Abstract | ||
---|---|---|
ABSTRACT The aging population problem is growing worldwide and physical therapists must be able to care for the elderly and injured efficiently. One burden facing physical therapists is determining the Functional Independence Measure (FIM), which measures the level of independence in activities of daily living for the elderly and injured. This measure is used in designing rehabilitation programs. Determining the FIM is a time-consuming task for both physical therapists and rehabilitation patients because it requires an interview. Some researchers have explored estimating FIM using wearable devices; however, it is burdensome for patients to wear such devices continuously. Therefore, we propose a method to estimate FIM motor items automatically using machine learning and sleep sensors. The proposed method classifies patients into three levels of FIM values that are referred to as s-FIM using vital data acquired through sleep sensors, personal data (age, gender, and Body Mass Index), and s-FIM values from the previous day. The results of a one-week study based on 19 patients showed that the proposed method had a mean accuracy of 0.87 for s-FIM motor items. The results were more accurate than continuing to use the s-FIM values determined by the physical therapist without updating them for one week. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1145/3460418.3479308 | Ubiquitous Computing |
Keywords | DocType | Citations |
Functional independence measure, Machine learning, Sleep data | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Eiji Kumakawa | 1 | 0 | 0.34 |
Wataru Yamada | 2 | 29 | 15.22 |
Keiichi Ochiai | 3 | 4 | 4.10 |
Yusuke Fukazawa | 4 | 137 | 19.28 |
Mizuki Shirai | 5 | 0 | 0.34 |
Hirokazu Masuda | 6 | 0 | 0.34 |