The previous post discussed the importance of the patient's narrative as the foundation of the medical process. We move on now to the step of documenting the patient's narrative. We must first acknowledge that the patient's narrative, despite its importance, is never recorded.
Let me repeat: we do not write down the patient's narrative.
The 'patient history' section of the medical record is a narrative written by the physician. This narrative is what the physician thinks that they heard the patient tell them.
We all know that what someone thinks they hear and what the person thinks they said are two different things. To prove this to yourself, ask the person helping you at tech support to repeat back what they wrote down as the problem you are having with your device.
In this way, it is important to realize that the 'patient history' section of the physician note is potentially an error-riddled document produced by the listener of the story, and not a credible first-hand artifact.
Because the written narrative is not written by its primary subject (the patient), error enters into the physician note during the steps of narrative translation and truncation.
Translation occurs when the physician converts expressions in the patient's narrative into specific medical terms. A patient may describe their chest pain, and the physician typically does not write word-for-word what the patient says but may shorten this description to a few words typically representing a medical term such as "angina" or "pleuritic chest pain."
This step introduced considerable diagnostic error. First, different physicians may use different medical terms to describe a series of patient symptoms. Although many terms have formal definitions, in practice, the use of words is looser and therefore we may not know if the word choice is based on a formal diagnostic definition or information colloquial usage.
Second, medical terminology is loaded. To say the patient's chest pain is 'angina' begins to frame the patient's narrative as a cardiac problem. To say their pain is 'pleuritic chest pain' is to begin to frame their problem as a respiratory problem. This framing bias starts for the physician the moment they begin to translate the patient's narrative into medical terminology. It furthermore then propagates the bias to whoever reads the note.
The truncation step occurs when the patient's narrative is shortened to what the physician perceives as the key points. In order to know what the 'key points' are, in theory, one needs to know what the final diagnosis is. Although a physician may have a diagnostic hunch and be building a real-time differential diagnosis of the final condition, ultimately, the final diagnosis is unlikely to be arrived at when the initial note is written. In this way, the truncation step, which allegedly retains the key aspects of the patient's narrative and disregards those that are not important, is prone to the error of over-including details that may be red-hearings and leaving out details that were not perceived at the start to be of importance.
Ways around these errors
The translation step can be avoided by recording in the note the patient's words as closely as possible when describing their symptoms. I've seen expert clinicians use direct quotations throughout their notes as a way to reflect this.
The truncation step is more difficult to mitigate. In theory, one could audio record an entire patient encounter and store it in the chart or transcribe an entire patient encounter word-for-word. This would retain the original 'patient's narrative'. However, it would also create an artifact that is unlikely to be reviewed, given its length. (I have a pending post on the utility of storing the entire patient encounter as a video, and how this would improve longitudinal medical documentation, but this is another post.)
It is reasonable to believe that truncation is a required step in the standard medical documentation process. Ideally, truncation would follow a semi-standardized approach, as Larry Weed would have proposed. This would mitigate against the risks of information being discarded early on, which turns out to have been important.
I recognize that there are a variety of startups and larger organizations working on transcribing the physician-patient clinical conversation into a structured note. However, that is a topic for another day.
The patient's narrative is never written down in clinical medicine. Instead, physicians translate what they think the patient said into medical jargon.
This introduces bias at the level of cognitive reasoning as well as diagnostic momentum in the word choice used. Furthermore, important details may be left out during narrative truncation, and unimportant details potentially overemphasized.
This explains why expert physicians typically pay little attention to the previous documentation of a patient's history in the clinical note and prefer to have the patient re-tell their story first-hand in undiagnosed and difficult cases. I've even seen some physicians go as far as to refuse to read the previous notes about a patient's history before speaking with the patient, as they are so worried that the other physicians' documentation of the patient's story will unduly influence their thought process.
The next post will discuss the errors derived from fitting the patient's narrative into the standard narrative structure.