Methodology, Evidence Standards & Author Transparency

Purpose

Measslainte is an independent evidence archive focused on mechanism synthesis in chronic post-viral and post-vaccination illness, immune dysregulation, spike-related pathology, and cellular clearance and recovery pathways.

The site has one methodological commitment: every claim is tied to a primary source, tagged with the strength of evidence behind it, and separated from inference that goes beyond what the source shows.

This page documents how that works in practice. It is the reference articles link to when a reader wants to know what a tag means, how a grade was assigned, or how errors are handled.

For the personal context behind this work (the analyst background, the 2017 MS diagnosis that started it, what the site is and is not), see About. The two pages are separate on purpose: About is the why, this is the how.

Evidence Sources

The default source hierarchy, in descending order of weight:

• Peer-reviewed human studies (RCTs, prospective cohorts, registry analyses, autopsy series) indexed in PubMed.
• Preprints (medRxiv, bioRxiv) when relevant and clearly flagged as preprints. Downgraded accordingly.
• Peer-reviewed mechanistic work (in vitro, animal, pathway-level studies) used to map biology and generate hypotheses. Necessary when human data does not yet exist for a question.
• Authoritative reviews (Nature Reviews, annual reviews, Cochrane) used for orientation, not as a substitute for the underlying primary literature.
• Regulatory documents, FOIA-produced emails, court filings used where they bear on specific claims (manufacturing changes, internal acknowledgments). Treated as primary sources for what they are, not for what they are interpreted to mean.

Search approach:

• PubMed is the default entry point, supplemented by Google Scholar for adjacent literature and PMC for open-access full text.
• 2025 and 2026 data is prioritised where it exists. Older work is cited when it is the definitive source on a question, and noted as such.
• Conflicts of interest at the study level (author funding, industry ties) are flagged in the article when material. The site itself has none to disclose.

Evidence Grading Framework

Every article carries an evidence tier at the top, and individual claims carry inline tags. The system has three dimensions. They answer different questions and are not interchangeable.

Dimension 1: Claim tier (article-level)

A claim tier summarises the weight of evidence behind the article's central thesis. It is a judgement call but it is bounded by the table above.

Dimension 2: Study type (inline badge on a specific statement)

This is the citation-level annotation. A single article will usually cite several study types. The badge on each statement tells the reader what kind of source backs that specific sentence.

Dimension 3: Confidence (inline badge)

Confidence is independent of claim tier. An article tagged [ESTABLISHED] at the top can contain individual statements tagged LOW confidence if the replications are sparse or contradictory on a specific point. The tier summarises the whole; the confidence flag warns the reader where the foundation is thin.

Local tiering

Some articles define local tiers for a specific purpose: test-ordering tiers in the stratification post, compound-priority tiers in compound-focused posts, protocol-intensity tiers in defence protocols. These are scoped to that article, defined on first use, and are not part of the site-wide vocabulary.

Downgrade and upgrade factors

A claim's grade is not mechanical. It adjusts up or down based on the same factors that drive GRADE assessments:

Downgrade for:

• Risk of bias (study design flaws, funding confounds, selective reporting).
• Inconsistency across studies (effect direction or magnitude varies substantially).
• Imprecision (small samples, wide confidence intervals).
• Indirectness (population, intervention, or outcome mismatch with the claim being graded).
• Publication bias (small-study effects, failure to publish null results).

Upgrade for:

• Large magnitude of effect (well above what confounding alone would produce).
• Dose-response gradient (biological gradient demonstrated across exposure levels).
• Plausible confounding would reduce the demonstrated effect (so the true effect is likely larger than observed).

When a grade is moved off the naive baseline by these factors, the article says so. The point of the tag system is not to perform rigour; it is to expose it.

Worked Examples

Real claims from the Spikeopathy Pathways article, with the grading reasoning shown explicitly.

Example 1: Spike protein persists at the skull-meninges-brain axis for up to four years post-infection

Claim. Spike protein is detectable in human skull, meninges, and brain tissue years after SARS-CoV-2 infection, with mechanistic links to neurological sequelae.

