At 6:40 in the morning in Ohio, an internist's phone buzzes. It is a WhatsApp message from a classmate she has not seen in eleven years, now practising in Lagos.
There is a photograph of a CT scan. A child. A mass. And a question.
Her classmate does not have a neurosurgeon to ask. Not at his hospital, not in his city, and, depending on where exactly he is, possibly not within a catchment of several hundred thousand people.
She is an internist. She is not a neurosurgeon either. But she is in a country that has them, and she knows people, and so she becomes what she has been for years without anyone naming it: a bridge.
She forwards the image to a neurosurgeon she trained near. He looks at it between cases and sends back three sentences. She relays them. Somewhere in Lagos, a plan changes.
No consent was documented. No jurisdiction was established. No record exists. Nobody is responsible for anything. The image is now on four phones in two countries.
And this is, by a wide margin, the most effective mechanism operating in that clinical situation.
The bridge exists at enormous scale
The population that carries this traffic is much larger than most people appreciate.
From the OECD's International Migration Outlook 2025:
- 606,000 foreign-trained doctors work in OECD countries, representing 18.4 percent of all doctors in those countries.
- That figure is up 62 percent since 2010.
- Roughly 89,000 of them trained in countries on the WHO health workforce crisis list.
In the United States specifically, about one in four physicians is an international medical graduate.
Every one of those physicians retains, to varying degrees, a professional network in the country where they trained: classmates, teachers, former colleagues. That network is a standing, high-trust, bilingual, clinically fluent bridge between health systems of radically different capability.
It carries a great deal of traffic already. And nobody has ever built any infrastructure on it.
The scarcity on the other side is severe
To understand why that traffic matters, look at the specialist density it is compensating for.
- Africa has approximately 1,974 neurosurgeons for 1.3 billion people, roughly 0.15 per 100,000. In parts of West Africa the figure is below one per million.
- In South Africa's Eastern Cape, four infectious disease specialists serve about seven million people.
- Across many low- and middle-income settings, entire subspecialties are represented by a handful of individuals nationally, concentrated in one or two teaching hospitals in the capital.
In that environment, the question is not whether a district hospital physician should consult a specialist. It is whether any specialist is reachable at all by any means.
And the evidence indicates that the channel doing the most work is peer-to-peer. A study of Kenyan physicians using telemedicine found that 51 percent of their telemedicine use was physician-to-physician, rather than physician-to-patient.
That is a striking finding and it inverts the usual assumption about what telemedicine is for in these settings. The scarce resource is not access to a doctor. It is access to a doctor who knows more than you do about this specific problem.
The informal channel is delivering real clinical value
It would be easy to frame all of this as a compliance problem. The evidence suggests the informal channel is doing substantial good.
Botswana teledermatology via WhatsApp: 811 consultation threads over four years, with 62 percent being provider-to-provider consultations about specific patients. Skin conditions are exceptionally well suited to image-based consultation, and the service substituted for specialist visits that would not otherwise have occurred.
South African district hospitals using similar arrangements report substantial time and cost savings.
Migration research documents the pattern directly. A Social Science and Medicine study of migrant doctors found they "continue to consult remotely" for patients in their countries of origin, across professional and national boundaries, as an ongoing feature of their working lives.
And the traffic runs in the other direction too. A study of recent Chinese immigrants in the United States found 15 percent used China-based telehealth applications as a first line of consultation, suggesting patients as well as clinicians maintain cross-border clinical relationships.
This is not a marginal phenomenon at the edges of global health. It is a large, functioning, unmeasured system.
What formal global health built instead
Institutional global health has produced genuinely valuable models, and it is worth being precise about what they do and do not cover.
Project ECHO is the outstanding example: hub-and-spoke tele-mentoring with, by its own published figures, more than 1,600 hubs, thousands of programs, and reach across a very large number of countries. It works, it has been evaluated extensively, and it has changed practice in many settings.
Its unit is a scheduled clinic. A weekly session on a defined topic. That is excellent for building capability over time and structurally unable to answer a question at 6:40 on a Tuesday morning.
