At the London School of Economics, I learned to look at healthcare through models: costs, outcomes, incentives and trade-offs.

At a Großvisite in a German care home, I learned what those models leave outside the frame.

A Großvisite is a scheduled visit in which the dental team moves through the care home and carries out routine checks for residents. The rhythm is practical. We knock, introduce ourselves, adapt to the room and to the person in front of us, examine, explain and move on. The dental equipment travels with the clinicians rather than asking residents to travel to it.

What stayed with me was not a dramatic procedure. It was the atmosphere around ordinary care. Frau Neumann brightened when the dentist asked how she was getting on with her prosthesis; she was proud that it felt comfortable again. Herr Müller had several questions ready, then stayed for the pleasure of the conversation. Their gratitude appeared in small forms: a smile, a joke, a careful description of what felt different when chewing.

On a spreadsheet, these moments are difficult to enter into a cell. In the room, they were part of the value of the visit.

A note on Frau Neumann and Herr Müller: both are composite characters created from recurring impressions during mobile-care visits. Their names and personal details are invented, and neither represents an identifiable patient. These scenes are illustrative, not an evaluation of outcomes.

The cost model changed at the resident’s door

Before the visit, I might have compared two delivery models by starting with the dental appointment itself. In one model, a patient goes to a practice. In the other, the dentist goes to the patient. The second can appear more expensive because professional time is spent travelling and equipment must be portable.

That comparison is too narrow.

For a resident with limited mobility, reaching a dental practice may require a care worker or relative, a suitable vehicle, coordination with medication and personal-care routines, and sometimes non-emergency patient transport. In the care home, I could see why that journey may be stressful for someone living with dementia, severe disability or several health conditions. It is a great deal of disruption for a routine check, and policies for older people increasingly recognise the value of bringing suitable care closer to where people live (OECD, 2025).

Those resources are real even when they do not appear on the dentist’s invoice. A relative’s time, a care worker’s absence from the ward, transport capacity, waiting and the resident’s distress all belong to the pathway. If an economic analysis counts only the provider’s bill, it can make a costly pathway look cheap by moving costs to families, care facilities or another budget.

Mobile care can avoid a journey and some of the time around it when an on-site visit is a safe substitute for a practice appointment (OECD, 2025). But the qualification matters. Not every dental procedure belongs in a care-home room, and not every resident can or should be treated there. Infection control, lighting, equipment, clinical risk, consent and the need for imaging or more complex intervention all shape the decision (Borg-Bartolo et al., 2020). Mobile care is not a universal replacement for the practice. It is an additional route that may create more value for the right patient and task.

The relevant economic question is therefore not, ‘Which appointment is cheaper?’ It is, ‘Which complete pathway achieves the outcome that matters, with the least burden and the best use of resources?’

Some outcomes become visible only when you are close enough

Economists like measurable outcomes for good reasons. Measurement makes alternatives comparable and claims contestable. Yet the Großvisite made me notice how quickly measurement can shrink the purpose of care.

Oral health is not only the absence of disease. The World Health Organization describes it in terms that include eating, speaking, self-confidence, well-being and the ability to socialise without pain or embarrassment (WHO, 2026). That wider definition matched what I saw. Frau Neumann’s prosthesis supported eating, but her ease in talking about it also suggested something more personal: confidence. Herr Müller’s questions reminded me that a routine visit can detect a problem and give a resident a direct conversation with a professional.

These are not sentimental extras added after the ‘real’ clinical outcome. They are reasons the clinical outcome matters.

The residents’ reactions also changed my understanding of access. Access is often measured by whether a service exists, is covered or was used. Those measures are necessary, but they do not fully describe care that reaches someone in a familiar place, at a manageable pace and with enough time for a question. A service can be technically available and still impose a pathway that a frail person cannot comfortably navigate.

Nor should one positive visit be turned into population evidence. I cannot infer improved quality of life from gratitude in a room. What I can infer is that our evaluation should ask about more than completed examinations. If we study mobile care, we should measure patient-reported comfort and experience, ability to eat, avoided transport, staff time, follow-up, clinical outcomes and the full cost of delivery. Proximity does not replace evidence. It tells us which evidence may be missing.

