most of these tests have not been shown to.

changes what someone does is worth its cost. Most of these have not been shown to.

## See also