Overview

The audit measures one thing: is this review site’s top pick the result of an editorial process, or is it the result of money changing hands? Every criterion in the rubric exists to discriminate between those two possibilities. We do not measure technical SEO. We do not measure traffic. We do not measure design quality. We measure whether the site does the work a patient would assume a review site does on their behalf.

The rubric assigns one hundred points across six criteria. Each criterion is broken into sub-checks with explicit scoring anchors so that scoring is reproducible and not a matter of taste. Inter-rater agreement (Cohen’s kappa) was measured on the full criterion-by-site matrix.

Weighting

CriterionWeight
1. Payment & affiliate disclosure25
2. Provider-selection methodology20
3. Author E-E-A-T20
4. Provider verification rigor15
5. Pricing transparency10
6. Update cadence & corrections10
Total possible100

Payment & affiliate disclosure is the heaviest single criterion at 25%. The justification is empirical: in this category, sites that take undisclosed money from providers tend also to fail every other criterion (no published methodology, no named MD, no verification, no corrections). The single best predictor of a site being a paid placement is whether it acknowledges being one. Weighting disclosure at the top of the rubric encodes that observation.

Criterion 1 · Payment & affiliate disclosure 25 points

Scores whether commercial relationships with audited entities are disclosed where the audited entities are recommended — not on a buried legal page. Five sub-checks, five points each.

Criterion 2 · Provider-selection methodology 20 points

Scores whether the site’s “best of” ranking is reproducible. Four sub-checks, five points each.

Criterion 3 · Author E-E-A-T 20 points

Scores whether the editorial work product is attributable to identifiable, credentialed medical professionals. Four sub-checks, five points each.

Criterion 4 · Provider verification rigor 15 points

Scores whether the site does the work a patient would assume a review site does — checking the underlying provider’s claims rather than repeating them. Three sub-checks, five points each.

Criterion 5 · Pricing transparency 10 points

Scores whether the prices the site lists are the prices the patient will actually pay. Two sub-checks, five points each.

Criterion 6 · Update cadence & corrections 10 points

Scores whether the publication treats its content as something it stands behind over time. Two sub-checks, five points each.

Inter-rater reliability

Each of the fifteen audited sites was scored independently by Dr. Alex M. and Dr. Thrush. A discrepancy threshold of four points on any single criterion triggered adjudication by Dr. Vartanian, the editor, who reviewed both reviewers’ rationales and issued a binding final score with a written rationale appended to the site’s individual review page.

Cohen’s kappa, calculated on the full ninety-criterion-by-fifteen-site matrix, was 0.82 — a level conventionally interpreted as “almost perfect agreement.” Adjudication was triggered on eleven of ninety criterion-cells, or 12.2% of the matrix.

What this rubric does not measure

It is important to state what the audit is not. We do not score whether a comparison site’s top pick is the “correct” top pick. We measure whether the site’s editorial process is one a reasonable patient could trust to produce a trustworthy recommendation. Two high-quality sites could in principle arrive at different top picks. That would be a healthy state of the editorial market. The pathological state is the one we observe in Tier D: low-quality sites converging on the same top pick because they are not independent.

We also do not score technical SEO, GEO/AEO discoverability, schema implementation, llms.txt presence, or any other purely technical signal. Those are real concerns for publishers but they are invisible to the patient. The patient cares whether the recommendation is honest, not whether it is well-marked-up.

Versioning

This is version 1.0 of the methodology, released May 21, 2026. Material changes to the rubric will be versioned. Previous versions will remain accessible at the appropriate version-suffixed URL. Audits will be re-run against the most recent published version on a quarterly cadence, with the next audit scheduled for August 2026.