Although it is important to focus on findings likely to be useful in distinguishing among the diseases in the differential diagnosis, it is also important to focus on unusual findings that can change the differential diagnosis dramatically.
Example to work through
This is tragically illustrated in the case of 18-year-old Libby Zion, whose misdiagnosis and death led to rules regulating work hours for residents. Zion had 4 key findings as stated in a New York Times article about her case:
… a high fever and mysterious jerking movements. She was alternately cooperative and agitated. She had a history of depression…
The Natural Language Processing (NLP) in the software processes this free-form combination of findings to yield the following 4 findings:
The differential diagnosis based on fever alone focuses on infectious diseases. The display of the other findings are shaded with green to denote their usefulness in distinguishing among those infectious disease. But instead of prioritizing by usefulness, the findings listed by the NLP are listed in order of rarity, with #1 being the myoclonic jerks (sudden, involuntary muscle contractions that can occur in various parts of the body).
Commenting on all the findings, including the myoclonic jerks, gives a very different differential diagnosis, and the display of pertinence on the Phenotype tab of the Diagnosis screen illustrates what is often the situation, that the rarest finding, here myoclonic jerks, has the highest pertinence to shaping the differential diagnosis. (Click the image below to jump into the software.)
Using other elements of diagnosis it is straightforward to proceed from here to determine that the diagnosis was Serotonin syndrome. Zion died after administration of meperidine, the Risk Warning that is seen after clicking the red asterisk next to the disease.
The 2009 New York Times article asserts that “the exact cause of Libby’s death was never found”, but that is not true. A 2005 New England Journal of Medicine article focused on the Libby Zion case as a classic example of serotonin syndrome. The NEJM article bemoaned the situation that “clinicians cannot diagnose a condition of which they are unaware, even though the serotonin syndrome is not rare”. However, with a focus on rarity of findings and the use of diagnostic assistance, clinicians can now diagnose a condition of which they are unaware.