E-E-A-T Amplification Per Content Piece: The 12-Signal Healthcare Checklist
E-E-A-T is not a vibe. It is a measurable set of on-page, schema-level and infrastructure-level signals. Most healthcare content teams stop at "we put a doctor byline" — which is one of twelve. This article documents the per-piece checklist we run inside Content HQ before a healthcare article is allowed to publish.
What Google actually evaluates in 2026
Two updates matter. First, the rater guidelines now distinguish Experience from Expertise — the extra "E". Experience means "has the writer personally lived through or directly observed this?" — relevant for patient stories, doctor case notes, before-and-after photography with consent. Expertise means formal credentialing. Both are needed for healthcare; neither substitutes the other.
Second, the AI Overview citation system filters aggressively for trust signals before it pulls a snippet. Watching our own GSC API data across 31 client domains, the correlation between trust-signal density and AI Overview pick-up is the single strongest variable — stronger than backlinks, stronger than freshness.
The 12-signal per-piece checklist
Every healthcare piece in Content HQ runs this checklist. The compliance gate blocks publish if the score is below 9/12.
| # | Signal | Pillar | How Content HQ enforces |
|---|---|---|---|
| 1 | Author bio with photo | Authority | Mandatory writer profile attached to ticket |
| 2 | Medical reviewer name + NMC reg number | Expertise | Reviewer field locked; ticket cannot publish without it |
| 3 | Last-reviewed date visible on page | Trust | Auto-stamped from publish event |
| 4 | Source citations (ICMR / WHO / peer-review) | Expertise | Minimum 3 sources for medical claims; less for cosmetic |
| 5 | MedicalWebPage / MedicalCondition schema | Trust | Auto-injected based on content type |
| 6 | Author schema with sameAs | Authority | Person schema with LinkedIn + hospital profile |
| 7 | First-person clinical observation | Experience | Editor flag — at least one "in my practice" or case anecdote |
| 8 | Conflict-of-interest disclosure | Trust | Pre-approved disclosure block per specialty |
| 9 | External authoritative outbound link | Authority | At least one link to ICMR, WHO, NHA or peer-review source |
| 10 | Patient cohort specificity | Experience | Brief mandates age band, severity, prior history |
| 11 | Risk + limitation paragraph | Trust | Pre-approved block from claim library, not writer-authored |
| 12 | Doctor-led video or audio embed | Experience | Optional but scored — pushes piece above 10/12 |
The Experience "E" — why this is the new bottleneck
Most healthcare content fails on Experience, not Expertise. Writers can cite ICMR. They cannot make the page sound like a doctor sat down and wrote it. The fix is operational, not stylistic.
For each piece, we mandate a 20-minute "clinical capture" call — the writer interviews the named reviewer and extracts two to three first-person observations. Things like "in my last 200 IVF cycles, the most common cause of failed implantation in the 36-40 cohort was undiagnosed endometritis" or "ortho patients who skip pre-op physio rehab take 3.4 weeks longer to return to walking". These observations cannot be googled. They are what make a piece feel earned. They are also what Google's AI Overview increasingly cites — because the observation is unique to your domain.
Schema isn't garnish — it's structural E-E-A-T
For healthcare, MedicalWebPage and MedicalCondition schema markups are the machine-readable version of trust. Author schema with sameAs links to LinkedIn, hospital staff page and PubMed author profile makes the credential graph traversable for Google. Content HQ injects these automatically based on the piece's specialty and reviewer assignment — the writer never sees them, and cannot break them.
A multi-city dermatology chain we audited had 240 published pieces with strong content and zero MedicalWebPage schema. After a six-week schema retrofit — no content changes — AI Overview citations grew by 280 percent. Schema was the missing leg.
The signals you should not fake
Two warnings. First, do not invent reviewers. Google now cross-references NMC numbers, LinkedIn profiles and hospital staff pages. A reviewer name that does not match a verifiable credential is worse than no reviewer name — it is grounds for manual action. Second, do not stuff schema with claims the page does not support. A MedicalProcedure schema declaring "outcome: 95 percent success" without the same claim being verifiable in the body sends a contradiction signal that suppresses the page.
E-E-A-T is not a checklist you tick once. It is a per-piece operational discipline. The pieces that compound rank are the pieces where every one of the 12 signals is real, verifiable, and consistent across the body, schema and metadata.
How E-E-A-T integrates with the per-piece score
Inside Content HQ, the E-E-A-T score is one of the seven dimensions of the per-piece performance score described in our earlier essay. A piece scoring 12/12 on E-E-A-T but failing on conversion is rewritten with a new CTA. A piece scoring 10/12 on E-E-A-T and 8/10 on engagement is the cluster's tentpole — that pattern is reverse-engineered into the brief template for the next ten pieces. Discipline compounds when the team can see which signals correlate with which outcomes.
An anonymised oncology platform we work with took their average E-E-A-T score from 6.4 to 10.1 in 90 days. Their AI Overview citation rate moved from one piece per fortnight to one piece per three days. No new traffic source. Same writers. Same publish cadence. The only change was the gate.
Audit your current pieces against the 12-signal checklist
We will run 20 of your published pieces through the E-E-A-T audit, score each one, and ship a per-piece remediation list in week one.
Book a free audit → WhatsApp ICG