The dental tourism trust gap

In 1970, George Akerlof published a fourteen-page paper about used cars. The buyer cannot tell, before purchase, whether a car is a good one or a “lemon.” The seller can. Knowing this, the rational buyer assumes the worst, refuses to pay good-car prices, and the good cars exit the market. The market is left with mostly lemons, not because anyone is dishonest, but because the information structure rewards the dishonest and punishes the honest [1] [2].

International dental care is structurally a used-car market with three differences, all of which make it worse. The buyer is a patient, not a consumer. The lemon is a treatment plan, not a sedan. The buyer rarely discovers the failure until eighteen to thirty-six months after they have flown home, by which point the seller has been paid, the relationship is gone, and the only person available to manage the failure is a domestic specialist with no relationship to the original treatment. The patient pays twice. The international clinic pays nothing. The price signal that should have warned the next patient does not propagate, because the next patient is not in the same country and is not searching the same review pool.

This is the trust gap. It is not a problem of bad clinics. It is a problem of a market in which a sophisticated patient (a journalist, an engineer, a clinician’s spouse) cannot, at the point of decision, distinguish a good clinic from a bad one. Every remedy that has been proposed for it leaks. None of them, deployed alone, closes the gap. This piece is an audit of why that is, what each remedy does and does not fix, and the case for the only response that has ever worked in any market with this shape: stacking imperfect remedies rather than waiting for a perfect one.

I write this as the editor of a publication that has built and published its own clinical-standards review framework, and I will say in plain language what that framework is and is not. It is one input, of equal severity to every other clinic the publication has reviewed. It is not, on its own, a solution. The opening of every review I publish includes that admission, and so does this one.

Part one — the structural problem

Why the patient cannot tell

A car has wheels and an engine, and an experienced buyer can drive it. A treatment plan for a full-arch zirconia restoration has neither. It is a written document (fixture brand, abutment type, prosthetic protocol, sterilisation cycle, lab provenance, occlusal scheme, follow-up cadence), most of which a patient is not equipped to evaluate, and most of which a clinic is not obligated to disclose in legible form. The asymmetry is built into the product, not bolted on by bad actors [1].

This is the textbook information asymmetry problem. The seller knows more than the buyer about the quality of the good. The buyer knows this. In financial markets, the response is regulation, audit, and disclosure mandates. In durable consumer goods, it is brand reputation, warranty, and serialised manufacturer accountability. In medical care delivered within one’s home jurisdiction, it is the licensing authority, the indemnity insurer, and the local malpractice system. Each is imperfect; each is a partial remedy.

In international dental care, all of these collapse simultaneously.

• The licensing authority operates only in the country where the treatment was provided. The Australian patient who flies home with peri-implantitis cannot file a complaint with AHPRA against a Turkish clinician, because AHPRA does not regulate practitioners outside Australia [9] [10].
• Indemnity insurance issued in the source country generally does not extend to elective treatment received abroad.
• The malpractice system in the destination country is, in many of the high-volume dental tourism markets, structurally inaccessible to a foreign patient: language, jurisdictional residency requirements, cost of representation, evidentiary standards, and the practical reality that the patient has flown home and the witnesses are 8,000 km away.

So the buyer cannot evaluate the product, and the buyer cannot escalate after the fact. Both arms of the asymmetry-correcting infrastructure are missing at once. This is why a rational patient, browsing twelve clinic websites in five countries, cannot in good faith tell which is good. They are not failing. The market is.

The principal–agent problem, doubled

Inside a single clinic, the dentist is the agent of the patient (the principal). The agent has more information; the principal has the interest. This is the classical principal–agent problem [3], and dentistry, like medicine, has spent a century building professional ethics, licensing, peer review, and adverse-event registries to bring agent and principal into alignment. None of these are perfect. All of them are present in a domestic system.

International dental tourism inserts a second agent. The marketplace, the medical-tourism facilitator, or the marketing aggregator stands between the patient and the clinic, paid by the clinic per booking or per converted lead. The marketplace’s economic interest is to convert the booking. The patient’s economic interest is to convert only if the clinic is right for their case. These are not the same interest, and there is no licensing body, no peer review, and no adverse-event registry that governs medical-tourism marketplaces. There is, in most jurisdictions, no licensing requirement at all for marketing oneself as a medical-tourism facilitator.

