The healthcare system we have is not the healthcare system we imagined. It was built with care and intelligence over decades, by people who genuinely wanted to improve lives. It has done so. But somewhere in that work — in the building of payment structures, clinical guidelines, diagnostic codes, and event-driven metrics — we built for the wrong thing. We built for clarity. For the moment when something definite happens, when a threshold is crossed and the system can act. We built, in short, for the hard event. And the people who drift — who decline slowly, who lose ground by degrees, who live in the long space between emergencies — learned to wait until they were sick enough for the system to see them.
The system we built — and what it cannot see
The event-based architecture of healthcare was not an accident. It was an achievement — a practical solution to the problem of how to organise, fund, and audit a complex system at scale. The myocardial infarction is precisely the kind of event that healthcare systems are extraordinarily good at managing. The acute collapse, the first exacerbation, the cancer diagnosis: hard transitions, with clean edges, codeable on both sides. The system mobilises at the threshold. It was designed to.
The problem is what happens in the long time before the threshold is crossed. The patient who quietly stops walking to the shops. The one whose grip strength falls by ten per cent over fourteen months. The one who recalibrates her breathing expectations so gradually that neither she nor her GP notices the recalibration itself. These are soft transitions — gradual, multi-signal, distributed in time — and the healthcare system is structurally unable to see them, let alone pay for attention to them.
This is not a simple resource problem. The NHS employs hundreds of thousands of skilled clinicians who are entirely capable of noticing a soft transition. The problem is that the payment architecture does not reward what they notice. A GP who spends forty minutes with a patient tracking a slow functional decline, adjusting a care plan, holding a trajectory in view — has nothing to bill. The same GP who manages a subsequent acute admission generates an encounter, a code, a discharge summary, and a 30-day readmission metric. The system sees the event because the event is what it was built to price. The drift is simply not legible.
Geoffrey Rose gave us the conceptual vocabulary of prevention — high-risk strategy versus population strategy — and it served well when applied to the discrete-outcome diseases it was designed for: blood pressure and stroke, cholesterol and heart disease. But applied to the modern burden of multimorbid chronic illness, Rose's framework hides more than it reveals. It anchors the concept of value to a downstream event, and builds a system that can only justify itself by reference to that event. Prevention, in this frame, is merely deferred reaction. The soft transition — the long, slow, socially embedded, biologically complex drift that constitutes most of living with illness — has no place in the ledger.
The Sussex integrated home renal unit merged peritoneal dialysis and home haemodialysis services and began pooling warm data about soft signals of decline. Patients were identified early enough for planned vascular access — bypassing emergency central venous catheters entirely, and reducing modality-transition hospitalisations to zero in compliant tracks. The soft drift was visible. The system was redesigned to see it. The results were unambiguous. This is not a theoretical alternative. It is operating care, now.
The Sussex model is the first hopeful thing in this essay. It proves what the structural argument predicts: that a system redesigned around the complex domain — probe, sense, respond — can achieve outcomes that the event-based system cannot. Not because it is better resourced. Because it is better oriented. The clinician who asks "would you be surprised if this patient transitioned to home haemodialysis in six months?" is practising a fundamentally different kind of attention than the one waiting for the peritonitis that makes the transition unavoidable.
The certainty reflex — how politics and medicine share the same mistake
Why does the system not change? The structural argument is understood by many of the people who work within it. Commissioners know the dashboard does not capture what matters. Clinicians feel the moral injury of the ten-minute appointment with the complex, multimorbid patient who needs forty. The researchers know their EVPI calculations are well-calibrated to epistemic uncertainty and structurally blind to the aleatory, the ontological, the existential — the uncertainties that matter most to the patient sitting in the car outside the hospital trying to understand what "slowing progression" means in practice.
The answer, partly, is political. Rainer Kattel's reading of Plato offers a diagnosis that reaches deeper than healthcare: those who seek power are, by the very fact of seeking it, seeking certainty. What they want is not the work of governance — the ongoing, adaptive, genuinely uncertain task of navigating complex social systems — but the resolution of that work into authority. The diagnosis is pronounced. The policy is announced. The uncertainty, which was always real and remains real, is simply no longer permitted to speak.
Healthcare does this with extraordinary sophistication. Harold arrives at his GP appointment carrying months of breathlessness, a stoicism about early symptoms that reflects his upbringing as a farmer, and a cough that does not fit the typical COPD presentation cleanly. He leaves with a COPD diagnosis and a prescription. The appointment has done its institutional work: a problem was presented, authority was exercised, a resolution was produced. The underlying uncertainty — which is genuine, biological, and in Harold's case attributable to idiopathic pulmonary fibrosis, not COPD — is now officially resolved. Two years of ineffective treatment follow.