Value is negotiated among several people

My training encouraged me to look beyond traditional decisions based only on clinical effectiveness or the smallest immediate expenditure. Value-based decision-making asks how outcomes that matter to patients relate to the resources used across the pathway. The European Commission’s expert panel widened the concept further, distinguishing personal, technical, allocative and societal value (European Commission Expert Panel, 2019).

The distinction matters in a care home because no single stakeholder can define the answer alone.

The resident can tell us whether the care is tolerable, understandable and relevant to daily life. The dentist judges what is clinically safe and feasible. Care-home staff understand routines, capacity and the resident’s behaviour over time. Relatives or legal representatives may support history, consent and continuity. Insurers and policymakers shape reimbursement. Operators see whether a theoretically attractive process can work repeatedly on a Tuesday morning.

These perspectives can conflict. A resident may value remaining in their room. A dentist may need equipment available only in the practice. A care home may welcome fewer external journeys but struggle to release staff during the visit. A payer may see the dental fee but not the transport or care time saved elsewhere. Value-based care does not make these trade-offs disappear. It gives us a better table around which to discuss them.

This is also why mobile care cannot be designed by dentists alone, or by economists alone. A qualitative study in two rural German care homes, involving eleven professional stakeholders, found that access, cooperation, incentives, information and equipment all shaped oral healthcare delivery (Gomez-Rossi et al., 2022). Its small, context-specific sample cannot tell us what every care home needs. It does show why an intervention aimed at one professional group can fail when the pathway depends on many.

Ageing turns a local lesson into a system question

The scale of the issue is growing. Germany’s Federal Statistical Office projects that one in four people in the country will be aged 67 or older by 2035, compared with one in five in 2024 (Destatis, 2025). Age does not itself determine dependency. Across OECD countries, however, more than half of people aged 65 and over were living with at least two chronic conditions in 2019, making adaptable, coordinated services increasingly important (OECD, 2025).

Mobile dentistry already exists in several forms. A Swiss survey described private providers using portable equipment and university services operating through different models, while calling for stronger collaboration (Borg-Bartolo et al., 2020). A later Swiss pilot examined 56 care-dependent older people, 47 of them care-home residents, and reported low programme uptake alongside substantial oral-health needs (Borg-Bartolo et al., 2021). That study was small and did not establish that one mobile model is cost-effective. I like it precisely because it resists an easy story: bringing a service closer does not automatically mean people will use it or that implementation will be simple.

For health economics, this creates a practical research agenda. We need comparisons that follow the whole pathway: who receives care, who is missed, which visits are safely completed on site, which require referral, what transport and staff time are avoided, what new costs appear, and whether outcomes and experience improve. We also need distributional questions. Does mobile care reduce unequal access, or does it mainly reach facilities that already organise services well?

The point is not to prove that mobile care always wins. It is to stop evaluating it with a model that ignores the reasons it exists.

Innovation can begin with a better trolley

The most immediate innovation I saw was not artificial intelligence or a new clinical device. It was a trolley.

Our team recently bought a transportable, foldable trolley for mobile visits. Previously, arriving at a care home could begin with a search for a suitable surface: Is a trolley available? Is it clean? Can it carry the instruments safely? The new trolley gives the dental team a predictable workspace, with equipment organised and ready to move from room to room. We have observed smoother set-up and fewer small obstacles before care can begin, although we have not yet measured the time saved.

That modest improvement matters because accessibility depends on more than bringing a dentist through the front door. The service must also be easy to set up and reliable enough to work around residents. Better organisation can help the team spend less attention on searching and carrying, and more attention on the person receiving care.

The trolley also hints at what supportive technology might make possible. A future version could help track supplies, guide packing or follow the dentist between rooms. Perhaps it could one day present a requested instrument. The purpose would be simple: remove practical friction without replacing clinical judgement or human contact. The economic question comes first: Which recurring tasks consume time, and where could technology create enough value to justify its cost?

What I brought back from the Großvisite

I left the care home more convinced of health economics, not less. Scarce resources still require comparison, evidence and trade-offs. But the visit changed what I think belongs in the comparison.

The value of mobile dental care may include a clinical problem detected, a prosthesis that works, a stressful journey avoided, a care worker who remains on the ward and a resident who can ask a question in a familiar room. Its cost includes the dentist’s travel, portable equipment, coordination and all the operational work needed to create a safe treatment environment. Both sides of the equation are broader than a fee schedule.