I am not arguing that all marketplaces act badly. I am arguing that the structure of the principal–agent relationship between patient and marketplace is unregulated in a way that the dentist–patient relationship is not. When a marketplace’s revenue depends on the booking and not on the outcome, the alignment is broken at the level of the contract, not the level of the personality.

Regulatory arbitrage as a feature, not a bug

The single largest structural driver of the trust gap is regulatory arbitrage: the practice of structuring an activity in the jurisdiction with the most favourable rules [4]. In dental tourism, this is not a bug. It is the central economic premise.

A clinic in a country with lighter advertising rules can publish before-and-after photos that an Australian-registered dentist cannot publish under AHPRA advertising guidelines. A clinic in a country with looser sterilisation reporting requirements is not lying when it says it meets local standards; it is operating in a regime where local standards do not require what Australian or EU standards require. A clinic in a country where short-implant and prosthetic-redesign protocols are not part of the standard-of-care training curriculum is not negligent for not offering them; it is delivering the local standard of care.

The patient flying in from Sydney is purchasing the price of the destination jurisdiction, but assuming the quality regime of the source jurisdiction. The price differential is real. The quality-regime differential is also real. Whether the destination’s regime is, in any given case, equivalent to the source’s is the question the patient is being asked to answer, and the patient is the least-equipped party in the transaction to answer it.

This is not a moral argument against dental tourism. The arbitrage runs both ways. There are international clinics that exceed their source-country domestic standards: better imaging, better fixtures, more time per patient, longer sterilisation cycles than a high-volume Australian general practice. The arbitrage is real in both directions. The patient cannot tell which direction it runs in any specific clinic in front of them.

Why the photos are not evidence

Every dental tourism clinic website features before-and-after photos. They are uniform across the industry and uniform in their structure: a smile that is darker, asymmetric, or shorter on the left, replaced by a smile that is brighter, regular, and longer on the right. This is, I am sorry to say, not evidence of anything except the existence of a camera and a lab.

A before-and-after photo cannot show:

• whether the prosthetic emergence profile was managed correctly, the variable that most strongly predicts long-term peri-implantitis incidence;
• whether the bite has been registered in centric relation or in a convenience occlusion that will overload the posterior fixtures;
• whether the lab work was milled to the tolerances the fixture brand specifies, or to the tolerances the fastest local lab can deliver;
• whether the radiographs at twelve months will show marginal bone loss;
• whether the patient is comfortable, or whether they are smiling because they are at the end of a holiday and the alternative (telling a stranger with a camera that something feels wrong) is socially expensive.

I have retreated cases where the patient sent me their before-and-after photo with their referral, and the photo looked good. The radiograph did not. The two artefacts told entirely different stories about the same case. The marketing artefact and the clinical artefact have different incentives behind them, and the patient is shown the marketing one because it is the one a clinic can produce on a phone.

The follow-up problem

The most subtle structural feature of the trust gap is that the failure mode is delayed. A poorly seated implant does not fail in week three. It fails in month eighteen. A poorly registered occlusion does not produce TMJ symptoms on day five; it produces them at month nine. A missed canal does not cause a periapical lesion at the post-op visit; it causes one at twenty-four months.

By the time the failure is visible, the relationship between patient and treating clinic has structurally ended. The patient is in another country. The follow-up appointment, if it occurs at all, is a phone call. The treating clinic has limited financial incentive to absorb the cost of a redo, because the redo is not visible to the next patient. The next patient is in another country too, and is searching a review pool that does not include the patient who flew home with bone loss eighteen months later. The price signal that ought to warn the next patient does not propagate. This is why the market does not self-correct over time the way Akerlof predicted some markets would: the signals from failed cases never reach prospective buyers in legible form.

This is not unique to dental tourism. It is the same structure that made early online consumer marketplaces (pre-eBay-rating, pre-Amazon-review) so prone to lemons. What dental tourism lacks is the infrastructure that fixed those markets. There is no equivalent of an eBay seller rating that survives across platforms, no equivalent of a Carfax history report, no equivalent of a Wirecutter that is paid by no one in the transaction. There are reviews on Google Maps and TrustPilot, and we will discuss in part two why those don’t work.

