Why create a searchable database of cases?

Diagnostic reasoning is an integral part of medical training. As a excitable intern eager to hone my clinical skills, I asked two master diagnosticians, Drs Rabih Geha, and Gurpreet Dhaliwal, what they would recommend to improve my diagnostic ability. They both recommended frequent reading of the NEJM Clinical pathological conference (CPC).

However, as a busy (and lazy) intern, I often felt like the randomness of the weekly cases made it difficult to motivate myself to read them each week. I wanted a way to find cases directly related to my patients. Ideally, I wanted to read a case that presented similarly to a patient I admitted, or whose differential diagnosis is similar to the differential diagnosis of a patient I was thinking about.

↓

And so I decided to create a tool to better utilize the deep, rich CPC case series for case-based learning.

 
sypmtom.png

Search by symptom: I’ve used this to help expand my differential and the way to think about a case that presents with a specific set of symptoms

ddx.png

Searching by differential diagnosis: I think of this as the “reverse DDx Search”. Sometimes I have had patients where I have thought “I think this looks like Tuberculosis. In cases where experts also had Tuberculosis high on their DDx list, what were some of the final diagnoses observed? This tool allows you to return all cases where a specific Dx was on the Differential.

final_dx.png

Searching by final Dx: In some cases, I already knew the final Dx of my patient, and wanted to read more about the pathophysiology or treatment recommendations for disease. CPC cases often have great discussions about underlying pathophysiology or cutting edge science in a field relating to the disease. This search allowed me to find and look up the times were a specific diagnosis was made. It also highlights cases that were very “similar” to cases with a specific Dx, so I could read a mystery case similar in nature to my Final Dx.

 

Interested in learning more?

Here Is the accompanying paper describing the methods and results: