Documentation and the Claim Describe the Same Visit, or They Do Not

An audit is not asking whether the visit went well. It is asking whether the record and the claim describe the same encounter, and those two documents are produced separately.

A claim is a set of assertions about a note written earlier

A claim says that this patient was seen on this date, that this was the problem, that this was done about it, at this level of complexity, by this provider. Every one of those is checkable against the record, and the record came first, written by someone with a different purpose in mind. The physician was writing to remember the patient and to tell whoever sees them next what happened. The claim describes the same twenty minutes in a payer's vocabulary.

Most of the time the two agree, because the same visit produced both. What is worth looking at is the ordinary case where they drift, since almost none of those cases involve anyone being careless.

Nobody has to be careless for the two to disagree

The shapes repeat. A code gets picked from a short list of favorites rather than from the note on the screen. A diagnosis comes off the problem list because it was sitting there, not because it was addressed that day. A charge is entered from the schedule instead of the chart, which is faster and usually right.

Templates do something similar with less noise. One that pre-populates a normal exam produces a record of an exam at a level of detail that may not match what happened, and it does that without anyone making a decision. Read later by a person who was not in the room, the template's sentences and the physician's sentences look the same.

Charge lag has its own version of this. A charge entered the same afternoon gets written next to an open note. A charge entered nine days later tends to get written next to the schedule, because the note is the slower document to open and read.

Documentation habits outlive the rules that shaped them

Office visit levels moved in 2021 and again in 2023, onto medical decision making or total time, and largely off the counting of history and exam elements. Plenty of notes still carry the older shape: a full review of systems, a complete exam, and then a few lines where the decision making sits.

That is now the part carrying the level, so a long note can be thin exactly where it is being read. The reverse happens too. A physician works through a genuinely complicated patient and writes four sentences about it. The thinking was done. Whether the record shows it was done is a separate question, and the record is the only part that travels.

Who performed the visit is the quietest disagreement

When a claim goes out under a physician's number for work an advanced practice provider took part in, it is asserting something specific about who did what, where, and with whom present. Incident-to and split or shared billing carry their own conditions, those conditions differ between Medicare and commercial plans, and the split or shared rules in particular have moved in recent years.

So the note is asked to carry information it was never built to carry. Clinically it rarely matters which of two clinicians performed which portion of an encounter, as long as the plan is right, and a good clinical note will often not say. On the claim that is the whole question.

What a records request can and cannot evaluate

A request from a payer, a TPE review or a RAC contractor is a request for the document behind the assertion. The person reading it has nothing except the chart, and no way to ask a follow-up question. What they are in a position to evaluate is whether the record supports what was billed, which means care that was excellent and written down sparsely reads, on paper, much like care that was not delivered.

The operational half of that is the part a practice manager can work on: where the two documents come from and how far apart they usually sit. What a particular rule requires of a particular practice, how it applies to a specific arrangement, and what to do about a request already in the mail are questions for a coding professional or a healthcare attorney reading the actual charts.

The comparison nobody makes until it is required

The comparison itself is cheap. It involves taking a small number of recent visits, reading the note and then reading the claim that went out for it, and asking whether someone holding only the note would arrive at that code, that diagnosis and that rendering provider. An afternoon covers it.

What comes back tends not to be a scattering of unrelated errors. It is usually one or two habits, belonging to one or two people, repeating: a template that overstates an exam, a diagnosis that rides along on every visit, a shortcut in how a code gets chosen. Those are findable in advance and they are findable during a review. The difference between the two is mostly who is asking.