ClinicRadar capability
Reputation intelligence
A deterministic Trust Health Score, a separate Coverage Confidence, and a themed breakdown of what the public actually says.
Reputation intelligence answers two questions that a star rating cannot: which theme is moving, and do we have enough public evidence to say so.
Public items are classified into themes from a clinic taxonomy reviewed in English, Turkish, German, and Arabic, so the same complaint is recognised as the same theme whichever language it was written in. The Trust Health Score is then computed by a deterministic, versioned formula over those classifications — the model that did the classifying never touches the number.
Coverage Confidence is reported next to the score, never folded into it. A clinic with two reviews and a clinic with two thousand must never look mathematically comparable, so an Insufficient coverage state suppresses the numeric score entirely.
What it produces
Concrete output
- A Trust Health Score, labelled with the formula version that produced it.
- A Coverage Confidence state shown alongside, and independently of, the score.
- A themed breakdown showing which topics are rising and falling over the period.
- An evidence list per theme, each item linking to the public source with its collection date.
- Score movement over time, comparable because the formula version is pinned.
Boundaries
What this capability will not do
Stated so the limits are known before you rely on it, not discovered afterwards.
- It does not produce a score when coverage is insufficient — the number is suppressed, not estimated.
- It does not let an AI model adjust, weight, or override the score.
- It does not profile reviewers; names are minimised in analytical screens.
- It does not read anything behind a login, and never patient data.
Related