Scientists have now engineered wearable sensors that may predict how patients with knee osteoarthritis might improve with physical therapy.
Osteoarthritis is a type of joint disease where surrounding cartilage and bone breaks down, leaving patients with severe joint pain, stiffness, and reduced mobility. It is one of the most common musculoskeletal diseases in the world and will impact over half of all people at some point in their lives. Specifically, knee osteoarthritis outnumbers all other mobility-related disabilities in people over the age of 65. While there is currently no cure, the only non-invasive treatment method available involves intensive muscle strengthening exercises. However, not all patients respond positively to this treatment. Research published in the Journal of NeuroEngineering and Rehabilitation describes a new tool that could help doctors and clinicians predict how patients might respond to physical therapy.
The study was led by Dr. Dylan Kobsar, at the University of Calgary. His team looked to see how data collected from wearable sensors attached to a patient’s back, thigh, shank, or foot, predicted physiotherapy outcomes. Thirty-nine adults radiographically diagnosed with knee osteoarthritis were vetted for corticosteroid use, systemic arthritic conditions, and high body mass index before being admitted into the study. Patients then underwent what researchers call a “gait analysis”, where the participants performed basic movement exercises on a treadmill with the inertial sensors placed on their back, thigh, shank, and foot to detect electrical stimulation. To compare improvement before and after physiotherapy, researchers had the participants fill out a Knee injury and Osteoarthritis Outcome Score (KOOS) – a standardized self-administered questionnaire to quantify symptoms such as pain and their ability to function in daily life.
The patients then completed a six-week hip strengthening exercise that targeted their hip and core. The exercises were administered by a Board Certified Athletic therapist who also monitored patient pain and exercise frequency. Overall the patients performed the exercises approximately five days per week.
At the end of the study, researchers then collated the data, looking to see if any information collected from the inertial sensors might be correlated with patient health outcomes. The scientists found that the most accurate test required data from the back, thigh, and shank sensors, with 81.7% accuracy. However, a simplified two-sensor system – back and thigh only – was statistically just as good, with 80.0% accuracy. Out of all the sensor placements, the thigh was the most predictive of improvement, at 74.4% accuracy, while the back was the worst, with only 66.7% accuracy.
Kobsar’s research team thinks that a more thorough analysis of the data collected by the wearable sensors could reveal even more diagnostic information. Nonetheless, the scientists argue that the use of the simplified two-sensor system they’ve devised could be a critical tool for clinicians and patients alike. By allowing doctors to predict how patients might respond to physiotherapy, patients can expect better care, enhanced treatments, and faster health improvements.
Written by Calvin J. Chan B. Sc.
Kobsar, D., Osis, S.T., Boyd, J.E., Hettinga, B.A. and Ferber, R. (2017). Wearable sensors to predict improvement following an exercise intervention in patients with knee osteoarthritis. Journal of NeuroEngineering and Rehabilitation. 14:94.