Summary of part one

The patient cannot evaluate the product, cannot escalate after the fact, is paired with an unregulated second agent in the form of the marketplace, is purchasing across a regulatory arbitrage gap they are not equipped to assess, is shown evidence (photos) that is not evidence, and discovers failure too late and too far away for the price signal to propagate. Every one of these is structural. None is a problem of bad clinics. The trust gap is a function of the market’s design, not its conduct.

This is what makes the problem hard. A bad-clinic problem can be solved by removing the bad clinics. A market-design problem requires market-design responses, and every available response leaks.

Part two — the limits of every remedy

The question of what to do about the trust gap is in some ways the only useful question. There are seven categories of remedy currently in circulation. I will say what each does, what each fails to do, and where each leaks. I include the publication’s own clinical-standards framework in the audit, with the same severity it applies to others.

Remedy 1: marketplace ratings

Google Maps reviews, TrustPilot, country-specific platforms, and the medical-tourism marketplaces themselves all run rating systems.

What they fix: they create a public artefact of patient experience. They surface acute negative experiences quickly. They add reputational cost for clinics that are unambiguously rude, unambiguously unhygienic, or unambiguously commit billing fraud.

What they don’t fix: rating systems measure the variable the patient can evaluate (the experience of receiving treatment), not the variable that determines outcome (the technical quality of the treatment). A patient who flew home twelve months ago, has not yet developed bone loss, and remembers a kind receptionist will write a five-star review. The rating that arrives is uncorrelated with the outcome that has not yet failed.

Worse, the high-volume rating platforms have proven structurally manipulable. Review fraud, incentivised reviewing, and the simple statistical reality that the vocal minority of reviewers is not representative of the patient population mean that aggregate marketplace ratings carry a known signal-to-noise ratio worse than randomised. I am not aware of any peer-reviewed evidence that aggregate online ratings predict five-year clinical outcomes in any healthcare service category.

Remedy 2: certifications and accreditations

International accreditation schemes (JCI, ISO, country-specific dental association memberships) issue badges that clinics display.

What they fix: they create a baseline floor. A clinic that has passed a real audit against a published standard is, on average, less likely to have basic process failures than a clinic that has not.

What they don’t fix: the floor is below the relevant question. JCI accreditation, for example, audits hospital management and patient-safety processes; it does not audit clinical-decision quality at the level of “should this patient have been recommended a sinus lift versus a short implant.” A clinic can be JCI-accredited and routinely overprescribe procedures the patient does not need, because overprescription is not a JCI-failure mode. ISO certification is similarly process-focused.

There is also the meta-problem that some certifications are paid-membership badges with thin or no audit, indistinguishable visually from rigorous certifications. The patient is shown a logo and is not equipped to research the standard behind it. (The WHO’s framing of quality of care is helpful here: it explicitly distinguishes the structural, process, and outcome dimensions of quality, and observes that most accreditation systems audit the first two and assume the third [7] [8]. The trust-gap problem is in the third.)

Remedy 3: destination-country regulators

Some destination countries have credible regulators. Some do not. Even where the regulator is credible, the patient cannot access it.

What they fix: they remove the worst offenders from the market, slowly. A clinic operating without a license, falsifying credentials, or running re-used instruments will, eventually, attract regulatory attention.

What they don’t fix: the regulator’s complaint mechanism is generally inaccessible to a foreign patient. Filing a regulatory complaint in a country where you are not a resident, in a language you may not speak, against a clinic that has been paid, with documentation that is in the language of the source country: this is not a remedy a normal patient can deploy. AHPRA itself documents the equivalent inverse problem: it does not regulate Australian practitioners’ conduct outside Australia, and it cannot act on practitioners not registered with it [9] [10].

Remedy 4: source-country regulators

The source country’s regulator (AHPRA in Australia, GDC in the UK, similar authorities elsewhere) governs domestic practitioners and continuity of care after return.

What they fix: they regulate the receiving end. A domestic dentist who provides post-treatment care to a returned international patient is bound by the source-country standards on documentation, advice, and clinical conduct.

What they don’t fix: they explicitly do not regulate the original treating clinic. AHPRA’s published scope is “registered health practitioners in Australia” [10]. A patient harmed by treatment received outside Australia has, from AHPRA’s perspective, received treatment that is outside AHPRA’s jurisdiction. The patient’s only AHPRA recourse is against the Australian clinician who provided post-treatment care, which is, in most cases, not the source of the original harm. The structural gap is named in the regulator’s own scope description.

