The Process of Doctoring

From the outside, doctoring looks like reaching a diagnosis. The diagnosis matters, but it is not what you are managing. Every patient has a latent true state: the real condition, trajectory, and reachable futures you cannot see directly, only infer from signals. What you can build is a working model of it, and you manage the patient through that model. The work is a loop. Build the model honestly, reduce the uncertainty that matters, act, and let the patient's response correct it. The diagnosis is only one handle inside the model. What follows is that loop, one turn at a time.

1 · Observe

A patient arrives short of breath. At first that is almost all you have: a complaint, an appearance, a set of vital signs. Behind them is the patient's latent true state, but you cannot see it directly. You see the shadow it throws. Pneumonia throws that shadow. So does heart failure, pulmonary embolism, a COPD flare, anemia, acidosis, ischemia, or two at once. The same presentation underdetermines the diagnosis: many states fit the few signs you can see. The true state is one thing. Your access to it is partial. Doctoring begins in that gap.

2 · Model

Because the true state is hidden, you cannot honestly name one diagnosis at the start. To do so is to invent certainty you don't have. What you build instead is a working model, and early on its honest form is a distribution: a weighted set of the states that could be producing these signals. That is what a differential diagnosis really is. Not a list of diseases to cross off one by one, but your current probability cloud over what may be happening. Probably pneumonia. Heart failure still plausible. Embolism less likely but dangerous. Malignancy unlikely today, though it would explain the longer story. The discipline is to carry that uncertainty without freezing, to act inside the cloud rather than wait for it to collapse.

3 · Personalize

Each item in that cloud is a generic model. "Pneumonia" is everything common to everyone who has it: the expected findings, the usual course, the standard treatment. That is its power and its limit. A label is, by construction, what cases share. To be a label at all, it must throw away whatever makes one case differ from another.

What it throws away is the patient. Severity, organ function, the drugs already on board, the goals this person holds, the home they will or won't go back to. None of that is in the word "pneumonia." These are not details added on top of the diagnosis. They are the residual it deleted, and they are usually where the action lives.

A healthy twenty-two-year-old with mild pneumonia and an immunosuppressed seventy-eight-year-old with pneumonia, hypoxia, acute kidney injury, delirium, and no one at home share the same label and almost nothing else. Same diagnosis, opposite management. Your working model is the generic disease model conditioned on this person, and the conditioning is what decides the action.

4 · Aim

The diagnosis is a handle, not the object. What you are moving is the patient's true state, and the target is the best achievable future state. That phrase is doing real work. The target isn't fixed at "normal." It is whatever future is still reachable given the biology, the clock, the resources, and the patient's goals. Cure, sometimes. Often something narrower: stabilization, a safe disposition, time bought for the picture to declare itself, comfort when biology has set the ceiling. The goal is a variable, not a constant.

5 · Probe

Between where the patient is and where you want them sits information, and not all of it is worth having. A test, a question, an exam maneuver, a stretch of observation earns its place only when it reshapes the cloud in a way that changes what you do next. The test is simple. Would your next action change if the result came back the other way? If not, the result is a fact you don't need. If so, it may be the only fact you need.

6 · Act & update

You act when the cloud is sharp enough to justify the move, or when waiting has become more dangerous than acting. In medicine that is often, because certainty is usually unavailable, unnecessary, or too slow to be safe. Then the action becomes information. The wheeze that breaks with a bronchodilator. The pressure that comes up with fluids. The pain that worsens through a reassuring workup. The patient who returns worse after a plausible discharge. Each is the world answering. Response is evidence. Trajectory is evidence. Failure to follow the expected course is the loudest evidence of all. You compare what the model predicted against what the patient did, and where they diverge, the model is wrong, not the patient. Then you update and go again: a sharper picture, a narrower aim, the next probe. A closed loop, not a one-time classification.

The same loop, run well

Novice and expert run the same loop. What separates them is how well each turn is done. The expert holds a wider range of possibilities, so the true state is more often somewhere in the model at all, and a possibility that never enters the model can never be reasoned toward. The expert weights it better, by probability and by danger both, keeping a low-odds catastrophe in view because missing it is unaffordable. The expert sees disease as a process in time, not a snapshot, and asks what the next hour should look like if the model is right. The expert reaches for the cheapest signal that discriminates, not every signal available. And the expert keeps a wider set of moves: stabilize, temporize, monitor, escalate, refer, safety-net, watch and wait. No one can choose a move they don't know exists.

What doctoring is

Strip it down, and doctoring is the building, updating, and use of an action-relevant working model of a patient's latent true state, in service of the best achievable future state. The loop runs until the patient reaches the best state still reachable, until care passes safely to someone else, or until no action can improve the futures left to them.

What you manage is the patient's true state, the part you never see directly. What you manage it with is the working model. The diagnosis is one handle inside that model, not the model, and not the patient. The model is what lets you act before the truth is in.

This is the clinical face of something the rest of this series treats more formally. The working model is the belief-state object of Part 7, written as narrative instead of structure. Reframing care as binding a general model to one specific person is the binding problem made constructive. And the question that decides which information is worth gathering, would the next action change?, is the same one that decides which distinctions matter in any retrieval or coordination system. The patient is just the substrate where the stakes are clearest.