Sources. Rong Z, et al. Persistence of spike protein at the skull-meninges-brain axis may contribute to the neurological sequelae of COVID-19. Cell Host & Microbe, 2024. DOI 10.1016/j.chom.2024.11.007 · PMID 39615487.

Grade applied. [HUMAN + MECHANISTIC], badge [AUTOPSY] and [AN], confidence MODERATE.

Reasoning. Peer-reviewed human post-mortem tissue series (skull-meninges-brain axis) directly demonstrates persistence. Upgraded from a pure mechanistic grade by the tissue-level evidence in humans. Downgraded from [ESTABLISHED] because the cohort is finite, persistence in some samples does not establish prevalence in the living population, and causation of neurological sequelae is suggested by mechanism, not demonstrated by intervention. Confidence sits at MODERATE pending larger longitudinal replication.

What would move it. Up: larger prospective cohorts with quantitative persistence assays and matched clinical phenotyping. Down: failure to replicate in independent post-mortem series, or persistence without correlative symptoms in well-powered cohorts.

Example 2: DTI-ALPS-measured glymphatic impairment tracks cognition and fatigue in COVID-recovered patients

Claim. The diffusion tensor image analysis along the perivascular space (DTI-ALPS) metric, an MRI proxy for glymphatic function, is reduced in COVID-recovered patients with cognitive symptoms, and the reduction correlates with symptom severity.

Sources. He C, et al. Dynamic brain glymphatic changes and cognitive function in COVID-19 recovered patients: a DTI-ALPS prospective cohort study. Frontiers in Psychology 16, 1465660 (2025). DOI 10.3389/fpsyg.2025.1465660 · PMID 40330302 · PMC12053492. Supporting: Asymmetrical glymphatic dysfunction in long COVID, BMC Neurology (2025), DOI 10.1186/s12883-025-04133-4.

Grade applied. [HUMAN-IMAGING], badge [PP], confidence MODERATE.

Reasoning. Multiple 2025 peer-reviewed cohorts with consistent direction (DTI-ALPS reduced in symptomatic patients). Not [ESTABLISHED] because DTI-ALPS is an indirect proxy for glymphatic flow rather than a direct measure, cohort sizes are still modest, and the causal direction (impairment drives symptoms versus shared upstream cause) is not yet disentangled. Confidence MODERATE pending larger multi-centre cohorts and longitudinal intervention studies.

What would move it. Up: prospective trials showing clearance-enhancing interventions move DTI-ALPS scores and symptoms together. Down: equivalent DTI-ALPS impairment in unexposed controls with sleep disorders (would make the signal non-specific to spikeopathy).

Example 3: Vaccine-derived spike detected in a small subset of post-vaccine syndrome patients (preprint)

Claim. Vaccine-derived spike protein is detectable in circulation in a subset of patients with chronic post-vaccination symptoms, with the longest documented detection around 700 days post-vaccination.

Source. Bhattacharjee B, et al. Immunological and Antigenic Signatures Associated with Chronic Illnesses after COVID-19 Vaccination. medRxiv preprint, 2025. DOI 10.1101/2025.02.18.25322379v1. Senior authors: Iwasaki, Medzhitov, Putrino.

Grade applied. [HUMAN, PREPRINT], badge [PP], confidence LOW-MODERATE.

Reasoning. The senior authorship is credible and the methods are rigorous, but the work is a preprint with a small cohort (42 participants, 22 controls), assay-threshold sensitivity, and no replication yet. Downgraded twice: once for preprint status (no peer review yet), once for imprecision (small n, wide CIs on any prevalence estimate). Cited because the question matters and the data is the best currently available, but the article explicitly flags this as hypothesis-generating rather than prevalence-confirming.

What would move it. Up: peer review acceptance without material methodological changes, plus an independent replication cohort. Down: failure to replicate in a larger better-powered study, or assay-threshold artefacts demonstrated by independent groups.

These three examples illustrate the range: a tissue-series human claim at MODERATE confidence, an imaging-cohort claim at MODERATE confidence, and a preprint claim at LOW-MODERATE confidence. The tags are doing different work in each case but the discipline is the same.

Falsifiability

Major model articles on this site carry a "What Would Change This Model" section. This is not decoration. The clearance-system failure hypothesis, the spike-persistence framework, and the working subtypes in the stratification post all make specific predictions about what future data should look like if the model is right, and what would weaken or kill it if it is wrong.

