Methodology
A daily reading of evidence — not a daily verdict
Microbiomer turns a defined review of new literature into a concise daily evidence update. The purpose is not to make every paper sound decisive. It is to preserve the distance between what a study reports, what its design can support and what remains unknown.
1. We begin with a defined scope
Each update is based on literature reviewed for a stated date and an established area of interest: food components, the microbiome and health. This boundary makes the work repeatable and the resulting count interpretable. “Today's literature” refers to the dated set reviewed by Microbiomer, not every paper published everywhere that day. We then identify which papers match the defined scope.
Scope is not a quality judgement. A paper can fall outside a daily update because it does not answer the question being reviewed, not because it is poor research.
2. We return to the source record
For every work included in an update, we identify the source and read the source information available to us. We record the population, intervention or exposure, comparison, duration, outcomes and study design when they are reported.
These details stay attached to the finding. “A probiotic worked” is not an adequate description if the strain, population, comparator or measured outcome changes what the sentence means. Where a detail is not reported in the source available to us, we say so or leave it out; we do not reconstruct it from expectation.
3. We describe the result before interpreting it
The wording follows the study design. A randomised trial may support a comparison between assigned groups. An observational study can show an association, but it does not by itself establish that an exposure caused the outcome. A review summarises a body of literature; it is not described as a new intervention trial.
We report direction and magnitude when they are available and relevant. Statistical significance alone does not tell a reader how large or important a difference is. If the magnitude needed for a clear description is not reported in the available source, the update makes that limit visible.
4. We keep different outcomes distinct
Microbiome research often measures several layers at once. Microbiomer separates four reader-facing outcome levels:
- Clinical outcome: how a person feels, functions or experiences health.
- Biomarker: a measured biological indicator, rather than a health outcome itself.
- Microbiome function: what microbial communities or their products are reported to do.
- Microbiome composition: which organisms are present, or their relative pattern.
A movement at one level does not automatically establish a movement at another. Changes in composition, diversity or a metabolite can be scientifically relevant without demonstrating a health benefit. The update names the endpoint that was measured instead of letting the word “microbiome” imply more than the study showed.
5. Editorial selection is not an evidence rating
A daily update is deliberately short. Studies are selected because they help explain the most relevant pattern, contrast or limit in the literature reviewed that day. Being included does not certify a claim; being absent does not reject one.
Study design matters, but no design label produces a conclusion on its own. Population, comparison, duration, endpoint, reporting and consistency with other evidence all affect what may reasonably be said. The daily update therefore remains a reading of that day's studies, not a categorical judgement about a product, ingredient or health claim.
6. Longer-term conclusions require a specified relationship
Evidence accumulates around a particular component × outcome relationship: a defined food component or intervention, considered against a defined outcome. Evidence for one relationship cannot be silently transferred to another. A finding about one strain does not automatically apply to a probiotic class; a change in microbial composition does not automatically become evidence for a clinical outcome.
When Microbiomer considers evidence cumulatively, the conclusion must stay attached to that specified relationship and to the studies relevant to it. On the static launch site, no public categorical evidence state is shown. The site does not turn a single daily update into a lasting verdict.
7. Zero, missing and unavailable are different
These states carry different meanings:
- Zero means the review found none within the defined measure or scope.
- Missing means a detail was not reported or is absent from the available record.
- Unavailable means the source, page or process cannot currently be checked.
Unavailable evidence is not zero evidence. Missing magnitude is not zero effect. A blank field is not an invitation to infer the most likely answer. Where the distinction matters to the conclusion, the update states it plainly.
8. Sources and method versions stay identifiable
Every canonical daily update links to its source records. Its publication date identifies what Microbiomer reported that day. Published updates remain as they were released; they are not silently rewritten under later evidence or a later method.
Potential inaccuracies follow the Accuracy process. We check the report and address the cause in the underlying data, method or safeguards. New evidence enters the current evidence process, while a method change appears in a newly dated method version.
Method version: launch draft 0.1 · 3 September 2026 Accountability: Microbiomer