If the note is the wrong atomic unit for clinical information, what is the right one? Chapter 7 gave the theoretical answer: the individual clinical fact, which allows maximal flexibility of storage, recombination, and display. But facts do not float free. They need an organizing container — something that gathers the potassium trend, the diuretic adjustments, the consultant’s opinion, and the clinician’s evolving reasoning into a coherent object a human can pick up and read. This chapter argues that the container is the medical problem — and works through what that implies about what should persist, what should carry forward, and what should stay behind.
11.1 The unit of meaning
The traditional note tries to do too much: legal record, billing artifact, communication handoff, memory aid, and workspace, all at once. Unsurprisingly it does none of them well. But its deepest mismatch is simpler: the note’s boundaries do not correspond to any boundary in the patient’s clinical reality.
A visit is an administrative event. It begins and ends by the clock. The patient’s problems do not. A patient with CKD stage 3 doesn’t have CKD stage 0 next week because a new encounter opened. The child with obesity doesn’t shed years of counseling because it’s a new visit. Chronic disease changed the calculus of documentation entirely: when a patient carries hypertension, diabetes, CKD, and depression — each with its own medication history, lab trajectory, and specialist involvement — no encounter is self-contained. Every visit is downstream of every prior visit. The clinical picture is inherently longitudinal, and a record organized by date is a poor container for longitudinal information.
Clinicians already know the real unit. Ask what’s going on with a patient and nobody answers in encounters; they answer in problems. The blood pressure is finally controlled on three agents; the A1c is drifting; the weight loss is still unexplained. The problem is the thread along which reasoning, evidence, and intervention accumulate. It is the unit of meaning — and the record should be organized around the units of meaning, with everything else (visits, notes, bills) derived as views.
This is not a new aspiration. Weed’s problem-oriented medical record made the case more than half a century ago, and generations of clinicians have been taught to think in problem-structured assessments. What has changed is feasibility. Under paper — and under paper’s digital reproduction — a problem-oriented record demanded constant manual curation, and it decayed the moment attention lapsed. A workspace with fact-level editing, version history, and (as Chapter 13 will discuss) modern language technology can maintain problem structure as a byproduct of ordinary documentation rather than an additional chore. The idea was never wrong. It was waiting for its infrastructure.
11.2 Persistence: you shouldn’t have to know to look
Organizing by problem enables the property that matters most. Call it information persistence: you should not need to know to look for something in order for it to be available to you.
If a patient has had a myocardial infarction, that fact shouldn’t have to be hunted for. It should be apparent the moment you open the chart — plainly, immediately, without a search, without knowing which note from which visit to dig into. Today that is not how it works. The place where the MI should live — problem list, medical history — is divorced from the record itself, which is really a pile of notes. And a note can be entirely accurate and still be functionally useless beyond the moment it was written.
Think about the cardiologist who does excellent work for a hospitalized patient. If that work’s key facts don’t persist somewhere durable and obvious — somewhere outside the body of a single note — then for all practical purposes they might as well not exist. Someone would have to know to look for that particular note, written at that particular time, about that particular problem. The knowledge lived in one clinician and in a document only that clinician knew to write. You can write something accurate, beautifully organized, fully reasoned — and if you leave the practice, or retire, or die, no one necessarily knows what’s going on with that patient. Even a perfect note can fail the patient.
Persistence is what the copy-paste behavior of Chapter 8 was trying to buy. Clinicians duplicate because duplication is the only persistence mechanism the note paradigm offers: if it isn’t re-pasted into something recent, it effectively vanishes. Give information a durable, problem-anchored home — document once, update in place, visible by default — and the rational motive for duplication disappears, along with the bloat, the conflict, and the chart lore it produces.
And medicine’s trajectory makes persistence more urgent every year. People live longer; guidelines multiply; goal-directed therapies grow more intricate. As medicine gets more complex, we need simpler ways to see what is going on with a patient — which means the record itself has to hold the complexity, rather than requiring every reader to reconstruct it.
11.3 What carries forward — and what stays behind
Problem-orientation forces a design question the note never had to answer, because the note treated every word identically: the results of a routine lab review got the same status as the updated heart failure plan — both just text, both buried together, both unranked when the next clinician opens the chart.
That uniformity is a structural mistake, because clinical information comes in two durably different kinds:
Visit-specific information belongs to the encounter. The purpose of today’s visit, who was in the room, the procedure performed, the orders placed that day, the acute complaint’s context. It matters for the record of the encounter — medicolegally and administratively — but it has done its job once the visit is documented. Forcing it forward clutters the chart and buries signal.
Patient-specific information belongs to the problems. The ongoing management of the hypertension, the adjusted metformin dosing, the evolving picture of the knee pain — and, crucially, the clinical thinking attached to each. This must carry: update, accumulate, and be present the next time anyone opens the chart.
Clinicians make this distinction constantly and automatically — in conversation, in their heads, in the visit itself, where the acute and the chronic are managed in the same breath. The failure of note-based systems is that they collapse it: everything in the note lives and dies together. A problem-oriented workspace can honor both kinds. The encounter keeps its own space — a defined record of what happened today. The problems keep theirs — continuing threads in which today’s assessment connects to last quarter’s and becomes the foundation for the next. Both coexist in the same visit because that is what a visit is: a moment of contact that touches some problems, opens others, and generates some purely local facts.
Getting the split right is harder than it sounds — you have to know what to carry, what to let go, and how to keep the flexibility for visits that don’t look like the last one. But it is exactly the distinction that makes the difference between a record that accumulates and a record that orients: the chart that answers “what is going on with this person?” before you’ve asked.
11.4 The note as byproduct
None of this abolishes documents. Referrals need letters; payers need encounter artifacts; the law needs frozen snapshots. The reversal is in the order of operations: instead of organizing the record around the note and bolting structure on afterward, organize the record around the patient’s problems and topics — and let the note fall out as a byproduct.
A visit’s “note,” in this world, is a rendered view: the problems addressed today, each with its update; the visit-specific context; the orders and attestations — assembled automatically from the workspace and frozen as the encounter’s legal artifact. Nothing about it requires the clinician to re-summarize, re-paste, or re-type. The same underlying structure can render the referral letter, the billing documentation, and the patient’s own summary, each shaped for its reader.
There is also a quiet payoff for the smallest units of clinical work: the follow-ups. Every clinician makes dozens of small commitments a day — the repeat BMP in two weeks, the colonoscopy to schedule, the referral to chase — and today they live in whatever ad-hoc system each clinician improvises, because the EHR’s task modules demand double entry (documented once in the note, again in the task list) and so go unused. Some clinicians solve this by making the note itself the task manager — a running list at the top of every note: things we’re working on, things outstanding, things to do. Written once, read first, one source of truth. It works precisely because it obeys this chapter’s principle: the plan and its tracking live in the same place. In a problem-oriented workspace the same follow-ups are structured objects born directly from the documented plan — no second entry — resurfacing when due. Closing the loop is a patient-safety issue before it is a productivity issue: every uncaptured task is a small bet that somebody will remember, and the lost bets are the abnormal result unaddressed, the titration that never happened.
The chart knows everything that was ever written into it, and yet it knows almost nothing about your patient. Those are not the same thing. The problem is the unit that closes the gap: it is how the record stops being an archive of what was said and becomes a description of what is true. Medicine has spent decades improving the quality of what gets documented. The next decades belong to what gets surfaced — to records that don’t just archive the patient’s story, but tell it.