The trolley reminded me that innovation can begin with close observation of friction. The residents reminded me that outcomes are lived before they are measured. And the care team reminded me that no model of value is complete until the people who deliver and receive care have helped define it.

A spreadsheet can support that decision. It should not be allowed to make the room disappear.

References

1. Borg-Bartolo, R., H. Amberg, O. Bieri, E. Schirrmann, and S. Essig. 2020. “The Provision of Mobile Dental Services to Dependent Elderly People in Switzerland.” Gerodontology 37: 395–410. https://doi.org/10.1111/ger.12490.

2. Borg-Bartolo, R., T. von Wyttenbach, M. J. Keller, et al. 2021. “Delivery of Mobile Dental Services to Dependent Elderly People: Results from a Pilot Study in Rural Switzerland.” Swiss Dental Journal 131: 584–593. https://pubmed.ncbi.nlm.nih.gov/33512790/.

3. Destatis. 2025. “By 2035, One Quarter of Germany’s Population Will Be Aged 67 or Over.” Press release, December 11, 2025. https://www.destatis.de/EN/Press/2025/12/PE25_446_12.html.

4. European Commission Expert Panel. 2019. Defining Value in ‘Value-Based Healthcare.’ Brussels: European Commission. https://health.ec.europa.eu/publications/defining-value-value-based-healthcare_en.

5. Gomez-Rossi, J., J. Schwartzkopff, A. Müller, et al. 2022. “Health Policy Analysis on Barriers and Facilitators for Better Oral Health in German Care Homes: A Qualitative Study.” BMJ Open 12: e049306. https://bmjopen.bmj.com/content/12/3/e049306.

6. OECD. 2025. The Economic Benefit of Promoting Healthy Ageing and Community Care. Paris: OECD Publishing. https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/10/the-economic-benefit-of-promoting-healthy-ageing-and-community-care_7f0491f3/0f7bc62b-en.pdf.

7. World Health Organization. 2026. “Oral Health.” Accessed August 6, 2026. https://www.who.int/health-topics/oral-health/.

On the clinic schedule, Frau Neumann’s appointment lasted twenty-five minutes.

For her, it began twelve days earlier.

That was when she first noticed that chewing on the left side had become painful. She waited for a few days, hoping it would settle. Then she looked for the card from her dentist, found two cards with similar names, and called the number she thought was right. The practice asked her to describe the problem, find her insurance card and bring an up-to-date medication list. Her daughter arranged time away from work to take her there. The appointment had to fit around a home-care visit and the hours during which Frau Neumann is usually most alert.

By the time she sat in the dental chair, several decisions had already been made and several small problems had already been solved. None of this appeared in the appointment book. The record began when she arrived.

Frau Neumann is a composite character, drawn from situations that recur in the care of older patients. She is not an identifiable patient. The details are invented, but the work is familiar.

In the first Care, Built article, Healthcare Is Discussed From Too Far Away, I wrote about the distance between a pathway on paper and care as it is actually delivered. One of the things that disappears at a distance is the work performed by the patient. We draw an arrow from “problem” to “appointment” and another from “appointment” to “treatment.” Someone still has to move the case along those arrows, and very often that work falls to the person who is ill.

The unofficial coordinator

We tend to divide healthcare work by profession. Doctors diagnose. Nurses monitor. Pharmacists dispense. Therapists rehabilitate. Administrators arrange and document. These categories are visible because people are employed to perform them, time is allocated to them and, in many cases, a code records that they occurred.

The patient’s role is described rather differently. The patient should “seek care,” “attend follow-up” and “adhere to treatment.” Those phrases sound like three simple actions. In practice, they can contain a surprising amount of work. Call it patient work.

The patient notices a change and decides whether it deserves attention. They work out which part of the system is responsible. They find a phone number, call during opening hours and explain the problem in the vocabulary the service expects. They compare appointments with work, childcare or home-care schedules. They obtain a referral, carry information between organisations, repeat a medical history, remember conflicting instructions and notice when nobody has called back.

Afterwards, they may need to collect a prescription, arrange another appointment, monitor symptoms and decide whether a change is expected or alarming. If several conditions are involved, they also become the only person with a view across the whole process.

For a healthy person with flexible work, good health literacy, reliable internet and confidence in the local language, this may be annoying but manageable. That is often the person implicitly imagined when a pathway is designed. The healthcare system gives them a sequence of tasks, and they complete it.