Harold's GP was not being dishonest. She was operating within a system calibrated to produce resolutions. The appointment structure, the ten-minute allocation, the pattern-matching logic of primary care, the institutional expectation that the patient leaves with a plan — all of these are incentives toward a confident answer. The uncertain answer — "I am not sure what this is, and here is what I am not sure about" — has no designated place in the encounter. It is structurally excluded.
The love of certainty is not a virtue. In a world as genuinely uncertain as ours, it is the performance of a virtue — and performance, sustained long enough, corrupts what it imitates.
George, the engineer in our uncertainty framework, brings a notebook to his appointments and asks questions the system was not designed to answer honestly. He wants to know why his prognosis is a population average from a trial population unlike him in significant respects. Francis, the civil servant, experiences an uncertainty that is existential — not about what is happening in his lungs but about what his life now means — and no clinical code captures it. Victor, the entrepreneur, rejects the prognostic estimate. The certainty-seeking clinical system reads this as non-compliance. It is, in fact, a philosophically coherent act of adaptive self-management: a deliberate choice about which information to hold and which to set aside, calibrated to his own values and capacity.
The same logic that produces the strongman in politics — the concentration of authority, the suppression of distributed intelligence, the mistaking of confidence for competence — produces the event-based healthcare system. They are not analogies. They are the same structural pathology, expressed in different institutional registers. And AI, deployed uncritically in either domain, does not transcend this pathology. It automates it. A diagnostic algorithm trained to project confidence, deployed in a system that rewards confidence and penalises acknowledged uncertainty, is not augmenting clinical intelligence. It is amplifying the certainty machine's most dangerous tendency.
What other traditions knew — and what we forgot to carry
Western modernity struck a deal. It traded the irreducible complexity of lived wisdom for the projected clarity of formal knowledge. The gains were real and enormous: penicillin, the welfare state, the randomised controlled trial. We have no interest in pretending otherwise. The deal was worth making.
But the deal had a hidden clause. In privileging knowledge that was propositional, verifiable, and formally transmissible — knowledge you could write down, audit, and replicate at scale — Western institutions systematically devalued forms of knowing that were embodied, distributed, relational, and narratively encoded. Not because those forms were inferior. Because they were inconvenient for institutions that needed to be legible, scalable, and governable from a distance. James Scott called this the destruction of metis: the practical, experience-based cunning of fishermen, navigators, farmers, and clinicians whose knowledge is primarily about how to respond to a complex, variable, context-dependent situation rather than what the situation is in the abstract.
The Polynesian navigator
The navigator crossing five hundred kilometres of open Pacific without instruments is not doing less sophisticated reasoning than a modern GPS user. According to Edwin Hutchins's analysis of distributed cognition, she is performing a computation of extraordinary complexity — one distributed across the body, the crew, the wave patterns, the star paths, and the navigational myths that encode routes learned across generations. Those myths are not decorative. They are the knowledge management system. The whole system knows — not the navigator alone.
This is not pre-scientific reasoning. It is post-GPS reasoning: an epistemology calibrated to genuine complexity rather than the simplified model that GPS requires. The navigator's etak system — in which the canoe is held stationary and the islands move around it — reduces the cognitive load of integrating velocity, drift, current, and direction over hundreds of miles. It is a beautiful computational shortcut for a genuinely hard problem. And it requires a quality that the certainty machine actively discourages: the willingness to hold a dynamic, multi-signal, relational picture rather than demanding a fixed coordinate.
The Talmudic tradition
The Babylonian Talmud's doctrine of elu v'elu — "these and those are both the words of the living God" — was formulated to address a dispute between the schools of Hillel and Shammai that neither school would concede. The resolution was not to establish which school was right. It was to declare both valid — and to rule in favour of Hillel not because Hillel was more correct but because Hillel's scholars demonstrated greater epistemic humility: they taught Shammai's positions before their own, and listened before they ruled.
The consequence is a tradition that has archived acknowledged dissent for two thousand years. Minority opinions are preserved alongside majority rulings. Irreconcilable interpretations are held in tension rather than collapsed. The uncertainty is the record. This is not an absence of rigour — it is a form of rigour specifically designed for systems where authoritative certainty would be a falsification, and where the archive of acknowledged disagreement has more long-term value than the apparent resolution of it.
