Skip to main content

Differential diagnosis

Suggests differential diagnoses from what you have recorded.

Using it​

Open the Differential Diagnosis panel from the clinical tools grid and generate. The list comes back with reasoning tied to the findings you actually documented.

What it reads​

The structured assessment: presenting complaint, history, examination findings, observation sets including NEWS2, completed scores, the Clinical Frailty Scale, any ingestions, and your free text.

It is built from structured state rather than the report preview text, and the patient is only ever "Pt". See What gets sent.

What it will not do​

  • It will not invent findings. A differential is argued from what you recorded.
  • It will not assume an observation you did not take. An unrecorded parameter is explicitly flagged as unknown to the model, not treated as normal.
  • It does not rank by what to do about it. This is a list to think against, not a disposition.

Getting a useful list​

The quality of the output tracks the completeness of the input almost exactly. A completed observation set and a few lines of free text about the history change the result far more than re-running it does.

Queueing​

If it fails for lack of signal, MedJot offers to queue it. It runs when the connection returns and the result appears here, with the tool card showing that something is waiting.

The obvious caveat, stated anyway​

This is a drafting aid working from a summary of one encounter. It has not seen the patient, has no access to their records, and can be confidently wrong. Read it as you would a suggestion from someone who was not on scene.