Reading Clinic Reviews Critically

This page is not about any particular clinic. It is a method, and it applies to every provider in every country.

The finding that settles the website question

Researchers collected every patient testimonial published on the websites of dental practices across a region of England. They found 1,798 testimonials.

All 1,798 were positive. Every venue averaged above four and a half out of five.

That is not evidence that the dentistry was excellent. It is evidence of what a self-published testimonial is: a selection made by the party being reviewed. The authors’ conclusion is the useful one, and it generalises: word of mouth online is often replete with bias, and third-party venues appear less curated and less susceptible to manipulation than practice websites.

Apply that to a treatment costing several thousand pounds, sold across a border, where the provider controls its own site entirely.

Ratings do not track whether people die

This is the sharpest result in the literature and it is worth stating in full.

Researchers took 614 cardiac surgeons in five American states with publicly reported risk-adjusted thirty-day mortality for coronary bypass surgery, and compared those figures against the surgeons’ online ratings.

There was no correlation.

The supporting detail explains why. Average rating 4.4 out of 5. Nearly four in five rated four or higher. Median number of reviews per rating: four. A third of ratings rested on two reviews or fewer.

A second study of surgeons found rating systems correlate with one another but do not accurately reflect physician quality, which is the worst possible combination: consistent and wrong.

The adjudicating systematic review of 28 studies found 7 supporting credibility, 6 finding no correlation, and 15 mixed. Its conclusion is this page’s thesis: rating sites seem credible when relying primarily on patients’ perception, and inadequate to represent the medical quality of physicians.

The problem specific to this treatment

Every method above applies to any clinic. One thing does not.

The people least able to leave a review are the ones for whom it went worst. Someone who died leaves no review. A family in the middle of a bereavement does not write one. Someone who relapsed within a month, and the best follow-up study found seventy per cent relapsed, is rarely writing a review at three months.

Meanwhile the effect on drug use measured across published cohorts is maximal at one month and lower afterwards. So the review written in the euphoric fortnight after treatment is capturing the peak of a curve that comes down, and it stays online permanently as though it were a final outcome.

That is not dishonesty by reviewers. It is a structural bias in when people write, and it points one way.

What the research says about fake reviews

Most advice here is guesswork. One study is not.

Researchers directly observed which sellers were buying fake reviews, rather than inferring it, and then looked at what that produced. Their finding is that the prevalence of fake reviews is arguably higher than ever, and that products buying them cluster tightly in the network linking products to reviewers.

The practical lesson: text is cheap to fake and network behaviour is not. Reviewer histories, timing clusters and the complete absence of negative reviews are more informative than the wording of any individual review.

The rules exist and cannot reach

Since October 2024 American regulation has prohibited fake reviews, buying positive or negative reviews, insider reviews without disclosure, company-controlled review sites, review suppression, and fake indicators of social media influence.

Since April 2025 British guidance has defined a fake review as one that purports to be, but is not, based on a person’s genuine experience, and has covered the whole supply chain including offering services that facilitate them.

Then read what researchers found when they interviewed people trying to enforce against addiction treatment marketing. The practices they named were patient brokering, deceptive marketing, unethical billing and insurance fraud. And the conclusion: without the ability to coordinate across state boundaries, state-level action alone would be insufficient.

If American states cannot police American treatment marketing between themselves, nobody is policing reviews of a clinic in another country.

What to ask instead

Replace sentiment with verifiable fact. Every one of these has an answer that is either true or false.

• Is there continuous cardiac monitoring, and who watches it overnight?
• What medical qualification does the person supervising hold, and where is it registered?
• How long is the drive to a hospital that can treat a cardiac arrhythmia?
• What tests are done before dosing, and what result would mean refusal?
• Has anyone ever been refused, and roughly how often?
• What happens if something goes wrong, in specific terms?

A provider who answers those precisely has told you more than a hundred five-star reviews. A provider who answers with testimonials has also told you something.

Reading iboga reports critically applies the same method to experience reports rather than to reviews.