Indigenous fire management
Indigenous Australian fire management represents millennia of adaptive engagement with an extraordinarily complex ecological system. The knowledge it contains is not primarily propositional — it cannot be captured in a protocol. It is relational: an ongoing, responsive relationship between human practice and a system that cannot be fully modelled. Seasonal burns are not executed from a plan. They are calibrated to the signals the country is emitting this year — this rainfall, this wind, these grasses, this time. The practice is the knowledge. The knowledge lives in the practice.
Western land management, convinced of the superiority of its formal models, dismantled this practice through the twentieth century. What was lost was not a primitive substitute for scientific forestry. It was a body of adaptive complexity science — honed over thousands of years of feedback — that Western institutions could not read because it was encoded in a form their epistemology was not designed to receive. The consequences are still burning.
These three traditions are not cultural curiosities to be appreciated at a distance. They are engineering specifications for epistemic systems that work under genuine complexity. Distributed cognition. Institutionalised tolerance of irreducible plurality. Adaptive iteration responsive to the system's own signals. These are the structural features that complexity-native epistemologies share — and they describe, precisely, what a well-designed MDT would look like, what a well-designed care pathway would do, and what a well-designed AI tool in healthcare would be trained to amplify rather than replace.
The specialist nurse who calls Francis every month is practising a form of metis. She holds knowledge about his situation — the slight change in his breathing on a difficult day, the thing he only says when he trusts her — that no guideline can capture because it is constituted by the relationship, not by the protocol. The healthcare system does not code this. It does not reward it. And when it designs AI-assisted care, it replaces it with an app.
Why understanding is not enough — the attractor and its stability
None of this is news to the people who run health systems. The soft drift has been named — in this work and in others. The certainty complex has been identified. The structural features of better practice are visible in places like the Sussex Kidney Unit, in the monthly specialist nurse call, in the MDT that records "genuinely unclear — monitoring and review" rather than forcing a diagnosis. So why does the system not change?
Complex adaptive systems settle into stable configurations — attractors — because the feedback loops reinforcing them are stronger than any single perturbation. The healthcare system's event-based attractor is extraordinarily stable, not because anyone designed it to be, but because it self-reinforces at every level simultaneously.
Each actor responds rationally to the incentives in front of them. No individual drives this. The system is the driver.
The payment architecture rewards what it can code, and hard transitions generate codes. Clinical trials are designed around hard endpoints because that is what the payment system will eventually act on. EVPI analysis — health economics' tool for pricing uncertainty — is precisely calibrated to epistemic uncertainty, the only kind it can resolve, and structurally blind to the aleatory, ontological, systemic, and existential uncertainties that constitute most of Harold's, George's, Francis's, and Victor's actual experience. Guidelines derived from event-based trials produce protocols that optimise for hard transitions. Clinicians trained in those guidelines optimise for what the protocol rewards. Commissioners fund what the guidelines specify. And the payment system rewards the funded events.
This is not a story about bad actors. It is not even primarily a story about underfunding. It is a story about a system stuck in an attractor, where every rational response to local incentives reproduces the global pattern. Knowing this does not dissolve it. The perturbations that can shift an attractor are not primarily arguments. They are structural changes — new feedback loops, new things being measured, new forms of accountability that make the invisible visible and reward different behaviours.
But there is a prior step, and it is also the more hopeful one. Before a system can redesign itself around something different, it needs a different frame — a different way of imagining what health and care are for. That is where regenerative thinking enters.
The regenerative turn — principles from living systems
Regenerative economics begins with a simple but radical reorientation: the economy is not the system. The economy is a subsystem — embedded in society, embedded in the biosphere — and its health depends on the health of the larger systems within which it operates. The extractive economy maximises outputs from the subsystem while externalising costs onto the larger systems. The regenerative economy asks a different question: what conditions support the whole system's capacity to generate life?
Kate Raworth's Doughnut Economics gives this a visual form: a floor of social foundation below which no one should fall, a ceiling of ecological limits beyond which we must not go, and the safe and just space in between where people can actually flourish. The goal of economic activity is not growth toward an edge. It is the maintenance of the conditions in which the whole can thrive.
Applied to healthcare, the same reorientation offers something genuinely new. The event-based system is extractive in a precise sense: it maximises measurable clinical outputs — encounters, admissions, procedures — while externalising the costs of the soft drift onto patients, carers, and the long-term reserve of the system itself. The regenerative alternative asks not "how do we manage more events more efficiently?" but "what conditions support the whole person's capacity for health over a lifetime?"
John Fullerton's eight principles of regenerative vitality, drawn from the patterns that complex living systems use to sustain themselves, translate with striking precision into the language of care.
Care as belonging, not delivery
The patient is not a recipient of services. She is a node in a web of biological, social, and relational systems. Regenerative care recognises this and designs for the whole web — not just the clinical encounter at its centre.
