Coding
Preventive, problem, or both
The fuzziest recurring billing call in a gynecology practice, and why a scribe that stops at the note leaves the hard part undone.
A patient comes in for her annual. Twenty minutes in, she mentions the bleeding that has been worrying her for two months. You take a history, you examine, you order something, you change a plan.
Was that one service or two?
This is the most common genuinely ambiguous billing decision in a gynecology practice, and it is decided many times a week, usually at the end of a long day, usually by memory.
Why a scribe that stops at prose does not help here
Turning a transcript into a fluent paragraph is close to a solved problem. It is also not the expensive part of documentation. The expensive part is the judgment sitting on top of it — and the judgment is invisible in the finished note unless something deliberately puts it there.
Both answers to the question above can be defensible. What is not defensible is a note that does not show which one you chose or why. That note is fine until somebody reads it back a year later, and a year later is exactly when it is read.
What the draft is asked to do instead
Not to pick a code. To make the reasoning explicit and leave the decision with the people who are accountable for it:
- Classify the visit — preventive only, preventive plus problem, or problem only.
- List what fell under the preventive service, so the components are not silently double-counted.
- List the separately identifiable problem work, quoting the visit rather than characterising it.
- Name what was considered and deliberately not billed separately — and keep that block in the signed note.
That last one is the one people find strange, and it is the most useful. A note that records what you decided against is a note that defends itself. It is also the honest form of the document: the decision happened, so the record should contain it.
Where it deliberately stops
A modifier that separates a problem visit from a preventive one is only appropriate when the problem work was significant and separately identifiable. That is a test, not a habit, and it gets applied as a test — with each part having to be answerable from the visit.
And where the billable code depends on something the software genuinely cannot see — a patient’s age band, whether she is new or established — it does not guess. It emits a placeholder that is visibly not a code, because a placeholder somebody has to resolve is safe and a real code that happens to be wrong is not.
Three plausible codes where only one is right is not helpful. It is a coin flip wearing a lab coat.