But the need for healthcare does not rise and fall with a person’s ability to manage those tasks. Quite often, the relationship runs in the opposite direction. Illness, pain, fatigue, dementia, depression, poor vision and limited mobility all increase the need for care while reducing the capacity to organise it.

This is particularly clear in mobile dentistry. A resident in a nursing home may not be able to identify the source of pain, call a practice, explain a broken denture, produce a medication list or consent to treatment without support. The work does not disappear. A care worker may notice that lunch has been left untouched. A nurse looks for the medication plan. A relative answers questions about previous treatment. Someone coordinates the visit, makes sure the resident is available and keeps track of a denture if it has to go to a laboratory.

When each of those things happens, the clinical encounter can look simple. When one does not happen, the treatment may never begin.

The same principle applies far beyond nursing homes. A parent trying to organise care for a disabled child, a person recovering from surgery, someone with a new cancer diagnosis or a patient seeing several specialists at once may all spend a significant part of the week managing healthcare. Family members often absorb the work, which makes it less visible rather than less real.

Who disappears before the appointment

There is a selection effect hidden in many healthcare workflows. We mostly see the people who reached the appointment.

The online form was apparently usable because the patients in the waiting room managed to complete it. The telephone system appeared adequate because the people speaking to the receptionist got through. The referral process seemed to function because the specialist saw patients with referrals.

The people who stopped earlier are harder to observe. They may appear later as a missed appointment, delayed treatment or an emergency. Sometimes they do not appear at all. A database showing no treatment cannot usually tell us whether no treatment was needed or whether the path to it quietly failed.

This matters because small obstacles do not affect everyone equally. An extra password, a morning spent calling, a form available only in one language or a request to upload a document may be a minor inconvenience for one person and the point at which another person gives up. Add several such steps together and access begins to depend on persistence, time, literacy and family support as much as on clinical need.

We often describe this as patient behaviour. The person “did not engage,” “failed to attend” or was “lost to follow-up.” Sometimes that description is fair; people can make choices that professionals would not recommend. But it can also conceal a design decision. Before concluding that a patient failed to comply with a pathway, it is worth asking how much project management the pathway demanded from them.

Formal coverage can therefore overstate practical access. A service may be included in insurance, located somewhere in the city and technically available. Yet receiving it still depends on a series of successful handovers, many of which are assigned to the patient without ever being named as such.

A case: the subsidy that must be applied for

Germany’s hardship protection for dental prostheses offers a particularly clear example.

As of 2026, statutory health insurance ordinarily pays a fixed subsidy equal to 60 per cent of the average cost of a defined standard treatment. For patients on very low incomes, the Härtefallregelung raises that subsidy to 100 per cent. If the patient chooses the standard treatment, the insurer covers its full cost. The arrangement is still often called the doppelter Festzuschuss, although the additional subsidy is now 40 rather than another 60 percentage points. It is a sensible protection: lack of money should not become lack of teeth. (KZBV, 2026)

But the protection does not activate itself. The insurer has to establish that the patient qualifies, which generally requires an application and evidence of income or receipt of a qualifying benefit. For an independent patient, this may mean one form and a few supporting documents. In a nursing home, the same request can expose a much larger problem: who is responsible for making it happen?

A resident with advanced dementia may not understand the subsidy, know where a pension or social-assistance notice is kept or be able to sign. A nurse may know that new dentures are needed but have neither the documents nor the legal authority to act. The dental practice may not know whether a relative, an authorised representative, a court-appointed guardian or the facility’s social service handles the resident’s finances. The relative may live in another city. The insurer, reasonably, waits for proof.

There may be only a few sheets of paper. The work lies in finding the person who is both able and authorised to submit them.

The available evidence suggests that the protection is useful once it has been activated. In an analysis of 404,610 AOK Nordost members aged 75 or older, having recognised social-hardship status was associated with higher use of prosthetic services (adjusted OR 1.19; 95% CI 1.17–1.21). The association is consistent with hardship protection reducing financial barriers once the status has been recognised. (Hempel et al., 2020)

It also changes how I read research I helped to conduct before building SwissMedAI. In a separate analysis of the same population, my co-authors and I examined how age, health, geography and social-hardship status were associated with wider dental-service use. The hardship variable appeared in the data as a clean administrative fact. It allowed us to study people whose status had been established; it could not distinguish a person who was ineligible from one who qualified but never completed the process. (Schwendicke et al., 2021)

Administrative data record recognised eligibility more readily than unrealised entitlement. If protection depends on a successful application, the people least able to apply can disappear from the variable intended to identify vulnerability.

