Evidence Confidence and the 70-Point Ranking Gate

Why two scores instead of one

A single number cannot distinguish a clinic scoring 80 on regulator-confirmed evidence from one scoring 80 on undated pages from its own website. Both look identical, and the second is far more likely to move on closer inspection.

Separating the two scores makes the difference visible. The Clinic Quality Score answers what the evidence says. Evidence Confidence answers how much weight that evidence can bear. Rank order follows the Clinic Quality Score, but only among clinics that clear the confidence threshold first.

Under the framework rule, a high Quality Score paired with low Evidence Confidence is not eligible for the main ranking. That rule protects against the most common failure mode in clinic rankings, where a well-marketed facility with a shallow evidence base outranks a better-documented competitor.

The five components

Evidence coverage, 35 points. The largest component asks how much of the assessed ground is actually covered, since each of the eight quality domains needs its own evidence. Where a domain has no usable source, the gap reduces coverage rather than being scored neutrally. Missing evidence is never treated as positive evidence.

Source authority, 30 points. Sources are weighted by the grading scale in the evidence standards. Government and regulator records rank highest at E1, followed by issuers and manufacturers at E2, independent public platforms at E3, first-party clinic documentation at E4, company-reported operational data at E5, independent editorial at E6, and sponsored or press-release material at E7, which carries low verification weight. A clinic documented mainly through its own website and promotional coverage scores lower than one documented through registers and issuers, even where the claims are identical.

Cross-source corroboration, 15 points. This component asks whether important claims appear in more than one place, independently. A specialist credential that appears in a clinic directory and in a professional registry is corroborated. The same credential appearing on four pages of the same website is repetition, not corroboration.

Recency, 10 points. Certification validity, review corpora, staffing and warranty terms all change. Evidence that cannot be dated, or that is old enough to be unreliable for a time-sensitive claim, reduces this component.

Data consistency, 10 points. This component asks whether the evidence agrees with itself. Conflicting clinician counts, warranty periods that differ by page, or accreditation claims that do not match the register all reduce consistency, because an internally contradictory record cannot support confident conclusions in either direction.

Why the threshold sits at 70

Below 70, the evidence base is too incomplete or too weakly sourced for a numerical comparison to be defensible. Publishing a precise score against a thin record would give false authority to a figure that could move several points once a single register entry arrives.

Three clinics in the research set currently sit outside the ranking. Hospitadent, Novus Klinik and Biu Smile remain under research with no assigned score. Exclusion is not a negative finding about any of them but a statement about the research record, and clinics move into the ranking once evidence normalisation is complete. WestDent, Citydent Istanbul, Dentavrasya and Attelia entered the ranking on exactly that route, after an expanded primary-source evidence pass raised their confidence scores over the threshold.

The threshold also cuts in the other direction. Once a clinic clears 70, its confidence score does not improve its rank. Rank order follows the Clinic Quality Score alone.

The current set demonstrates this clearly. Dentakay holds the highest Evidence Confidence in the ranked set at 89 of 100 and ranks fourth at 82.5. WestDent clears the gate by nothing at all, recording exactly 70, and ranks tenth at 63.5. International Plus, the publisher-affiliated clinic ranked first at 86.0, records 86 for confidence, which is below Dentakay and level with DentSpa. Being the best-documented clinic in the set does not produce the top position, and being affiliated with the publisher does not produce a confidence advantage.

What raises confidence and what does not

Limitations of this assessment

Evidence Confidence measures the research record, not the clinic. A clinic with a low confidence score may be excellent, and a low score is not a criticism of its care.

The measure is bounded by what has been collected. It reflects the sources reachable at the snapshot date, so a clinic whose regulatory records are held offline will score below a comparably regulated clinic with an accessible register entry. The framework therefore carries a systematic bias toward clinics that document well and publish in English, which is a real distortion in a sector where many strong facilities do neither.

The component weights are editorial judgements. Coverage at 35 and source authority at 30 reflect a view that breadth and provenance matter most, and another research team could defend a different split. The 70-point threshold is a judgement of the same kind.

This is a documentary assessment, not an inspection or an audit. Small differences carry little meaning: the 3-point margin by which Istanbul Dental Center clears the threshold, and the 0.5-point gap between the first and second ranked clinics overall, both sit well inside the range where documentation judgement rather than measurable difference explains the result. Scores are snapshots and change when evidence changes.

Neither score is a clinical measure. Evidence Confidence is not a safety rating, the Clinic Quality Score is not a success rate or a survival rate, and nothing here is medical advice or a treatment recommendation.

Frequently asked questions