Create a Bulleted List of Asthma Control

Created October 3, 2025

A

by Adam Abbas

Software Engineer at Evidently

Note from the author:

I have this prompt saved in a document and use it almost daily. It works best in Evidently, because Evidently is the best. I don’t recommend using it in generic LLMs that don’t have access to clinical data, as there is nothing to summarize.

You are a pediatric pulmonologist. create a bulleted summary of asthma control since the last clinic visit with me. This includes dates of steroid courses (prednisone, prednisolone, methylpred, and dexamethasone), emergency department visits for breathing issues, urgent care visits for breathing issues, hospitalizations for breathing issues. Also include what inhalers they are prescribed, most recent chest xray results, any chest CT results, any swallow study results, and any flexible bronchoscopy results. Finally, include a social history bullet that includes who is living at home, whether there are smoking or vaping exposures, what pets live in the house, whether they were born at full term, and whether they had any neonatal respiratory issues

Specialty: Pulmonary Disease

AI Tool: Evidently

Run this prompt on a synthetic patient:

Profile Icon

Jane Doe, 66 year-old female

See what the response to this prompt looks like in Evidently by running this prompt on Jane, one of our synthetic, de-identified patients.

Run this prompt on a synthetic patient

See what the response to this prompt looks like in Evidently by running this prompt on Jane, one of our synthetic, de-identified patients.

Profile Icon

Jane Doe, 66 year-old female

who is jane doe

Profile Icon

Jane Doe

66 YEAR OLD

Anytown, USA

Jane Doe is a synthetic patient from a publicly de-identified, PHI-compliant dataset. Jane has a very complex medical history, which provides a great sandbox for exploring how prompts can affect information retrieval in complex patient record data. Jane's patient record contains 1,500 scanned documents, 400 imaging documents, 400 data elements, and 18 faxes.

Evidently's prompt responses pair the patient record with an expansive medical knowledge graph consisting of millions of medical concepts and the billions of connections between them.