Evidence from a neighbouring German benefit shows why that blind spot deserves attention. Buslei and colleagues estimated that around three in five households eligible for old-age basic-income support were not receiving it, with non-take-up especially high among older and widowed people. They discussed limited information, small expected benefits, stigma and procedural complexity as possible explanations. (Buslei et al., 2019)

That figure is not an estimate for the dental Härtefallregelung. I have not found a robust dental-specific estimate of how many eligible patients fail to obtain it. The comparison establishes a narrower point: a legal entitlement that depends on an application will not necessarily reach everyone who qualifies. In a qualitative study I co-authored in German care homes, dementia, disability, information exchange and cooperation among residents, carers, relatives and dental professionals all shaped whether oral care could be delivered. The study did not examine subsidy applications, but it described the human network through which one has to pass. (Gomez-Rossi et al., 2022)

Seen from the insurer’s office, asking for a few documents is proportionate. Seen from the resident’s room, those documents may sit at the end of a chain involving the dentist, nursing staff, social services, a relative and a legal representative. If nobody owns the chain, a benefit designed for the most vulnerable becomes easiest to obtain for the most administratively capable among them.

The answer is not to abandon eligibility checks. Where another public authority has already established receipt of Grundsicherung im Alter or a qualifying benefit, hardship status should, as far as legally and technically possible, be recognised without asking a cognitively impaired resident to prove the same vulnerability again. Where more evidence is unavoidable, the process needs a named owner and a fallback: someone who contacts the authorised representative, sees that a document is missing and follows the case when nobody replies.

A right that works only after the patient has coordinated several organisations is not equally available to every eligible patient.

From an economic perspective, these tasks are part of the cost of care even when no invoice is issued. Time away from work, transport, waiting, repeated phone calls and unpaid help from relatives are resources. So are attention and emotional energy, although they are harder to price. A pathway does not become inexpensive merely because part of its cost has been moved outside the healthcare budget.

When convenience means self-service

Digital healthcare can reduce this burden. Online appointments are often easier than calling. A well-designed portal can make results available immediately, preserve instructions and spare patients a journey. Remote monitoring may replace visits that provide little value.

The difficulty is that a digital service is not necessarily a convenient one.

Some digital services remove work; others transfer it. The patient creates another account, locates a document, scans it, enters information already held elsewhere and waits for a verification message. Each provider has its own portal. The general practitioner cannot see the hospital’s messages, so the patient downloads a PDF and brings it to the next appointment. What looks like automation from the organisation’s side may be unpaid administrative work from the patient’s side.

This is not an argument for keeping every process on paper or routing everything through a receptionist. Many patients prefer self-service and should have it. The problem is assuming that a self-service route is neutral, or that its existence removes the need for another route. A system designed around the most digitally confident users will produce very favourable feedback from the people who can use it and very little feedback from those who cannot enter it.

Artificial intelligence will create a similar choice. It can help patients interpret instructions, prepare questions, summarise records and navigate services. It can also generate more alerts, more messages and more things a patient is expected to monitor. Whether it reduces burden depends less on the sophistication of the model than on what the patient no longer has to do after it is introduced.

There is, however, an easy way to take this argument too far. Patients are not passive recipients, and a good healthcare system should not treat them as such.

Many parts of care cannot and should not be delegated. People have to describe what matters to them, make decisions, manage medicines, perform exercises, change habits and live with the consequences of treatment. Participation can increase autonomy and improve care. For chronic conditions, much of the clinically meaningful work will always happen at home rather than in a consultation room.

The useful distinction is between participation in one’s care and compensation for a fragmented system.

Learning how to manage diabetes is part of living with and treating diabetes. Repeating the same medication list to four organisations because their systems do not communicate is not. Choosing between treatment options is an exercise of autonomy. Discovering which office lost a referral is clerical work. Monitoring a symptom after surgery may be necessary; carrying an imaging result between two departments in the same health system should not be.

