How we analyse reviews
Almost every doctor on popular platforms is rated 4.8–5.0. So we ignore the stars, read every review in full and count what it actually says.
1. Every review gets tags
We read the full text and tag it. Reviews with no text and reviews with generic praise only (“liked everything”, “competent doctor”) are not counted as substantive — they weigh 0.
| Tag | Meaning | Weight |
|---|---|---|
| R | result described: got better, it helped, diagnosis confirmed | 3 |
| P | returning patient | 2 |
| W | came by referral | 2 |
| X | visit specifics: examination, explanation, plan, tests | 2 |
| C | specific complaint about the doctor | 2 |
| H | negative marked “Situation resolved” | 2 |
| L | non-medical remark: queue, price, service | 1 |
| U | result not known yet | 1 |
| G | generic praise only | 0 |
| E | no text | 0 |
2. Review weight
A review weighs as much as its strongest tag. A review with a treatment result and a return visit weighs 3. A detailed complaint weighs as much as detailed praise, so criticism does not get lost among thank-yous.
3. Doctor metrics
- Shares of results, returning or referred patients, specifics and generic praise are calculated from reviews with text.
- Complaints are all reviews with a specific complaint, including those marked “Situation resolved”.
- A burst is 3+ reviews in one day or 5+ in 10 days, when that is at least three times the doctor’s usual pace.
- A repeated name is the same name on one day, or a name behind at least 8% of a doctor’s reviews.
4. Verdict: compared with others
The overall score from 0 to 100 has two equal halves. Quality is the average review weight (positive substantive weight minus critical weight, divided by reviews with text), blended with 10 “average” reviews from the sample, so a doctor with few reviews stays near the average instead of jumping up on two or three good ones. Quantity is how many substantive reviews the doctor has (each complaint subtracts two), on a log scale. We only compare doctors of the same specialty in the same city with at least 5 reviews with text. The top third is “Above others”, the bottom third “Below others”, the rest “On par”.
5. Price
Price is assessed separately and does not change a doctor’s rank. We compare the visit price with the sample median; within ±5% counts as typical. Review content and price then give one label, from “Good value” to “Pricey for such reviews”.
| Cheaper | Typical | Pricier | |
|---|---|---|---|
| Above others | Good value | Good value | Pricier, stronger reviews |
| On par | Good value | Typical price and reviews | Pricier, typical reviews |
| Below others | Cheaper, weaker reviews | Weaker reviews, typical price | Pricey for such reviews |
6. What we do not do
We do not claim anyone buys reviews: bursts and repeated names are signals that often have ordinary causes. We do not judge a doctor’s qualifications. We do not copy reviews in full — only a summary and a short quote linking to the original. Author names are shown as on the source site.