Institutional twinning partners a hospital in a high-income country with one elsewhere. Real value, and it is institution-to-institution, dependent on grant cycles and leadership continuity, and covers whichever pairs happen to exist.
Mission and outreach trips deliver concentrated capability for a defined period, then leave.
Each of these is scheduled, institutional, and planned.
The demand is unscheduled, peer-to-peer, and immediate. So it routes to WhatsApp, which is available, free, instant, and entirely ungoverned.
The problems with the current arrangement, stated honestly
An article celebrating the informal channel without naming its failures would be irresponsible, because the failures are serious.
No verification. The person answering is whoever the sender knows. Their competence for this specific question is assumed from a decade-old training relationship.
No consent framework. Patient images and clinical details cross borders with consent practices that vary from careful to nonexistent. Studies of clinical messaging groups have found consent obtained for only a minority of posted images.
No routing beyond the friend graph. If your classmate does not know a neurosurgeon, the chain ends. The physician with no diaspora connections has nothing at all, and that is a large population.
No record. The advice is not in the patient's chart. The next clinician cannot see it. Nobody can review whether it was right.
No liability clarity. The advising physician is operating outside their licensure jurisdiction, offering an opinion on a patient they have not examined, in a legal environment neither party has examined. Most participants have simply not thought about it, which is not the same as it being safe.
And an asymmetry of burden. The diaspora physician carries a moral obligation that can become genuinely heavy: messages at all hours, questions beyond their competence, and the knowledge that declining may mean nobody answers. The home-country clinician carries the clinical risk with no recourse if the advice is wrong.
Why the answer is governance, not replacement
The instinctive institutional response is to prohibit or replace the informal channel. Both fail, for the same reason: the informal channel exists because it is the only thing that works at the speed and specificity required.
Prohibition removes the only available option and produces nothing in its place. Replacement by a scheduled institutional program addresses a different need.
The productive framing is to keep the trust graph and add the missing layers underneath it.
Verification of both parties. Who is asking, who is answering, what are they actually qualified in, and are they currently practising it. This is the piece that converts "my classmate from eleven years ago" into a defensible clinical interaction.
Routing beyond the friend graph. The single largest improvement available. The physician in Lagos should be able to reach a neurosurgeon whether or not his particular classmate happens to know one. The diaspora network's value multiplies enormously when it becomes searchable rather than personal.
De-identification and documented consent, by default. Built into the act of asking rather than left to individual practice, with consent obtained according to the patient's own jurisdiction.
Explicit advisory framing. Non-binding opinion, treating clinician retains full responsibility, no prescribing across borders, no assumption of care. Stated in the structure of the interaction and reviewed under both legal systems, because the ambiguity currently protects nobody.
A record for the asking clinician. So the advice can enter care properly and be reviewed later.
And a bounded obligation for the answerer. The diaspora physician who declares availability for specific question types, with limits, is far more sustainable than one receiving unbounded personal appeals.
Why now
Several things have changed in the last few years that make this more tractable than it was.
The bridge population is growing quickly, up 62 percent since 2010 by OECD figures.
Conflict-driven migration from Ukraine, Sudan, Gaza, and elsewhere has created new diaspora physician cohorts with strong ongoing ties and acute need at the other end.
Technical friction has collapsed. Automatic translation is good enough for clinical exchange in many contexts. Image de-identification can be automated. Video is universal.
And funder priorities have shifted toward locally led models that prize peer networks over external delivery, which is precisely what a diaspora consultation graph is.
What you can do now
If you are a diaspora physician
Set boundaries and state them. "I can look at dermatology questions, within 48 hours, and I cannot advise on management" is far more sustainable than open-ended availability, and it makes you more useful rather than less.
Insist on de-identification. Ask for images cropped, names removed, identifiers stripped, before you look. You are the person best placed to set that norm and the request is easy to make.
Be honest about the limits of your competence. The role you frequently occupy is broker rather than expert, and saying "I am not the right person, let me find you someone" is more valuable than an uncertain opinion.
Build the second hop deliberately. Your value is largely in who you can reach. Knowing which colleagues will look at an image for you multiplies what you can offer.
