New in EM 001: What does a patient’s smile tell you about their chance of having a PE?

Originally Published on RCEM Learning Podcast April 2017

As part of the RCEM Learning Podcast I record reviews of recent literature with Dave McCreary. We’ve been doing this for about a year now and you can hear them all on the RCEM Learning Podcast each month. I’d like to have them here and searchable on this site too so I’ll be posting the ones I find most relevant here. You can hear the newest ones by subscribing to the free RCEM Learning Podcast. 

Audio:

Clinical Question:

• What does the patient’s smile tell you about their chance of having a PE?

Title of Paper:

• Role of physician perception of patient smile on pretest probability assessment for acute pulmonary embolism

Journal and Year:

• EMJ, 2017

Lead Author:

• Jeff Kline

Overview of study methods:

• Patients Studied:
  • anyone undergoing CTPA in two EDs in the states ask to take part.
  • the doc looking after them had to do a structured questionnaire pre CT including a “gestalt” pre test probability and a direct question “did the patient smile”
  • the patient sat in front of a laptop and viewed a series of images and videos while the web camera on it recorded their responses to the images. (of note they have a control video of a cat getting angry and a dog falling in the pool from America’s funniest home movies as a way to make sure the patient did smile at some point)
  • A software programme then “reads” the facial expressions
  • They wanted to know whether the perception of smile affected the pre test probability of disease in either direction

Summary of Results:

• 208 patients, 13% rule in rate
• the doc was more likely to recall a smile the patients who ruled in for PE (63 v40%)
• the software agreed with this (PE positive patients were more likely to demonstrate a happy affect while watching funny cat videos…)

Thoughts

• this is the second of Kline’s papers looking at this, the first looked at chest pain overall using similar facial recognition software. it found somewhat different results in that patients with less emotional response were more likely to have pathology (though it covered a large range of pathologies not just PE)

Clinical Bottom Line:

• that “gut” feeling you have about patients likely has some validity and emotional response and its perception is part of that. But we’re not yet at a place we can use it clinically.

Other #FOAMed Resources:

• Jeff is on Twitter and a must follow
• Go read the first paper too
• And if you haven’t go read Kahneman too and listen to Gary Klein on EMCrit for the alternative.