Removing unnecessary coordination does not infantilise patients. It leaves more capacity for the decisions and actions in which their involvement actually matters. It also allows people who want to manage more of the process themselves to do so without making that ability a condition of receiving care.

Where the process starts

Healthcare organisations usually define the beginning of a process at the point where they first touch it: an incoming referral, a booked appointment, a patient at reception. The patient’s process often began much earlier.

At SwissMedAI, this changes how a mobile-care pathway has to be understood. A dental visit in a nursing home does not begin when the team enters the building. It begins when someone notices a possible need and has a reliable way to report it. Before treatment, there may need to be information about medicines, consent, a conversation with a relative and coordination with nursing staff. After treatment, somebody must understand what was done, what to watch for and whether another visit is required.

The practical question is not whether all of this work can be eliminated. It cannot. The question is whether every necessary step has an owner and whether that owner has the information and capacity to complete it. “The patient will arrange it” is sometimes an appropriate answer. It should not be an automatic one.

This also changes what should be measured. Appointment activity tells us about people who have already crossed much of the pathway. To understand access, we also need to know where requests stop, why follow-up is delayed and which groups repeatedly need someone else to rescue the process. A no-show rate alone says very little. It combines forgetfulness, illness, transport failure, misunderstanding, a change of mind and poor scheduling into one convenient number.

None of this requires a grand new institution. Often the improvements are ordinary: fewer duplicate questions, one clear contact, status information that can be understood, follow-up that does not depend on memory, and an explicit fallback when the patient cannot complete the next step. The main change is to treat coordination as part of care rather than as the empty space between clinical events.

Frau Neumann eventually reaches the dental practice because her daughter takes responsibility for most of the process. The painful area is examined and treated. The record contains the examination, the diagnosis and the procedure. It does not contain the search for the phone number, the calls, the rearranged workday, the medication query or the uncertainty about whether the appointment was really necessary.

If her daughter had been unavailable, the clinical record might contain nothing at all.

That absence would be easy to interpret as an absence of care. It might instead be the place where care was needed but never quite became an appointment.

When we say that a service is accessible, then, we should ask more than whether it exists and who is entitled to it. We should also ask what a person has to notice, understand, remember, organise and endure before the service can begin. The answer tells us how available the care really is.

Leave a comment

References and notes

1. Kassenzahnärztliche Bundesvereinigung. Festzuschuss und Eigenanteil. Current description of the fixed-subsidy and hardship rules, including full coverage of standard care in recognised hardship cases.

2. Hempel FM, Krois J, Paris S, et al. Prosthetic treatment patterns in the very old: an insurance database analysis from Northeast Germany. Clinical Oral Investigations. 2020;24:3981–3995.

3. Schwendicke F, Krasowski A, Gomez Rossi J, et al. Dental service utilization in the very old: an insurance database analysis from northeast Germany. Clinical Oral Investigations. 2021;25:2765–2777.

4. Buslei H, Geyer J, Haan P, Harnisch M. Starke Nichtinanspruchnahme von Grundsicherung deutet auf hohe verdeckte Altersarmut. DIW Wochenbericht. 2019;86(49):909–917.

5. Gomez-Rossi J, Schwartzkopff J, Müller A, et al. Health policy analysis on barriers and facilitators for better oral health in German care homes: a qualitative study. BMJ Open. 2022;12:e049306.

If you wanted to understand why healthcare fails, where would you go?

The usual answers are some version of: a ministry, an insurer, a university, a technology conference. There you would find intelligent people discussing reimbursement systems, clinical guidelines, interoperability and artificial intelligence. These are reasonable places to look, and I have sat in a number of them myself, nodding at diagrams.

I would take you somewhere else. I would take you to Room 214 of a nursing home.

There you might meet someone like Frau Keller, who is 86, has moderate dementia, and has broken her lower denture. The problem sounds straightforward enough: repair the denture. Clinically, it may even be straightforward.

But first, someone has to notice that she is eating less. Someone has to work out whether the denture is broken, painful or simply missing. Her daughter may need to be contacted. Her medical history and medication list must be available. A member of the nursing staff has to accompany her. The dentist needs the right equipment in the building. The denture may need to travel to a laboratory and come back without getting lost, and everyone involved needs to know what happens next.

The actual dental procedure might take twenty minutes. Delivering those twenty minutes can take several days and half a dozen people.

