‘Robots capable of expressing pain can reduce doctors’ error and bias’: Report

A team of UK researchers has developed a method for creating more realistic and accurate expressions of pain when medical training robots are examining painful areas. The new approach by the Imperial College London team may reduce errors and bias during physical examinations by doctors.

In addition, the findings in Scientific Reports suggest this could teach trainee doctors to use clues hidden in facial expressions to minimize the force needed during physical examinations, and might also help detect and correct bias in students by exposing them to a wider variety of patient identities.

Sibylle Rerolle, from Imperial’s Dyson School of Design Engineering, says that improving the accuracy of facial expressions of pain on these robots is an important step forward in improving the quality of physical examination training for students. Students were asked to examine a robotic patient’s abdomen as part of the study. MorphFace – a robotic face capable of emulating pain-related facial expressions – was programmed using data about the force applied to the abdomen.

These experiments were conducted in order to find out what order the activation units (AUs) of a robot face must trigger in order to convey pain accurately.   A study was also conducted to determine the optimal speed and magnitude of AU activation. Researchers observed a more realistic facial expression when the upper face AUs (around the eyes) were activated first, followed by the lower face AUs (around the mouth). A longer delay in activating the Jaw Drop AU produced the most natural results.

In physical examinations of painful areas, doctors pay attention to patient facial expressions. Most medical training simulators, however, do not display real-time facial expressions of pain in addition to having a limited number of ethnicities and genders for the patients. Studies have already found a racial bias in the recognition of facial expressions of pain based on these limitations, according to the researchers.

According to co-author Thilina Lalitharatne from the Dyson School of Design Engineering, underlying biases may cause doctors to misinterpret the discomfort of patients, leading to mistreatment, damaging doctor-patient relations, and even causing death. Medical students could be trained to normalize their perceptions of pain by using a robot-assisted approach in the future.