Health as multi-capital
Biological reserve, relational continuity, functional capacity, and the patient's own agency are all forms of capital. A system that depletes relational capital to generate clinical throughput is not efficient — it is extractive. The depletion simply arrives later, as a hard event.
The patient as sensing node
The patient is the most sensitive instrument available for detecting soft transitions. George's notebook, Francis's monthly call, Harold's wife's observation of his breathing on the farm — these are not soft data. They are distributed intelligence in a sensing network the system has not yet learned to receive.
The margins are where the richness is
In ecology, the greatest biodiversity is found at the boundaries between systems. In care, the specialist nurse at the interface between the clinical and the personal, the community pharmacist at the boundary of the formal system, the voluntary sector advocate who sees what the system looks like from the patient's side — these edge nodes hold the most valuable intelligence. The system routinely strips them of recognition and resource.
Probe, sense, respond — not plan and execute
Complex systems cannot be governed by fixed plans. They can only be navigated by rapid, attentive iteration: act, observe what happens, revise. The Sussex model did not implement a strategy. It redesigned a feedback loop. That is the scale at which complex systems actually change.
Holding tension rather than resolving it
The system that resolves clinical uncertainty into a confident diagnosis — whether or not the complexity warrants the confidence — is not managing well. It is managing efficiently. The difference matters. Good care, like good governance, requires the capacity to hold what is not yet known and act appropriately within that holding.
Commonland's 4 Returns Framework — developed for landscape restoration but structurally portable — offers a way of thinking about what "return" means in a healthcare context. It argues that sustainable restoration of degraded systems requires investment across four inseparable forms of return: Inspiration (hope and purpose), Social (relationships and connections), Natural (the substrate that makes future life possible), and Financial (sustainable economic activity). Each depends on the others. Optimising for one alone produces a system that collapses what it claimed to restore.
The return of hope
Harold, George, Francis, and Victor all need something the event-based system rarely offers: an honest account of what is happening, held with enough warmth and skill that it does not become a sentence. The clinician who can model a functioning relationship with uncertainty — who can show, by their own practice, that it is possible to act well in the presence of irreducible not-knowing — is offering something that changes lives. That is a return. The system does not currently measure it.
The return of relationship
Relational continuity — being known over time by the same clinician or care team — is the single most efficient mechanism for detecting soft transitions. It is also the resource the system most reliably destroys through fragmentation, rotations, and the administrative pressure that drives clinician time away from the patient and toward the screen. Measuring and protecting relational continuity is not a sentimental gesture. It is a structural requirement of the regenerative model.
The return of biological reserve
The soft transition is, biologically, the loss of reserve — the slow depletion of the functional margin that allows recovery from acute insult. Managing the soft transition is not preventing disease. It is protecting reserve. The patient who arrives at an acute event with preserved functional capacity, intact relational support, and a care team that knows their trajectory is a different person from the one who arrives depleted on every dimension. That difference is the natural return on a decade of better-oriented care.
The return of sustainability
The event-based economy is, in the long run, not economically sustainable. It optimises for short-term throughput and externalises the costs of soft drift onto later, more expensive events. The regenerative alternative — trajectory-based contracts, long-horizon population responsibility, payment for relational continuity — is financially more durable precisely because it manages the processes that produce the events the current system pays to treat. This is not idealism. It is a different actuarial calculation.
Wayfinding — a practice, not a destination
This is the point at which essays like this one are expected to produce a conclusion. A prescription. A framework with numbered steps. We are not going to do that — not because the conclusion is elusive, but because the conclusion of this body of work is that the demand for such a prescription is itself part of the problem. The impulse that wants a roadmap is the same impulse that built the event-based system. Give me certainty. Give me the plan. Give me the thing I can audit.
What we can offer instead is a bearing. The Polynesian navigator does not have a map of the Pacific. She has a practice of attention — a set of relationships with the ocean, the sky, the swell patterns, and the accumulated experience of every voyage her tradition has made before her. She does not know which wave comes next. She knows how to read the ocean. That is the model.
Wayfinding in healthcare looks like this: clinical teams that track trajectories, not just thresholds. MDTs that hold acknowledged uncertainty as information rather than collapsing it into a confident label. Care systems that reach the knowledge held outside the clinic — in the carer's observation, the patient's self-generated data, the specialist nurse's monthly call — and treat it as the distributed sensing it is. Research and commissioning that ask not only what it is worth to resolve uncertainty but what it costs when the system projects certainty it does not have. Contracts that pay for preserved reserve, not managed events.
