From 1 to 10, how much pain are you in right now?
When trying to help a patient, most doctors ask for their pain score and look at their facial expressions to have an idea of how much pain they are in. However, this isn't very accurate and could lead to unwanted consequences.
Given that people feel and show pain differently, a doctor’s estimation of the patient’s pain sometimes differ from the patient’s own pain score or even the "real" pain they are in. This creates confusion on whether prescribing high doses of potentially addictive painkillers is a must or not.
Hence, Dianbo Liu of the Massachusetts Institute of Technology and colleagues created a new algorithm to objectively measure one's pain levels just by looking at their facial expressions. This could help in catching out potential fakers, consequently leading to better diagnoses.
First, they trained the algorithm on videos of people experiencing pain. Each video contains a person who specifically has shoulder pain. Those people were asked to move at certain directions while rating their pain. The resulting algorithm appeared to be able to utilize just minute facial movements to guess how a person is truly currently feeling.
Specific parts of the face are particularly revealing, Liu says. Movements in nose and mouth, for instance, are particularly common in higher self-reported pain scores.
The team’s algorithm can also be modified to include a person’s sex, age, and skin complexion to have a more accurate result, with age being the biggest factor on their pain expression. Liu said that this creates an algorithm better at estimating pain levels than the one-size-fits-all system.
Although this is still in its infancy, Liu says the system will hopefully become an app for the doctors to use eventually. So, fakers, go ahead and practice your poker face or your award-winning acting now! Bring it on!
Get weekly science updates in your inbox!