Remedy 5: independent specialist review

This category, of which the publication you are reading is one example, promises to act as an Akerlof-correcting third party: paid by no one in the transaction, applying a published clinical-standards framework with the same severity to every reviewed clinic, sourcing every claim, disclosing every relationship in paragraph one.

What it fixes: it produces a reviewable artefact for cases that have been formally reviewed. It models, publicly, what a defensible review looks like. It documents specific gaps and specific strengths in named clinics under conditions where the patient can verify the methodology, the disclosures, the source list, and the absence of paid placement.

What it doesn’t fix:

1. Coverage. Even an aggressive specialist-review publication can review tens of clinics per year, not the thousands of clinics operating in international dental markets. Most clinics will never be reviewed. A patient choosing among unreviewed clinics has not been helped.
2. Recency. A clinic reviewed in 2026 is not the same clinic in 2029. Practice-acquisitions, lead-clinician departures, protocol drift, and ownership changes can all materially alter clinical quality without any visible signal. We commit to re-reviewing every reviewed clinic on a documented cadence, but the cadence is annual at best, and a clinic that changes hands in February is misrepresented until November.
3. Specialist coverage by topic. A specialist endodontist reviewing an oral surgery practice is reviewing across the boundary of her own specialty. Multi-specialty review boards mitigate this, but every reviewer’s coverage is bounded by their training. A review framework that is rigorous in implant prosthetics may be thin on orthodontic outcomes, and vice versa.
4. Reader access. A defensible review only helps the patient who finds it. SEO, AI citation, and direct distribution all leak. A patient who Googles a clinic name and lands on a marketplace listing before they land on the review has not been reached.
5. The reviewer is a single point of failure. If the reviewer’s framework is wrong, every review is wrong. If the reviewer’s disclosures are incomplete, every review is suspect. The publication’s defence against this is published methodology, public sourcing, named corrections, and the editorial integrity ratchet that pairs every affiliated-entity review with a non-affiliated review of equal severity within seven days. None of these defences is enough on its own.

I do not list this remedy first, and I do not list it as the strongest remedy. I list it as one remedy among several, with named limits.

Remedy 6: domestic clinician second opinion

The most underused remedy is the simplest. A patient with a quote from an international clinic shows that quote, and the supporting CBCT, to a domestic specialist, who reviews the treatment plan and gives a structured second opinion. Some patients do this. Most do not, because it is expensive, because the local specialist may be perceived as biased, and because the patient has often already booked.

What it fixes: it inserts a credentialed agent into the patient’s principal–agent chain whose interest is not the booking. A domestic specialist reviewing a treatment plan can see whether a sinus lift was justified by the imaging, whether short implants were considered, whether the prosthetic redesign was offered. This is the highest-leverage protective step a patient can take.

What it doesn’t fix: the consultation costs $200–400 AUD. The patient has often committed to dates. And a small minority of domestic clinicians will, when asked to review a quote from an international clinic, give a review that is biased against the international plan because of perceived competitive interest. This is rare in my professional experience but it is real. The remedy is structurally sound; the implementation is variable.

Remedy 7: patient-side education

The publication has invested in patient-education content because educated patients ask better questions. Frameworks for evaluating an implant treatment plan that includes bone grafting [link below], frameworks for evaluating a failed root canal recommendation [link below], cost references with what’s-included specifications [link below]: all are designed to give the patient the structural questions to ask.

What it fixes: it raises the floor of patient knowledge. A patient who knows to ask “was a short implant considered?” will, in most clinics, get a more honest answer than a patient who does not.

What it doesn’t fix: even an excellent patient cannot run an audit. The patient who asks the right questions can detect the most blatant misalignment, but cannot detect a moderately well-trained sales process, cannot evaluate the actual quality of the proposed work, and cannot tell a clinic that has updated its protocols against the current literature from a clinic that has the right answers cached for the right questions. Patient education is necessary; it is not sufficient.

The honest conclusion

No single remedy closes the gap. Every one of them leaks at a different point. The marketplace ratings leak on outcome correlation. The certifications leak on what they audit. The destination regulators leak on accessibility. The source regulators leak on jurisdiction. The independent specialist review leaks on coverage and recency. The domestic second opinion leaks on cost and patient willingness. The patient education leaks on what a patient can verify. The trust gap is not closed by any of them on its own.