None of this requires a national transformation programme. Each of these moves is available now, within existing institutional authority, to clinical teams and commissioners willing to make them. The Sussex Kidney Unit did not wait for a national strategy. It redesigned a feedback loop. The MDT that records "genuinely unclear — monitoring and review" is practising Talmudic epistemological humility in clinical governance. The specialist nurse who calls Francis every month is a Polynesian sensing node in a distributed care network. These things are already happening. The task is not to invent them. It is to name them, fund them, and design for them intentionally rather than despite the system.
It is not the abandonment of evidence. The Polynesian navigator's navigational myths were empirically grounded, revised across generations of feedback, and capable of navigating routes that instrument-free Western sailors could not replicate. Adaptive complexity science is science. Talmudic hermeneutics is rigorous. Indigenous fire management is not superstition — it is adaptive ecology. The question is not evidence or complexity. It is which kinds of evidence, gathered by which methods, are adequate to the complexity of the systems they describe.
Wayfinding in healthcare is not anti-scientific. It is better-calibrated science: science that knows the limits of the RCT for aleatory uncertainty, that designs trials with trajectories rather than hard endpoints where those are the right question, that includes the patient's longitudinal self-observation as evidence rather than treating it as anecdote, that measures relational continuity alongside 30-day readmissions and takes both seriously.
The hopeful thing — and this essay is intended to be hopeful — is that the structural analysis points directly at the structural levers. If the system is stuck because payment rewards hard events, then the lever is payment reform: trajectory-based contracts, long-horizon capitation, explicit funding for relational continuity. If the system is stuck because trials are designed around hard endpoints, then the lever is trial design: adaptive trials, longitudinal patient-generated evidence, explicit assessment of soft-transition pathways. If the system is stuck because AI amplifies certainty-seeking, then the lever is procurement: tools that probe rather than alert, that show trajectory rather than threshold, that protect rather than replace the relational capacity of the clinician.
None of these levers is hidden. What has been missing is not the knowledge of where they are, but the confidence to reach for them — the confidence that comes from a coherent account of why the current system is failing, what better would look like, and why the alternatives are not naïve idealism but the better-calibrated response to a genuinely complex reality.
The inheritance is open. The Polynesian star compass, the Babylonian archive of irresolved disagreement, the adaptive fire management of a country that has been burning and regenerating for forty thousand years — these are design principles, available now, waiting to be translated into the institutions that need them most.
A bearing, not a map — what we are navigating toward
Harold is still in the car. His daughter drove, and he has been told something he does not yet have words for. His wife, at home, has been noticing changes in his breathing for eighteen months that no MDT has yet asked her about. That knowledge is available. The system is not yet designed to receive it.
George is still reading the trial data. He understands numbers well enough to know that the mean does not describe him, and he is looking for the individual-level information that the population-average framework cannot supply. He does not need false reassurance. He needs a clinician who can say "I do not know how this will progress in your case, and here is why that is genuinely unknowable, and here is how we navigate together in that uncertainty." He needs a practice of good attention, not a performance of confident prognosis.
Francis is still preparing his list of questions for the next appointment. He trusts the system more than it has always deserved, and he is asking, beneath the clinical questions, a different one: what does a good life look like from here? The existential uncertainty he carries has no clinical code. It will not appear in an EVPI calculation. But it is the uncertainty that shapes every decision he makes, every morning he gets up and chooses how to spend the hours he has. A care system designed for the whole person would hold space for that question. It would not resolve it. It would navigate it alongside him.
Victor is still at dinner. He has chosen to live in the present tense, and the system regards this as non-compliance. He would recognise himself immediately in the Polynesian navigator's etak system — the island is stationary, the ocean moves — as a philosophical strategy for managing irreducible uncertainty that serves him far better than the population statistics he has set aside. His clinician does not need to agree with his choice. She needs to respect that it is his to make, and to navigate the relationship accordingly.
These four people are not edge cases. They are the expected outputs of a complex biological system meeting a complex social one. Most patients are Harold, or George, or Francis, or Victor — living in the soft transitions, carrying uncertainties that the event-based system cannot price, making adaptive choices that the certainty machine cannot recognise as rational. A healthcare system worthy of them is not one that has resolved these difficulties. It is one that has learned to hold them honestly, navigate them attentively, and sustain the relational capacity through which they can be lived with dignity.
That is the regenerative turn — not the elimination of uncertainty but the restoration of the conditions under which uncertainty can be navigated well. Not the map, but the practice. Not the destination, but the bearing. The ocean is genuinely uncertain. It has always been. The question is whether we are willing to learn, at last, to read it.