If you practise where specialists are scarce
Ask about consent and documentation explicitly, according to your own jurisdiction's rules, before sending anything.
Record what you were told and who told you, in the patient's chart, as advice received from a named colleague. This protects the patient's continuity and your own position.
Ask for reasoning, not just conclusions. "What would make you change that advice" produces something you can apply to the next patient, which is the compounding benefit.
If you fund or run global health programs
Study the informal channel before designing around it. It is carrying substantial volume that appears in no evaluation, and any program that ignores it is competing with something free, instant, and trusted.
Fund routing rather than only scheduled sessions. Structured tele-mentoring builds capability over time. It does not answer Tuesday's question, and both are needed.
Treat the diaspora as infrastructure. Six hundred thousand physicians with dual clinical fluency and standing relationships in under-resourced systems is one of the largest untapped assets in global health, and essentially no program is built around it.
Frequently asked questions
How many foreign-trained doctors work in high-income countries? The OECD's International Migration Outlook 2025 reports 606,000 foreign-trained doctors working in OECD countries, representing 18.4 percent of all doctors and up 62 percent since 2010, with roughly 89,000 trained in countries on the WHO health workforce crisis list. About one in four US physicians is an international medical graduate.
How scarce are specialists in low and middle income countries? Extremely, in many specialties. Published figures indicate Africa has approximately 1,974 neurosurgeons for 1.3 billion people, roughly 0.15 per 100,000, with parts of West Africa below one per million. In one South African province, four infectious disease specialists serve about seven million people.
Is telemedicine in these settings mostly for patients? Not necessarily. A study of Kenyan physicians using telemedicine found 51 percent of their use was physician-to-physician consultation rather than physician-to-patient, indicating the scarce resource is specialist expertise rather than access to a clinician.
Does WhatsApp consultation actually help? The published evidence suggests it delivers real clinical value in specialist-scarce settings. A Botswana teledermatology service handled 811 consultation threads over four years with 62 percent being provider-to-provider consultations about specific patients, substituting for specialist visits that would not otherwise have occurred. The same studies document serious gaps in consent, record keeping, and verification.
What are the risks of informal cross-border consultation? No verification of the advising clinician's current competence, inconsistent patient consent, patient images crossing borders without governance, no record entering the patient's chart, no routing when the personal network cannot answer, and unclear liability for a physician advising outside their licensure jurisdiction on a patient they have not examined.
Why doesn't Project ECHO or institutional twinning solve this? They address a different need. Structured tele-mentoring and institutional partnerships build capability through scheduled sessions and long-term relationships, which is valuable. Neither answers an unscheduled, patient-specific question at the moment it arises, which is what the informal channel is used for.
The bottom line
Six hundred thousand physicians in wealthy countries maintain living professional relationships with the health systems that trained them, many of which have a handful of specialists for populations of millions.
That is one of the largest and best-positioned clinical bridges in existence. It is bilingual, clinically fluent, built on genuine trust, and already carrying traffic every day.
Global health has spent decades building institutional partnerships, scheduled tele-mentoring, and outreach programs, all of which do real good on their own terms and none of which answers the message that arrives at 6:40 on a Tuesday with a CT image attached.
So that message goes to WhatsApp, where it is answered by whoever the sender happens to know, with no verification, no consent framework, no record, no routing when the chain runs out, and no clarity about who is responsible for anything.
The bridge is there. It has been there for years, carrying more weight every year.
Nobody has built a road on it.
Part of a series on the missing professional infrastructure of healthcare. Previously: Mentors Give Advice. Sponsors Spend Capital.
Evidence note: migration figures come from the OECD International Migration Outlook 2025. Neurosurgical density figures come from published global surgery literature and vary by source and by definition of the workforce counted. Kenyan telemedicine use comes from a 2023 study with a sample of 157. Botswana teledermatology figures come from JMIR Dermatology (2023). Diaspora consultation behavior comes from Social Science and Medicine (2025), a qualitative study with 21 participants. Cross-border patient telehealth use comes from a 2026 Digital Health study with 227 participants. Regulatory and liability positions for cross-border consultation vary substantially by jurisdiction and nothing in this article is legal advice.