What does, in markets with this shape, demonstrably reduce the gap is the stacking of imperfect remedies. Each one’s failure mode is partially covered by another’s success mode. A patient who uses a published specialist review and obtains a domestic second opinion and asks the structural questions from patient-education content and checks that the destination clinic holds a real (not paid-membership) certification and documents in writing what is and is not included in the price has not closed the gap. They have stacked five imperfect remedies. The remaining residual is what no system, including this publication, can fix on its own.

Stacking is not a satisfying answer. It is more work than the patient should have to do. The fact that it is the honest answer is part of what we owe the patient. The alternative, telling the patient there is one source they can rely on without verification, is what the marketplace marketing copy says, and it is wrong.

What would change my view

I hold this position because of (a) the structural features of the market documented in part one, (b) the observed failure modes of each remedy in part two, and (c) my own clinical caseload of returned-from-overseas cases over the last several years. The evidence that would update it:

• A peer-reviewed cohort study, multi-country, N>2,000, comparing five-year outcomes of patients who used a single remedy (any of the seven above) against patients who stacked three or more, with implant survival, marginal bone loss, peri-implantitis incidence, and patient-reported outcomes as endpoints. No such study exists. The closest comparators are single-country surveys of patient experience, which do not measure outcome.
• A demonstration that a major international accreditation scheme has begun auditing clinical-decision quality (the third dimension in the WHO quality framework [7]) with a published methodology and a public list of decertifications. None currently does, to my knowledge.
• A demonstration that an international medical-tourism marketplace operates with a structural separation between marketing revenue and review-publishing function: a real Chinese wall, with audited compliance and visible disagreements between the two arms. None currently does, to my knowledge.

If those exist, the audit above shifts. Until they exist, the structural argument stands.

What you can do, today, if you are the patient

Stacked remedies, in practical order:

1. Obtain a domestic specialist second opinion on the treatment plan and the CBCT. This is the single highest-leverage step.
2. Ask the structural questions documented in the implant-without-bone-grafting framework and, for endodontic cases, in the framework for failed root canal treatment. A clinic that answers them clearly is providing information you can act on; a clinic that deflects them is providing information of a different kind.
3. Verify cost figures against the publication’s implant cost-by-country reference, with currency, date, and what’s-included specified. If a quote is materially below or above the published range with no documented reason, that is a question worth asking.
4. Read the procedure reference for endodontic retreatment or the equivalent for whichever procedure is being recommended, before agreeing to the plan. If the plan does not match the procedure reference’s standard-of-care description, you have something specific to ask.
5. Confirm the destination clinic’s certification is a real audit, not a paid badge. Read the standard the certification claims to enforce. If the standard does not match the question that matters to your case, the certification is not protective on the dimension you need.
6. Read the publication’s methodology page before relying on any of its reviews. The methodology is published precisely so you can decide whether to trust the review. A review framework that does not publish its methodology should not be relied on, including this one.
7. Confirm the publication’s disclosure status for the clinic, marketplace, or manufacturer mentioned in any review you read here. The default state is no commercial relationship; if a relationship exists, it is documented in paragraph one of the relevant piece.

None of these alone closes the gap. Together they reduce it. The reduction is real and is, in my professional view, worth the work, both because it materially lowers the probability of a failed treatment, and because the act of stacking remedies is itself an education in the structure of the market you are buying into.

The trust gap is not a problem the patient can solve. It is a problem the patient can navigate. The difference is the difference between a closed gap and a managed one, and the managed version is the most honest answer this publication can offer.

I do not know how long it will take for the structural problems described in part one to be solved. I suspect, but cannot prove, that they will not be solved by any of the parties currently in the market — clinics, marketplaces, certifiers, regulators — because the alignment is wrong at the level of the contract for each of them. I suspect the most likely path to a real solution is some combination of source-country regulators extending continuity-of-care standards, peer-reviewed five-year outcome registries that are open across borders, and the slow accumulation of independent specialist review at sufficient density that AI engines and search engines surface the review before they surface the marketplace. Each of these is partial. Each is also possible.

Until then, the patient is the integrator. The publication’s role is to make the integration easier, to publish the framework, to disclose the relationships, to admit the limits, and to keep doing it under a byline that has a license on the line. That is the work. It is not enough, on its own, to close the trust gap. It is, when stacked with the other imperfect remedies, the most honest available response.