Formulating a Knee Pain Risk Metric Using an In-Shoe Motion Sensor Compatible With the Implicit Knowledge of Clinical Experts

Early identification of knee pain and mitigation of its risk are inevitable in one's daily life. The aim of this letter was to formulate an effective, long-term quantitative metric compatible with the implicit knowledge of clinical experts and suitable for daily knee pain risk assessment using an in-shoe motion sensor (IMS). We operationalized a knee pain risk metric derived from six distinct IMS-estimated knee joint parameters, utilizing principal component analysis. These knee joint parameters were determined via models established preliminarily. To ascertain validity, we first obtained clinical expert-rated knee pain risk scores (KPRSs) via gait observation, which indicate the risk of visiting primary care due to knee pain over the next five years, and then we examined the formulated knee risk metric outcomes against KPRSs from a sample population of 25 older, healthy adults using Pearson's correlation. The results yielded a large effect-size correlation (
r
= −0.816) between the constructed metric and the KPRS, thereby suggesting the validity of the constructed metric as well as the potential applicability in knee pain prevention using IMS.