I arrived at the neuro ICU at 5:30 a.m. to read up on my new patient before rounds. The resident on duty the day before had accepted a transfer at 7:00 p.m. and documented the following in his note:
Mr. X is a 72-year-old male with a past medical history of severe bilateral carotid artery stenosis who had a devastating right MCA ischemic stroke at home this morning, confirmed by CT at an outlying hospital. He was transferred to us for an MRI and further workup. The patient is intubated and completely unresponsive: GCS of 3. Pupils sluggishly reactive to light. 1+ reflexes in upper and lower extremities. Patient stable. On propofol drip. MRI ordered.
For students entering the wards, the information in a patient’s electronic health record (EHR) looks like a question stem. The style is the same and it is, after all, written by someone with more medical knowledge and training than we possess. As medical students, we read thousands of these stems in our first two years in the classroom. They are written to get us to think about differential diagnoses, diagnostic work-ups, and treatments, but there is one thing we are never meant to question — the reliability of the information provided in the question itself. For example, I must assume that the 15-year-old girl with photophobia and nuchal rigidity I encounter on Step 1 does in fact have those physical exam findings in order to diagnose her correctly. In this way, we become conditioned to trust the patient information provided to us.
This means that many of us start our third year of medical school relying on an EHR to do most of the work of discovery, ignoring the patient right in front of us. Not only can this cause us to overlook the humanity of our patients, but it presents very real risks to their wellbeing. Mr. X’s case helped teach me this lesson.
After reading the initial note, I noticed a more recent progress note, written by the night resident later that evening. It used the exact same language, likely copied and pasted:
Mr. X is a 72-year-old male … Patient stable. On propofol drip. MRI ordered.
Nearly identical wording would be used three more times in the next four hours. I read it off the page when presenting to my intern, my intern used it in his progress note, and our upper level resident quoted it directly during rounds. By then, I realized that, to us, Mr. X had become eight sentences of copied and pasted text. In hindsight, I think we were all waiting for the MRI results to guide our treatment plan, but we all failed to take a hard look at Mr. X to confirm his condition. We trusted that the person before us had gotten it right. But they hadn’t.
This process of “treating the chart, not the patient,” happens more than we like to think, though less commonly with patients as critical as Mr. X. I hoped at the time that we all reused the same language because Mr. X was indeed stable.
But during rounds, a neuro ICU attending walked into Mr. X’s room, turned off his propofol drip, flipped through the patient’s records, and began to examine him. Within 10 seconds of starting her exam, she turned to the team behind her and said, “Did anyone notice this? His eyelids are twitching. This patient is in non-convulsive status epilepticus. Who has seen this patient since he’s been here?”
Four of us raised our hands, but none could say whether the subtle eyelid fluttering was present on our exam.
An EEG confirmed our attending’s suspicion. There is no telling how long Mr. X had been seizing. And it’s unclear how much we could have done for him by identifying and treating his seizure activity earlier; his brain had already been devastated by the massive stroke he had experienced the day before. But what was clear to me was that nearly our entire team had based our individual assessments of Mr. X on a single note, rather than on Mr. X himself.
During my second year of medical school, I told a friend that I wanted the patients in my review books and exam questions “to come to life.” At the time, I was proud of the line. And I meant it. I had hoped that the patient I encountered on Step 1 would one day be a living, dying patient in my care that I would diagnose and save. I had wanted to leave the text behind and touch real patients with my own hands and heart.
But instead of seeing text come to life, I now find myself fighting the temptation to turn my patients into mere words.
As a doctor in training, I need to remind myself that everyone makes mistakes — even some which find their way onto medical records. Therefore, a note in the EHR should not be read as a question stem or fact, but as a fallible assessment of a patient’s condition at a past point in time. Patients are people, not notes; their health is dynamic, and those who are critically ill can deteriorate quickly. My experience with Mr. X changed the way I approach patient care, and his memory is a constant reminder to treat the patient, not the chart.