Improving Accelerometry-Based Measurement of Functional Use of the Upper Extremity After Stroke: Machine Learning Versus Counts Threshold Method.
Peter S Lum, Liqi Shu, Elaine M Bochniewicz and 4 others
PMID 33150830WHAT IT FOUND
Machine learning models accurately measured functional arm use in stroke survivors, closely matching video evidence.
The standard counts threshold method failed, grossly overestimating functional use by including nonfunctional movements like arm swing. This suggests current accelerometry metrics may be misleading for assessing real-world recovery.
Key findings
01Machine learning intrasubject models estimated functional use with high accuracy, showing a strong correlation with video ground truth.
02The standard counts threshold method poorly represented functional use, showing no significant correlation with video ground truth and high error rates.
03The counts threshold method detected more nonfunctional than functional movement in the paretic limb, leading to gross overestimation of functional use.
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 included a small sample size (n=20). The activity script was limited to four specific instrumental activities and may not represent a full day of real-world behavior, where more rest periods occur. The intrasubject machine learning models require extensive video annotation (6.9 person-hours per participant), limiting clinical feasibility. One stroke participant with high functional ability was a clear outlier for intersubject models, suggesting these models may struggle with patients who have near-normal movement patterns.
Declared interests
The authors declared no potential conflicts of interest. Funding was provided by the Department of Health and Human Services (NIDILRR), the US Army Medical Research and Material Command, the Department of Veterans Affairs, and The MITRE Corporation.
The easy way to misread this
Do not assume that standard 'usage' metrics from wrist accelerometers accurately reflect functional arm use in stroke patients. This study shows these metrics grossly overestimate functional use because they fail to distinguish purposeful movement from nonfunctional arm swing, especially in the paretic limb.