Actually, pretty much everybody who’s been on EHR for a while talks about it. Sometimes loudly, to anyone who will listen.
F-R-A-U-D.
As in, “Your documentation says you did that, but you didn’t really.”
Did you?
You Didn’t Mean To, Honest
We know.
And hopefully, you’re not even guilty — yet.
But it’s an occupational hazard that you must be aware of in order to guard against: stating in your documentation that you did much more than you actually did, or reasonably could have. And perversely, the more adept you get at your record keeping, the more likely you are to fall into this trap.
It starts with one of the foundational tenets of patient encounter workflows: documenting by exception.
Most systems are based — have to be based — on some version of this. You can always document each and every aspect of a patient visit, button click by button click, checkbox by checkbox, word by word.
That’s accurate, but excruciating. And the vast majority of visits, unless they are completely novel, are based on clustered common data elements, that can be pre-populated in the chart to save you time.
For example, the common cold usually gives patients some collection of the following:
- nasal congestion
- rhinitis
- mild dry or slightly productive cough
- mild to moderate sore throat
- no high fever or shaking chills
The exact details, like whom the patient caught it from, and how many days they’ve had symptoms before presenting to the office, will vary from person to person, and these can be added individually, as well as particular elements (say, if a particular patient actually had a 101 degree fever).
Documenting by exception (DBE) takes advantage of this data clustering tendency of most conditions, diagnoses, and syndromes, to insert the cluster as a block into the medical record with one button click. This enables you to concentrate on the few corrections — the exceptions — that make the visit unique.
Whether you use it for the history, physical exam, systems review, or plan, DBE can save you major time when dealing with problems that closely fit the expected description. You can “correct the mold” when things don’t fit, or go into unique, totally-from-scratch documenting at any time (”A 135 y.o. Tibetan monk got off the plane and slipped on foil Doritos bag, straining his left orbicularis oculi muscle and causing an apparent subungual hematoma on his right 2nd toe”).
So What’s The Problem?
One problem is, even if the DBE-generated documentation is 100% spot on, nobody believes it.
That may be a bit of an overstatement, but face it: if you’ve seen one auto-generated description of a head cold, you’ve seen them all.
Working with EHR’s, even the most pro-tech admit that reading their own notes can be like reading someone else’s auto parts list. There’s a certain lack of identifiable individuality when you read an account like this:
The symptoms began 3 days ago. Onset from ill contact. Symptoms are rated as mild. Symptoms are gradually worsening. Exacerbating factors include laying down and cold air. Pertinent positives include dry cough, sore throat, rhinitis. Pertinent negatives include fever, chills.
As opposed to this:
Mr. Smith’s noted a dry, mild cough 3 days ago, followed immediately by a runny nose and mild sore throat on day 2. Recalls coming back from Denver on a commuter flight last week, “everybody was coughing in the cabin.” Denies fever or chills; but cough does worsen at night, especially when laying down. The ST and rhinitis are gone as of yesterday.
You’d be hard pressed to recall the visit based on account #1; you could probably visualize Mr. Smith’s face and the entire visit just by reading account #2.
Back in the day when artificial intelligence and expert systems were in the realm of science fiction, the Turing Test was pretty commonly quoted. If you couldn’t tell the output on a computer screen between a human typing in another room, and computer replying to your questions, then the AI had reached a stage where the difference in thinking ability/intelligence was moot.
With electronic medical records, we’re kind of stepping backwards if our human output is being compared to a machine’s. And the perceived difference isn’t a question of intelligence; it’s potentially a much more serious matter of calling into question your professional integrity.
Problem #2: There’s Just No Way, Dude
You can make a good case that, exciting prose be darned, you documented exactly how things went with Mr. Smith, even though you clicked the “head cold” history button to get 95% there. Medico-legally, you’re obligated to document accurately and thoroughly, not melodramatically. Dry is OK.
A potentially more serious problem is when you get a little fast and loose with the button clicks.
Whipping around that mouse cursor, clicking everything in sight as being broadly normal, in other words.
Now really, when Mr. Smith came in with his probable head cold, did you really do a systems review that included the following?
- Neurological: negative for headaches, dizziness, seizures, loss of consciousness
Did you really ask him if he got dizzy or had headaches with his common cold? Much less if he fainted or seized on the trip?
Did you document what amounts to a complete annual physical for a 6 minute urgent care visit?
So What’s A Conscientious Person To Do?
First, you’ve got to familiarize yourself with the DBE premises and get comfortable with its limitations.
There’s no way around it — you will need to use it to get through your day. And if you think about it, you’ll admit to using it in your ol’ pen and paper days, too: does “nl” followed by a vertical line going down the length of the PE section, plus the few non-normal details scribbled here and there sound familiar?
DBE isn’t an artifact of EHR’s, it’s an artifact of busy human beings.
The simplest fix is to carefully read what those normal checkboxes put into the chart, and to make sure you ask or examine to match that documentation. If you don’t routinely percuss the lung borders, maybe it’s time you should.
If your system allows you to reword the verbiage assigned to those buttons and checkboxes, the next level fix is to pick words that accurately match the questions and examinations that you actually perform.
Next, consider individualizing where it counts the most: the HPI and the Plan.
Does it really matter that you percuss lung borders using the 3-fingered South Bend technique? Probably not. But a reader (or auditor) will naturally want to “hear the story” in the History section, and how you’re planning to resolve it in the Plan. That’s where the individual touches will be missed the most — especially when it comes to clarifying your thought processes for concluding what you did, and taking the actions that you did.
Free text is AOK here.
There’s an eternal argument that will rage until the sun burns out, between the “I hate those busy pages with all those buttons next to signs and symptoms” camp, and the “I love those buttons, I can click a lot faster than you can possibly type” camp. You can join in the fun at any time in your medical career, depending on your control issues or lack thereof.
This isn’t about that. The folks arguing that matter are by definition not at real risk of fraudulent documentation — both sides are being hyper conscientious, in their own way.
You’re at risk if you’re using needed shortcuts, but not paying periodic attention to what is being cut. You will drift, as all complex systems do, unless you look down at the road and correct your course, once in a while.
