The Importance of Real-World Validation of Machine Learning Systems in Wearable Exercise Biofeedback Platforms: A Case Study.
biofeedback
biomedical technology
exercise therapy
human factors
inertial measurement unit
machine learning
wearables
Journal
Sensors (Basel, Switzerland)
ISSN: 1424-8220
Titre abrégé: Sensors (Basel)
Pays: Switzerland
ID NLM: 101204366
Informations de publication
Date de publication:
27 Mar 2021
27 Mar 2021
Historique:
received:
26
02
2021
revised:
22
03
2021
accepted:
25
03
2021
entrez:
3
4
2021
pubmed:
4
4
2021
medline:
28
4
2021
Statut:
epublish
Résumé
Machine learning models are being utilized to provide wearable sensor-based exercise biofeedback to patients undertaking physical therapy. However, most systems are validated at a technical level using lab-based cross validation approaches. These results do not necessarily reflect the performance levels that patients and clinicians can expect in the real-world environment. This study aimed to conduct a thorough evaluation of an example wearable exercise biofeedback system from laboratory testing through to clinical validation in the target setting, illustrating the importance of context when validating such systems. Each of the various components of the system were evaluated independently, and then in combination as the system is designed to be deployed. The results show a reduction in overall system accuracy between lab-based cross validation (>94%), testing on healthy participants (
Identifiants
pubmed: 33801763
pii: s21072346
doi: 10.3390/s21072346
pmc: PMC8037109
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : H2020 Marie Skłodowska-Curie Actions
ID : 676201
Organisme : Science Foundation Ireland
ID : 12/RC/2289_P2
Pays : Ireland
Références
Annu Int Conf IEEE Eng Med Biol Soc. 2016 Aug;2016:4686-4689
pubmed: 28269318
Sports Biomech. 2017 Sep;16(3):342-360
pubmed: 28523981
Stud Health Technol Inform. 2009;145:231-48
pubmed: 19592797
Musculoskelet Sci Pract. 2019 Feb;39:164-169
pubmed: 30502096
IEEE Trans Neural Syst Rehabil Eng. 2014 Jan;22(1):168-80
pubmed: 23661321
Annu Int Conf IEEE Eng Med Biol Soc. 2012;2012:2214-8
pubmed: 23366363
J Bone Joint Surg Am. 2015 Sep 2;97(17):1386-97
pubmed: 26333733
J Biomech. 2018 Nov 16;81:1-11
pubmed: 30279002
Phys Ther. 2019 Apr 1;99(4):478-486
pubmed: 30657981
JMIR Rehabil Assist Technol. 2017 Aug 21;4(2):e9
pubmed: 28827210
J Rehabil Assist Technol Eng. 2018 Mar 09;5:2055668318761523
pubmed: 31191926
J Neuroeng Rehabil. 2014 Nov 27;11:158
pubmed: 25431092
Sci Rep. 2018 Jul 26;8(1):11299
pubmed: 30050087
JMIR Rehabil Assist Technol. 2017 Apr 05;4(1):e4
pubmed: 28582253
Physiol Meas. 2009 Apr;30(4):R1-33
pubmed: 19342767
J Strength Cond Res. 2017 Aug;31(8):2303-2312
pubmed: 28731981
Annu Int Conf IEEE Eng Med Biol Soc. 2016 Aug;2016:659-662
pubmed: 28268414
Sports Med. 2018 May;48(5):1221-1246
pubmed: 29476427
Sensors (Basel). 2019 Jan 21;19(2):
pubmed: 30669657
Sensors (Basel). 2015 Feb 12;15(2):4193-211
pubmed: 25686308
Healthc Technol Lett. 2015 Aug 03;2(4):79-88
pubmed: 26609411
NPJ Digit Med. 2019 Jan 7;2:1
pubmed: 31304351
Disabil Rehabil. 2009;31(6):427-47
pubmed: 18720118
J Neuroeng Rehabil. 2013 Jun 18;10:60
pubmed: 23777436