Evidence and editorial methodology

This page explains exactly how the information on this site is produced, what our labels mean, where our money comes from, and what we do when we get something wrong. If any of it does not match what you see on a page, that is a fault worth reporting.

Who publishes this site

Dental Implants Guide is published by IDAOPS CONSULTING LTD, a UK company registered at Companies House under number 16927251.

We do not describe this site as independent, because it is not. Taki Dent, a dental clinic in Antalya, Turkey, is our disclosed clinical partner, and we may receive a commercial benefit when a patient chooses treatment there through this website. That relationship is stated in the footer of every page and again before any enquiry is passed on.

We would rather tell you that plainly than claim an independence we cannot demonstrate. What we can offer instead is verifiability: a named publisher with a company number, named clinical reviewers with registration numbers you can check, and a source next to each factual claim.

What our labels mean

Every factual statement on an entity page carries a label describing what kind of claim it is and what backs it. There are six.

Verified spec: A fact we checked against a source that can be inspected independently — a materials standard, a regulator record, a company registration, a peer-reviewed finding. The source is listed on the page.
Manufacturer-reported: Something the manufacturer states about its own product. We label these separately because a company describing its own implant is not independent evidence, however accurate it may be.
Clinical evidence: A finding drawn from a published study, cited with its design and follow-up period so you can weigh it yourself.
Derived: A value we calculated from other sourced fields on the same page — for example the narrowest diameter in a published range. We record what it was derived from.
Editorial: Our own explanation or interpretation. It carries no source because it is not claiming to be a specification. It must never contradict a sourced fact on the same page.
Not verified: We looked and could not confirm it, so we left it empty. The page states which of these applies: not publicly verified, manufacturer documentation not located, insufficient evidence, sources conflict, or requires manual verification.

Why a missing value is a feature

The most important rule we work to is that a missing value is acceptable and a fabricated value is not. If we cannot verify the diameters a particular implant is made in, the page says so. It does not quietly print a plausible range.

This is enforced mechanically rather than left to good intentions. Our build refuses to publish an entry that claims a verified fact without a source attached, and refuses to mark an entry as publicly listed unless at least one Tier 1 source supports it. An entry that fails those checks stays in the database but is not published to search engines.

How we rank sources

Tier 1 — primary authoritative: Manufacturer instructions for use and official product documentation, regulatory agencies and government databases, peer-reviewed original studies, systematic reviews, and official clinical guidelines.
Tier 2 — strong secondary: Universities, professional dental organisations, recognised journals, and academic reference works.
Tier 3 — supporting: Reputable dental trade publications and specialist educational resources.
Tier 4 — discovery only: Clinic blogs, distributors, forums and social media. These sometimes tell us which questions patients are asking. They are not permitted to establish a product specification.

How we grade clinical evidence

We do not give implants a score out of ten. A number like "9.7 for clinical evidence" implies a published quantitative methodology, and inventing one would make the site look more precise than it can honestly be. We use defined categories instead, based on how many human clinical studies we can identify for that specific system, their design, their follow-up duration, and whether independent groups have replicated the findings.

One distinction matters more than any other here. Survival and success are not the same endpoint. An implant that is still in the mouth has survived; whether it has succeeded depends on bone levels, soft tissue health, the state of the restoration and the absence of complications. Studies frequently report the first and are quoted as if they proved the second. Where we cite a figure we state which endpoint it measured, over what period, and in whom.

We are also careful not to attribute brand-level evidence to a specific system. A large literature on one manufacturer's older implant line tells you relatively little about a product it launched three years ago.

Extensive: Multiple systematic reviews or meta-analyses, with long-term follow-up and independent replication across research groups.
Strong: Several well-designed human clinical studies including randomised trials or substantial prospective cohorts, with medium-to-long-term follow-up.
Moderate: A reasonable body of human clinical evidence, but limited by follow-up duration, sample size, or reliance on a small number of research groups.
Limited: A small number of human clinical studies, often short-term, retrospective, or manufacturer-funded.
Emerging: Early human data only. The system may be too recent for meaningful long-term evidence to exist yet — which is a statement about time, not about quality.
Insufficient system-specific evidence: We could not identify clinical research on this particular system, as distinct from research on the brand or on implants in general.

How artificial intelligence is used, and how it is checked

Artificial intelligence tools assist our research and drafting workflow. Product specifications and material factual claims are checked against cited sources before publication. Clinical guidance is reviewed according to our editorial and medical review policy.

In practice we run two research passes on each entity. One works from primary sources directly — manufacturer documentation, instructions for use, regulator records and the published literature. A second, independent pass uses a separate AI system that is deliberately not shown the first pass's conclusions, so that agreement between them means something. Where the two disagree, we record the disagreement rather than picking whichever answer is more convenient, and the field stays unverified until a human settles it.

An AI system is never itself treated as a source. Anything it returns must resolve to a real document that we have fetched. We check candidate links automatically and distinguish a page that merely blocked our request from one that does not exist — the second is how a fabricated citation reveals itself.

This is not a theoretical safeguard. While building this site we asked a general-purpose AI model to describe a widely used implant system; it confidently reported the wrong implant level and the wrong connection type, and expressed no uncertainty about either. That is the failure mode the whole process above exists to catch.

Authorship and clinical review

Pages are written by Dental Implants Guide — Editorial. Clinical content is reviewed for accuracy by GDC-registered UK dentists — currently a panel of 24 — each of whom is named with a registration number verifiable on the General Dental Council register.

Authorship and clinical review are deliberately separate roles, and we do not blur them. A reviewer checking that a description of osseointegration is clinically sound has not thereby authored a comparison of two implant systems. No clinician from our commercial partner reviews or authors content on this site.

What we will not publish

A specification we could not trace to documentation.
A survival or success percentage without its endpoint, follow-up period and population.
A claim that one implant system is universally better than another.
A star rating or numeric quality score for an implant or a clinic.
A favourable comparison in exchange for payment. Manufacturers cannot buy placement, and if sponsored content ever appears it will be labelled as such.

Corrections

If something here is wrong, we want to know — from patients, clinicians and manufacturers alike. You can report an error. We verify corrections before applying them rather than accepting them automatically, including corrections from a manufacturer about its own product.

Entity pages record when they were last reviewed. That date means a human checked the entry against its sources on that day — not merely that the file was edited.