Wearable fall risk assessment by discriminating recessive weak foot individual.
Zhen Song, Jianlin Ou, Shibin Wu and 3 others
PMID 40114179WHAT IT FOUND
A new algorithm flags patients with inconsistent gait patterns that confuse fall-risk models.
By separating these individuals, the system improved fall-risk classification accuracy from 81.2% to 85.4%. This tool aids quality control but requires further testing before clinical use.
Key findings
01The two-stage model achieved an accuracy of 85.4% with a sensitivity of 87.5%, compared to the 81.2% baseline accuracy.
02The adaptive threshold method identified individuals with 'Recessive Weak Foot' gait patterns, which are characterized by discontinuous high-risk gait on the weak foot side.
03The method improved accuracy in Parkinson's Disease classification from 67.2% to approximately 71% by discriminating between individual types.
STILL TO COME
How it was doneWhat they foundWhat it means for PTs
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 is a technical validation of an algorithm, not a clinical trial. It did not measure patient outcomes or compare the tool's effectiveness against standard clinical assessment. The sample size was small (48 participants), and the 'Recessive Weak Foot' subgroup had too little data for the specific model to train effectively, resulting in low accuracy for that group. The method relies on specific plantar pressure data and is not yet a 'plug-and-play' solution for other types of wearable sensors or clinical settings. The definition of fall risk was based on the Berg Balance Scale, which is a clinical measure, but the study's primary focus was on model accuracy rather than clinical predictive validity for actual falls.
Declared interests
The study was supported by the National Key Research and Development Program of China. No conflicts of interest were declared.
The easy way to misread this
Do not assume this algorithm is ready for clinical use. The study validates a technical improvement in data classification, not a proven reduction in patient falls. Furthermore, the model performed poorly for the specific subgroup of patients with inconsistent gait (66.7% accuracy), meaning it may still misclassify the very patients it was designed to identify.