Identifying changes in dynamic plantar pressure associated with radiological knee osteoarthritis based on machine learning and wearable devices.
Gege Li, Shilin Li, Junan Xie and 7 others
PMID 38570841WHAT IT FOUND
Wearable insoles and machine learning correctly distinguished between people with radiological knee osteoarthritis and those without in 82.61% of cases.
The key differences were lower peak pressures in the big toe and heel, and greater variability in foot balance during walking.
Key findings
01A random forest model using age and three specific plantar pressure features distinguished between radiological knee osteoarthritis and non-radiological cases with 82.61% accuracy on the testing set.
02People with radiological knee osteoarthritis showed lower absolute peak plantar pressures in the big toe and heel compared to those without.
03Increased variability in the center of pressure and standard deviation of peak pressures were the most prominent features associated with radiological knee osteoarthritis, suggesting decreased gait stability.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study did not externally validate the model on a separate population, raising the risk of overfitting. The ROA group was significantly older than the non-ROA group, so it is unclear if the plantar pressure differences are due to the knee condition itself or age-related changes. The study only classified participants into two broad groups (ROA vs non-ROA) rather than four Kellgren–Lawrence grades, limiting insight into disease progression. No subgroup analysis was conducted based on the side of the affected knee.
Declared interests
The authors declare no competing interests. The study was supported by grants from the National Natural Science Foundation of China and other regional bodies.
The easy way to misread this
Do not assume the machine learning model is ready for clinical diagnosis or that the pressure changes are solely caused by knee osteoarthritis. The model was not externally validated, and the ROA group was significantly older than the control group, meaning age could be driving the observed gait differences.