Supplement-Company Testing or Independent Testing: Who Should Interpret
When the same organisation measures you and sells you the answer, the measurement can be excellent and the recommendation still systematically skewed.
The Short Answer
A growing number of services combine measurement with supply: they run your panel, interpret it, and sell the products the interpretation implies. The assay in that arrangement is often genuinely good, run by an accredited laboratory to the same standard as anywhere else. The problem is downstream. Interpretation involves judgement, judgement responds to incentives, and the incentive in a combined model runs toward findings that require a purchase.
Where the Skew Enters
It is worth being precise, because the vague version of this criticism is easy to dismiss.
Not in the assay. A regulated laboratory measuring apoB produces the same number regardless of who ordered it. Content and purity testing is objective.
In the panel design. Which analytes are measured is a choice, and a panel including analytes with poor status validity or unvalidated interpretive frameworks will generate more actionable-looking findings than a deliberate one.
In the reference range used. Some services apply narrower optimal ranges than laboratory reference intervals. Sometimes that is defensible, since reference ranges describe populations with substantial disease prevalence. It also converts a normal result into a finding, and the direction of that conversion is not neutral.
In the interpretation of borderline results. Judgement calls resolved consistently in one direction across thousands of reports produce a pattern.
In what is omitted. The correct answer is frequently no change, or removal. A report that never says so is not interpreting.
In the follow-up cadence. Retesting intervals shorter than the marker's biology serve subscription retention rather than the member.
The Counter-Signals Worth Looking For
| Signal | What it indicates |
|---|---|
| Ever recommends removing something | The interpretation is not purely additive |
| Ever concludes no change is needed | The system is evaluating rather than generating |
| Publishes a list of things it recommends against | A filter is being applied |
| Uses standard laboratory reference ranges, or explains any deviation | Range selection is not doing hidden work |
| Returns raw numbers with units | You can check the interpretation yourself |
| Retest intervals matched to marker biology | Cadence set by evidence rather than retention |
| Asks about medications | Performing the highest-value safety function |
| Routes serious results to clinical care | Recognises its own boundary |
The first three are the strongest, and they cost the provider something, which is what makes them informative. Any service can claim rigour; only one applying a filter has a list of exclusions.
The raw numbers requirement is the practical protection. With numbers and units you can read the panel yourself, take it to a clinician, or compare it to a previous laboratory. With a score you can do none of those, and you are dependent on the interpretation you were given.
The Separation Worth Buying
The alternative to a combined service is not a single provider but a deliberate separation of functions.
Measurement from a laboratory, ideally the same one clinicians use, returning raw numbers with units and assay-specific reference ranges.
Interpretation from someone without a catalogue. A clinician appointment specifically to review results, or your own reading, which is genuinely learnable.
Products from wherever is cheapest and best verified, chosen against the trial form and dose, third-party tested, single ingredients.
Record-keeping in a format you control, since the record is the asset that compounds and a provider-held record ends when the subscription does.
That separation costs more effort and usually less money, and it removes the structural skew entirely. It is also how medicine has traditionally been organised, with prescribing separated from dispensing in many systems for precisely this reason.
The trade-off honestly stated: combined services are more convenient, and convenience has real value for people who would otherwise do nothing. The separation is better and requires more from you.
What Independent Does Not Mean
Independence is not a guarantee of quality, and overreading it is its own error.
An independent laboratory can still run a poorly chosen panel. Analyte selection is your responsibility if you order it yourself.
A clinician can be wrong, and clinicians vary in familiarity with apoB, lipoprotein(a) and fasting insulin, which are still not routinely ordered in many settings.
Your own interpretation can be wrong, and confidently so. Reading your own panel is learnable and the failure modes are real: over-reading a single flagged value, under-reading within-range drift, missing the pattern, and not accounting for draw conditions.
Independent product sources vary enormously in quality, and buying from a marketplace rather than a curated service shifts the verification burden to you.
Free advice has incentives too, including engagement, ideology and simple confidence unsupported by evidence.
The realistic position: separation removes one specific and predictable skew, and it introduces the burden of doing the work. It is better for anyone willing to do it and not automatically better for everyone.