This is the first thing distance hides: the treatment is often the smallest part of delivering the treatment.

The altitude problem

Healthcare looks different depending on the altitude from which you observe it.

From 10,000 metres, it consists of populations, budgets, institutions and policies. From 1,000 metres, it consists of pathways, professional groups, contracts and information systems. At ground level, it consists of a particular person, in a particular room, needing someone to do something.

Each altitude shows something true. None of them shows enough on its own, and the trouble begins when we design healthcare at one altitude and assume it will behave the same way at another.

A policy document might contain an arrow between “assessment” and “treatment”. On paper, the arrow occupies two centimetres. In reality, that arrow may contain three phone calls, a missing consent form, an unavailable caregiver, incompatible software, an unanswered email and a patient who no longer remembers why a dentist is standing in her room.

Healthcare accumulates complexity inside its arrows.

From far away, the arrows look free.

Three maps of the same system

My own view of healthcare has changed mostly because I have kept changing professions inside it.

As a dentist, I was trained to read the clinical map: diagnosis, treatment options, risks and outcomes. As a health economist, I learned the incentive map: who pays, what is reimbursed, what gets measured, and which behaviours the system quietly encourages. As a data scientist, I learned the information map: which events become data, which variables disappear, and how confidently we can infer anything from the record that remains.

Then I built and operated a healthcare organisation, and discovered a fourth map that nobody had trained me to read. Call it the operational map. It holds the things that almost never appear in scientific papers or policy presentations: who calls the patient’s daughter, where the equipment is stored, whether the nursing home has a working elevator, who notices an incomplete form, and what happens when the only person who understands a process is on holiday.

Each of these maps can be accurate and still dangerously incomplete. A clinically excellent intervention fails because nobody is paid to coordinate it. A financially attractive programme fails because it does not fit into anyone’s working day. A beautiful dataset describes only the parts of care that generate a billable event, and a technically impressive product dies because it adds forty seconds to a task performed two hundred times a day.

Healthcare does not happen on any one of these maps. It happens where they overlap.

Why a new reimbursement code does not create care

Imagine that a health system wants more preventive examinations in nursing homes. A new service is defined, eligibility criteria are written, a reimbursement code is created, and a presentation is produced showing the expected improvement in access.

From a distance, the causal chain appears obvious:

New benefit → providers perform it → patients receive better care.

A reimbursement code, however, is not a care-delivery system. Someone must identify eligible patients. Consent may need to be obtained. A visit must be coordinated with the facility. Documentation requirements must be understood. The service has to fit among other clinical responsibilities, and information must move between organisations that may use entirely different systems. If the work surrounding the reimbursed activity is more difficult than the activity itself, utilisation stays low.

When that happens, it is sometimes read as resistance from professionals. Occasionally it is. More often it is a design problem: the system has funded the visible clinical act while leaving the surrounding coordination unpaid, unowned and invisible.

Incentives do not decide whether healthcare workers care about their patients. They decide which forms of caring can survive contact with a crowded schedule. Those are different questions, and a great deal of policy disappointment lives in the gap between them.

Data records events, not reality

The same distance problem runs through healthcare data.

Suppose a dataset shows that a patient received no dental treatment during a particular year. What does that mean? Perhaps the patient was healthy. Perhaps the patient declined treatment. Perhaps nobody examined her. Perhaps an examination took place but was documented elsewhere. Perhaps treatment was needed, but transport could not be organised. Perhaps the care was delivered and simply is not represented by the code being analysed.

In the dataset, these very different realities look identical: nothing happened.

That is one of healthcare data’s most seductive qualities. It presents a clean surface over an untidy world. The numbers may be correct while the interpretation is wrong, which is a more dangerous combination than an honest error.

Claims data can tell us a great deal about what a system paid for. It is far less reliable about what patients needed, what professionals attempted, or why an apparently indicated intervention never took place. None of this makes the data useless. It makes proximity essential: researchers need to understand how the data were produced, clinicians need to understand what disappears when care becomes data, and policymakers need to know which parts of the system are visible only because they happen to be easy to count.

Otherwise we improve the metric and leave the patient in Room 214 exactly where she was.

Technology has an altitude problem too

Healthcare technology is usually demonstrated from very far away.

A product can analyse an image, draft a clinical note or predict a risk with extraordinary accuracy. The capabilities are real, and they are getting more impressive by the quarter. But capability is not the same thing as usefulness.

