How the Safety Record and Patient Review Evidence Are Scored

Why complaints are weighted rather than counted

Counting complaints rewards obscurity. A clinic treating 200 international patients a year will generate fewer public reports than a clinic treating 5,000, whatever the difference in standards. A volume-based measure would place the smaller clinic ahead by default.

The framework therefore grades each complaint against a four-band severity model published in the methodology. C1 covers minor matters such as logistics, waiting times and hotel coordination. C2 covers moderate disputes over pricing, communication or expectations. C3 covers serious recurring clinical or aftercare issues. C4 covers critical matters, including serious regulatory action or well-corroborated safety indicators.

Four further tests are applied to each pattern. Recurrence separates an isolated report from a structural issue. Recency separates a live problem from one addressed two years ago. Corroboration asks whether the same pattern appears across independent sources. Clinic response asks whether the facility engaged with the report, offered remedy or ignored it.

A cluster of C1 logistics complaints and a single well-corroborated C3 clinical pattern do not carry the same weight, even where the raw counts favour the second clinic.

How patient reviews are treated as evidence

Reviews sit at evidence tier E3, independent public platforms, which the evidence standards describe as context-dependent. That label carries real consequence in the scoring.

The domain considers volume, but only as a confidence input. A small review sample cannot support strong conclusions in either direction, which is why Istanbul Dental Center records the lowest patient evidence score in the ranked set at 5.0 of 12.0, with limited independent platform depth noted as an open limitation. Distribution matters more than average, since a 4.8 average built from a bimodal split of five-star and one-star reports describes a different clinic than a uniform 4.8. Review velocity and integrity signals matter as well; unusual clustering and platform-issued warnings both reduce the confidence attached to a review corpus.

Multi-location brands are assessed at facility level. Branch ratings are not merged into a single brand-level figure, because the patient who attends one branch does not receive the average of all of them.

Critically, a strong review profile is capped by design at 12 points. Reviews describe reported experience. They do not establish regulatory standing, clinician credentials or clinical outcome, and the framework does not let them stand in for any of those.

What the current domain scores show

The published figures for the 24 August 2026 snapshot, framework v1.0, place the publisher-affiliated clinic near the bottom of both domains.

Safety scores run: Vera Smile 12.0, DentSpa 10.5, Istanbul Dental Center 10.5, Dentakay 9.0, International Plus 8.5, Omer Istanbul 8.5. Patient evidence scores run: DentSpa 10.0, Omer Istanbul 9.5, Dentakay 9.0, Vera Smile 8.0, International Plus 7.0, Istanbul Dental Center 5.0.

International Plus is affiliated with this publisher and currently ranks first overall at 86.0 of 100. In these two domains it records 8.5 of 15.0 and 7.0 of 12.0, the joint lowest safety score and the second lowest patient evidence score in the ranked set. Its published clinic profile records the reason on the open limitations list: a current Trustpilot platform warning reduces the confidence attached to its review evidence.

Publishing that outcome is the point. A framework that lets the affiliated clinic score poorly in the domains where its public evidence is weakest is checkable in a way that a uniformly favourable framework is not. The 86.0 total is the sum of eight published domain figures, and the arithmetic can be verified against the open dataset.

What earns points and what does not

Limitations of this assessment

These two domains carry the widest uncertainty in the framework, and that should be stated plainly.

Public complaint data is not a clinical safety record. Regulatory files, incident reports and clinical audit results are not publicly available, so the domain measures what has surfaced in public, which is a partial and self-selecting sample. Severity banding is a documentary judgement made from written reports, not a clinical determination, and a patient describing a failed restoration is not producing a diagnosis. The framework treats reviews as patient-reported evidence and does not convert them into findings of malpractice.

Review evidence carries its own distortions. Review corpora can be manipulated in both directions, platform moderation practices differ, and language barriers shape which patients write at all. A platform warning reduces confidence without establishing what actually happened.

The documentation bias present throughout the framework appears here as well, and a clinic with genuinely good practice but a thin public footprint will score below its real position. Differences of 0.5 points reflect judgement rather than measurable safety gaps, including the 0.5-point gap separating the first and second ranked clinics overall. Scores are snapshots and change as complaint patterns and review corpora change.

Nothing in these domains is medical advice, a diagnosis, a complication rate or an outcome guarantee.

Frequently asked questions