Lessons Learned from Consulting: Institutional Improvement of ED Clinical Documentation

Providers should have the choice of bringing into the record what they deem to be relevant and important.

I am currently doing a project assessing emergency physicians’ documentation from disparate sites, and I am finding that they each have a different electronic health record (EHR) and/or template.

Some are pleasing, visually and/or informationally, to me, and some are not. I have no formal training in document design, but I thought I would share my observations and thoughts with you. I am going to avoid saying “in my opinion” or, “I think…” for every point; consider it implied.

We must recognize that the printed version of a record may look quite different from the electronic version. You also navigate differently in the digital world. I was reviewing PDFs, but so would an auditor or an out-of-network caregiver.

I will preface with this: I have certain biases. I was trained in the days of paper, and there was a very specific order in which you recorded information, I suppose by convention. Documentation is intended to convey clinical information to a reader. Anything that frustrates or distracts the reader is counterproductive. One of the challenges of today’s documentation is that there are still old-school and new-age readers and their preferences may be contradictory. 

The importance of written cues is to assist the reader to find specific information quickly, so some institutions have taken to frontloading their impressions. The history and physical examination (H&P) was intended to lay out the information obtained from the patient and the conclusions drawn by the provider, based on that information. It should be expository and logical. I know readers are often pressed for time, but putting conclusions up front seems to diminish the importance of their basis. Who reads the last chapter of a book first?

I liked the systems for which the sections were easily identifiable and in a block-text narrative format. Titles of a different font, size, or offset in bold making it easy for you to find that for which you are looking.

The arrangement or alignment of text in relation to the margins can be important. Some systems had chief complaint, history of present illness (HPI), review of systems (ROS), physical exam (PE), medical decision-making (MDM), and the assessment and plan in a column on the left, and medications, allergies, past medical/surgical/family/social history, and orders in a column on the right. I liked the way my eye could scan down and find the information I sought quickly and easily. Those readily identifiable titles marked each section.

I gravitate to the narrative. I prefer a paragraph telling a story. I appreciate setting the table with pertinent past medical history (PMH), but listing every superfluous condition the patient has ever had in either the HPI or the PMH section is distracting.

I presumed that some systems have a built-in template of fields to ensure that providers hit their elements (quality, severity, now and maximum, timing, duration, etc.). It did not offend my sensibilities if the provider told a narrative story and those elements were embedded in the paragraph at the end, adding some detail. It was offensive when that served as the entire story.

I hate embedding huge, extraneous sections of results into H&Ps. If you must pull in a radiology impression, go ahead. I don’t need to know what the patient’s radiation exposure was or what kind of contrast was used. I don’t even need to know who the radiologist was. There are parts that are not usually useful. Stop note bloat!

Interestingly, though, it also irked me when the computer only brought in abnormal values. It can’t know what a pertinent negative is. Providers should have the choice of bringing into the record what they deem to be relevant and important. In the electronic world, I would just click on the lab results section and peruse, myself.

White space can be helpful or detrimental. Double-spacing of the entire document was irritating. Double-spacing between sections was helpful.

The PE normals should be reasonable. Don’t waste my reading time with “right external ear normal. Left external ear normal,” especially if there is no point to examining the external ears. If you use template normals, mindfully edit. Trustworthiness goes out the window when your conclusion is “chest wall pain” and you never document examining the chest wall for tenderness.

Care should be taken to avoid orphaning. This is when a sentence or thought ends up on the next page. They can be overlooked and might contain important information.

I recommend that caveats and attestations be enterprise-standardized, with compliance’s input.

• “At least 10-point ROS (review of systems)” does not mean full credit to the Centers for Medicare & Medicaid Services (CMS). You must say “complete” or “14-point.”
• There should not be “0 minutes of critical care time” asserted with a sentence of “there was a high probability of clinically significant/life-threatening deterioration in the patient’s condition, which required my urgent intervention.” You can have a critical patient for whom you do not claim CCT (carve out procedures, < 30 minutes), so this attestation can exaggerate how sick a patient is when you are just trying to say that the patient was not a critical care patient.
• Providers must read and edit their voice recognition. No disclaimer nullifies lack of editing or revision. You are always responsible for your output. Instructing a reader to contact you for questions is silly if you don’t give them your contact information. And do you really want a call at 2 a.m. for clarification? Just fix it!

I’m going to make a plug for robust MDM. The best records explained how the provider interpreted the studies and why they thought what they thought. They detailed repeat examinations in response to treatment. They included discussions with the admitting physician and with the patient and/or family. If a patient was being admitted for a condition that could be addressed with observation or admission, explanation of thought processes supports medical necessity of the disposition. MDM may soon be the CPT® level-of-service determinant.

Some systems force the provider to select an ICD-10-CM code in the impression section. The providers need to be educated that you can’t affix (rule out) after the impression. In those systems, uncertain diagnoses belong in the MDM section, and only certain diagnoses should be coded. For others, I always recommend marrying signs and symptoms with uncertain diagnoses. This explains your thought process and gives both inpatient, outpatient, and professional coders something to pick up.

The best documenters included comorbid conditions in their impression list. Opportunity to query for nuanced ICD-10 verbiage could be detected by the combination of the HPI, the MDM, and the impression list (e.g., patients with osteoporosis with fall from standing sustaining multiple pelvic fractures).

The best documenters also crafted hearty supervisory or attending notes. Whether it is a split/shared visit with an APP or an attending note for a resident, I recommend adding your own documentation to demonstrate that you had a face-to-face with the patient and were value-added to the decision-making.

Quantifying the quality of documentation can be difficult. It’s easy to recognize and remediate bad documentation or commend excellent documentation. It is more challenging to take mediocre documentation and transform it into excellent documentation. You must identify what is lacking or suboptimal and give the provider concrete examples and suggestions.

 Just telling a provider that their documentation needs to improve is inadequate. If they knew how they likely would be doing it better. At the very least, your institutions need to do what they can to facilitate the process by making good templates and formats.