Take an AI system that reviews a dental radiograph in seconds and performs beautifully in a controlled evaluation. Now place it inside a working clinic. Can it access the image without a manual upload? Does it know which patient the image belongs to? Can its result enter the clinical record? Is the dentist expected to check a separate screen? What happens when the internet fails? Who is responsible when the model is uncertain? Does the system remove work, or merely move it to someone who appears on no slide?

Nothing about the model’s accuracy has changed. Its value has.

I have come to hold every healthcare technology, very much including our own, to three questions:

1. Where, exactly, does it enter the workflow?

2. Whose work becomes easier or unnecessary?

3. What happens when it is wrong, unavailable or ignored?

A product without good answers to these questions may be an impressive technology. It is not yet a functioning part of care.

Responsible AI, in other words, is not only a matter of model performance, bias and data protection, although all of those matter. It is also a matter of operational honesty. A system should not claim to save clinicians time when it merely transfers work to assistants. It should not claim to improve access when it serves only the patients who already navigate the system successfully. And nobody should call a prediction “actionable” unless someone has both the authority and the capacity to act on it.

The real unit of innovation is not the algorithm. It is the completed workflow.

What mobile care makes visible

At SwissMedAI, much of our work involves delivering dentistry to older people, including residents of nursing homes. I used to think of mobile dentistry as conventional dentistry performed in a different location. The field cured me of that idea fairly quickly. Mobile care is a different operating environment altogether.

In a conventional practice, most of the conditions required for care are already in place before anyone picks up an instrument. The patient has arrived. The chair, the instruments and the records are within reach. The building itself is designed around treatment, and the surrounding organisation stays invisible mainly because it is familiar.

Mobile care removes that illusion. The clinical team must bring not only the treatment but a portion of the system that makes treatment possible: equipment, information, coordination, continuity. This is what makes mobile care difficult, and it is also what makes it revealing. Weaknesses that stay hidden in a conventional setting have nowhere to hide in a nursing home corridor. Poor information exchange shows up immediately. So does ambiguous responsibility, and inadequate reimbursement, and technology that was designed without anyone ever watching the user work.

Nursing homes are commonly treated as a marginal corner of healthcare. I have come to see them as the opposite: stress tests for healthcare design. If a process only works for people who can organise appointments, follow complex instructions, travel independently and coordinate several professionals, it is not a robust process; it is a process that quietly depends on healthy patients doing part of the system’s work.

Proximity is not the same as truth

There is, however, a danger in romanticising the frontline, and I would rather name it myself before someone else does. Being close to care does not automatically make a person right.

A clinician can mistake personal experience for general evidence. An operator can optimise a local process while missing its wider consequences. A single compelling patient story can drown out data from thousands of patients nobody tells stories about. Distance has real advantages: comparison, abstraction, perspective.

The remedy is not to replace policy with anecdote, or research with intuition, but to build a better elevator between the altitudes. Policy informed by what implementation actually requires. Research that tests whether frontline impressions generalise. Technology teams that observe the work before redesigning it. Clinicians who understand the incentives and constraints shaping the systems around them.

We need both the map and the room.

Starting from the room

This publication will be written from that intersection.

I will write about clinical care, incentives, data, regulation and technology, but I intend to begin, each time, with what is actually happening: not with what the workflow is supposed to look like, not with what the database implies must have happened, and not with what a product presentation promises will happen. What actually happens when a policy reaches a clinic? When an algorithm meets a professional workflow? When a theoretically available service reaches a person who cannot independently access it?

Sometimes the answer will come from evidence. Sometimes it will come from operational experience. Sometimes it will remain a hypothesis worth testing, and I will try to be clear about which is which.

I am not arguing that healthcare is impossibly complicated. I am trying to locate where its complexity actually lives, so that we can build systems that deal with it honestly.

Frau Keller eventually receives her repaired denture.

She eats lunch.

It is not a spectacular outcome. Nobody announces it at a technology conference, it appears on no executive dashboard, and depending on the data, we may not even be able to distinguish it from a hundred other repairs. But this is what healthcare is for: making ordinary human outcomes reliably possible.

To build it better, we need to get closer.

A note on Frau Keller: she is a composite, assembled from situations that recur in mobile care. No detail in this essay identifies a real patient.