A Practical Test You Can Run
Before trusting any interpretation, combined or not, three checks are available to you.
Check whether the reference range used is standard. If a result is flagged against a narrower optimal range, note that and reinterpret against the laboratory range. Both readings can be informative; only one is standard.
Check whether the recommendation follows from the finding. A raised hs-CRP with central adiposity and raised triglycerides points at adiposity, sleep and activity. A recommendation for three anti-inflammatory supplements instead is addressing the marker rather than the cause.
Check what was not recommended. If the panel showed low ferritin in a man and the response was iron supplementation without mentioning that the cause needs investigating, the interpretation is incomplete in a way that matters.
And one general check: does the report's list of suggested actions look like the provider's product range? If the correlation is high, the interpretation is at least partly a catalogue.
Running these takes minutes and it converts a passive report into something you can evaluate, which is the practical answer to the whole problem.
The Broader Principle
This is a specific instance of a general rule worth applying across health purchases: separate the party that measures from the party that profits from the answer, or at minimum know when they are the same.
It applies to a testing service selling supplements, a clinic selling procedures it recommends, a device company producing the score it optimises, and a platform whose subscription depends on continued findings. None of those arrangements is fraudulent and all of them contain a predictable pressure.
The protections are the same in every case: insist on raw data, check whether the recommendation follows from the finding rather than from the catalogue, note whether the source ever concludes against a purchase, and keep the record in a format you control.
Applied consistently, that turns a set of provider relationships into something you can audit, which is a considerably better position than trusting any single one of them.
The AEONNN Perspective
AEONNN is subject to this criticism and should be assessed against these counter-signals rather than exempted from them. The platform recommends removal, concludes that no change is needed where that is the answer, and publishes an exclusion list, which is what this Journal's negative conclusions amount to.
The precision matters, though. The skew in a combined model does not enter the assay, which is objective, but the panel design, the reference range chosen, the resolution of borderline results, what is omitted and the retest cadence. Each is a judgement call, and judgement calls resolved consistently in one direction across thousands of reports produce a pattern.
The platform's practical protection is the one it insists on generally: raw numbers with units, so a member can read the panel themselves, take it to a clinician or compare it to another laboratory. With a score they can do none of those. And the honest statement of the alternative is that separating measurement, interpretation, products and record-keeping removes the skew entirely, costs more effort and usually less money, and is better for anyone willing to do it.
Pillar Matrix mapping
Database Matrix layers
- Meta / Consensus Layer (JAMA, BMJ, specialty society positions)
- Quality / Formulation Layer (ConsumerLab, Labdoor)
- Evidence Layer (PubMed, Cochrane, ClinicalTrials.gov)
- Regulatory Layer (EFSA, FDA, EMA)
Frequently Asked
Where does bias enter a combined testing and supply service?
Not in the assay, which is objective, but in panel design, the reference range used, the resolution of borderline results, what is omitted and the retest cadence. Each is a judgement call.
What are the strongest counter-signals?
Whether a service ever recommends removing something, ever concludes no change is needed, and publishes a list of things it recommends against. All three cost the provider something.
Why do raw numbers matter so much?
With numbers and units you can read the panel yourself, take it to a clinician or compare it to a previous laboratory. With a score you can do none of those.
What is the alternative to a combined service?
Separating the functions: measurement from a laboratory, interpretation from someone without a catalogue, products from wherever is best verified, and record-keeping in a format you control.
Does independent mean better?
Not automatically. An independent laboratory can run a poorly chosen panel, clinicians vary in familiarity with apoB and fasting insulin, and your own interpretation can be confidently wrong.
How can I check an interpretation?
Check whether the reference range used is standard, whether the recommendation follows from the finding rather than the catalogue, and what was not recommended when it should have been.
What is the general principle?
Separate the party that measures from the party that profits from the answer, or at minimum know when they are the same, and insist on raw data either way.
Evidence and review
Any dosage ranges cited here reflect the ranges used in published human trials, not personal recommendations. Evidence in this field moves, so this article is reviewed quarterly and carries its last-updated date above. Nothing here is intended as medical advice, and supplementation should be discussed with a qualified clinician, particularly alongside prescribed medication or an existing condition.