Pain is experienced more intensely among certain groups of people. Studies have shown that individuals from minority and underprivileged groups tend to experience more physical pain than the general population due to stress, overall health, and other factors. However, misconceptions about how minority groups experience pain among healthcare providers have resulted in significant disparities in their quality of care. For instance, a 2016 study found that a subset of medical trainees believed that black people are not as sensitive to pain as white people and were less likely to address black people’s pain issues appropriately.
A team of Stanford researchers has turned to artificial intelligence to level the playing field when it comes to managing severe knee pain. The team, led by Jure Leskovec, has published a study demonstrating the deep learning system's capabilities in the journal Nature Medicine.
Standard clinical practices for measuring knee pain use a radiographic grading system developed 70 years ago called KLG. This technique does not consider other factors, such as stress and underlying medical conditions that could contribute to the worsening of the patient’s symptoms. The team’s new machine learning algorithm was designed to improve this, measuring and assessing knee X-rays more objectively and accurately than the current gold standard.
“By using X-rays exclusively, we show the pain is, in fact, in the knee, not somewhere else,” said Leskovec. “What’s more, X-rays contain these patterns loud and clear, but KLG cannot read them. We developed an AI-based solution that can learn to read these previously unknown patterns.”
Leskovec and colleagues validated the algorithm in a database of over 35,000 knee X-ray images from 4,000 patients, 20 percent of whom came from underserved and minority groups. The platform takes multiple parameters into account when predicting patient pain levels, such as health data, race, income, and body mass index. Ultimately, the model was able to scrub out some of the racial and socioeconomic biases that influence the formulation of pain scores.
“The pain is in the knee,” Leskovec explained. “Still useful as it is, KLG was developed in the 1950s using a not very diverse population and, consequently, it overlooks important knee pain indicators. This shows the importance to AI of using diverse and representative data.”
No, robots will not replace doctors, says Leskovec, who notes that he views the AI-driven system as complementing and supporting physicians’ decision-making process. By generating accurate pain scores and visual aids such as “heat maps” of painful areas in the knee, healthcare staff are better equipped to design the most appropriate treatment strategies for patients. Organizations like Shaip are helping AI companies improve their healthcare AI models with data licensing & accurately annotating healthcare datasets from various categories i.e. pathology, neurology, radiology etc.
“We think AI could become a powerful tool in the treatment of pain across all parts of society,” Leskovec said.