Walking stability in patients with benign paroxysmal positional vertigo: an objective assessment using wearable accelerometers and machine learning.
Yuqian Zhang, He Wang, Yifei Yao and 3 others
PMID 33789693WHAT IT FOUND
People with benign paroxysmal positional vertigo walked slower, took shorter steps, and had less smooth, more variable gait than healthy controls.
Wearable sensors captured these walking changes in a clinic corridor.
Key findings
01BPPV patients walked more slowly, with shorter steps and lower cadence, than healthy controls.
02BPPV patients had less regular and more variable walking, including lower step regularity in head motion and higher gait variability in head and trunk motion.
03A machine-learning model using gait variables reported area-under-curve values of 0.78 for separating BPPV patients from controls and 0.83, 0.85, and 0.96 for mild, moderate, and severe dizziness-handicap groups.
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 included only 27 BPPV patients and 27 healthy controls, and the DHI subgroups were small (12 mild, 9 moderate, 6 severe). Only posterior canal BPPV was studied, so the findings may not apply to other canal types. The machine-learning model was tested only with repeated 5-fold cross-validation on the same dataset, not external validation or a prospective clinical test. The model used 4 principal components from gait variables, so it is not clear which specific walking features drive the classification.
Declared interests
Funding was provided by the National Nature Science Foundation of China and the Key Technology Support Project of Shanghai Municipal Science and Technology Commission. The supplied text does not state any competing interests beyond these funders.
The easy way to misread this
Do not read this as a validated wearable screening test for BPPV. The model was built on 27 patients and 27 controls and tested only with repeated 5-fold cross-validation, and the authors describe it as a preliminary attempt.