For a worked example, see Spikeopathy Pathways → What Would Change This Model. Predictions are grouped into three sets: observations that would strengthen the model, observations that would weaken it, and observations that would kill it outright.

A model that cannot specify what would change its mind is not a scientific model. It is a narrative. The site's policy is to write the former and label the latter.

Sourcing Standard

• Every mechanistic claim cites a PMID, PMC identifier, or DOI.
• Every human claim cites a peer-reviewed source or a high-quality cohort, with preprint status flagged where applicable.
• Every hypothesis is tagged as such, with the inference path exposed.
• 2025 and 2026 data is preferred where it exists; older citations are used when they remain the definitive source on a question, and noted as such.
• Where claims rest on regulatory documents, FOIA-released correspondence, or legal filings, the document is cited for what it literally says, not for what an interpretation layer adds. Interpretation is in the article body and is labelled as interpretation.

Correction and Version Policy

Content on this site is treated as living. Articles carry a date (original publication) and a lastmod (most recent substantive update) in their frontmatter, both visible in the article header.

What triggers an update:

• A cited study is retracted, corrected, or substantially amended.
• New evidence materially shifts the weight behind a claim (upgrade or downgrade).
• A reader raises a substantive objection that, on review, has merit.
• An error of fact or citation is found.

How updates are recorded:

• Minor corrections (typographical, broken links, wording that does not change a claim) are made silently. The lastmod date updates.
• Substantive corrections (a claim is restated, a citation is replaced, a grade is moved) are recorded in a revision note at the foot of the article, with date and reason.
• Major rewrites produce a new dated version and a note explaining what changed from the prior version.

Hypothetical worked correction. Suppose a future multi-centre DTI-ALPS study finds no glymphatic signal in PASC patients versus matched controls, contradicting Example 2 above. The article citing the He et al. result would be updated: the grade would move from [HUMAN-IMAGING] at MODERATE confidence down to [HUMAN-IMAGING, REPLICATED-IN-SOME-COHORTS-ONLY] at LOW-MODERATE, the new study would be cited alongside the original, and a dated revision note would record the change. The original claim would not be deleted; it would be visibly downgraded, and the reasoning would be exposed.

Reporting errors. Substantive errors can be reported via X/Twitter or the contact details on the About page. Constructive criticism with citations gets priority. Complaints without sources are read but do not trigger changes.

Limitations

This site has stated limits and the reader should know them.

• Mechanism focus, not exhaustive systematic review. Most articles are structured syntheses around a working model. They are not Cochrane-style systematic reviews and should not be cited as such.
• Solo operation. Coverage is selective. The site goes deep on clusters where the synthesis is original (spikeopathy, DNA contamination, natural compounds with mechanism overlap) and is thinner elsewhere.
• Judgement is involved. Tagging involves assessment. Two competent reviewers could land on different grades for the same claim. The system exposes the reasoning so disagreement is productive rather than rhetorical.
• Evolving evidence base. Post-COVID and post-vaccine literature is moving fast. What is [HYPOTHESIS] today may be [ESTABLISHED] next year, or may be refuted. The lastmod date is the index of recency.
• Not medical advice. This is repeated in the footer of every page and in every article disclaimer. The site is a synthesis and framework for informed discussion between patients and clinicians who know the literature. It is not a substitute for clinical care.

Independence and Funding

• No industry funding. Self-funded.
• No supplement brand affiliations, paid promotions, sponsored content, or affiliate links.
• No advertising.

Where a researcher or clinician whose work is cited has a disclosed commercial interest (for example, Theoharides and a luteolin/eriodictyol/sulforaphane blend), the article says so when the citation appears. Cited work is cited for the biology, not for the product.

What This Site Is and Is Not

The list on the About page applies site-wide. Short version: this is an evidence archive, research compilation, educational resource, transparency tool, and support for informed consent. It is not medical advice, treatment recommendations, supplement sales, opinion blog, or conspiracy theory site.

Scope of This Methodology

This methodology is itself a living document. If the tagging system is revised, the change will be recorded in the lastmod date and explained on this page. Suggestions for improvement are welcome.