Comparison of an Artificial Intelligence-Assisted Rehabilitation Program for Shoulder Musculoskeletal Disorders and the Clinical Decision Making of Therapists

People with shoulder musculoskeletal disorders among middle-aged and older adults have the highest need of rehabilitation services. The population growth and aging society subsequently increase the number of disabled people, the healthcare costs and the needs for healthcare professionals. The evidence exists to support the beneficial effect of exercises on function and quality of life. Traditionally, a rehabilitation program is designed by therapists for each patient depending on their conditions. In recent years, AI is increasingly being employed in the field of physical and rehabilitation medicine, however, there is no study of applying AI in predicting rehabilitation programs for shoulder musculoskeletal disorders. The main purpose of this study is to explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders. Twenty-three features are identified based on shoulder range of motion, pain, whether or not perform surgical procedure. Each exercise is considered as a label with a total of twenty-five exercises. Dataset is collected by clinical therapists to develop and train the model. Each patient has to receive at least two months of rehabilitation and two times of evaluation. Logistic regression, support vector machine and random forest are used to build the computational model. Accuracy, precision, recall, F-1 score and AUC are used to evaluate the performance of the computational model in machine learning. After training, we compare the consistency of rehabilitation programs predicted by using machine learning model and the clinical decision making of therapists.

Study Overview

• Status: Recruiting
• Conditions: Machine Learning; Shoulder Musculoskeletal Disorders; Rehabilitation Programs
• Intervention / Treatment: Other: usual care

• Study Type: Observational
• Enrollment (Anticipated): 80

Contacts and Locations

• Study Contact: Name: Hanyun Hsiao, master; Phone Number: 1624 +88622490088; Email: 10252@s.tmu.edu.tw

Study Locations

• Taiwan
  • New Taipei City, Taiwan, 235
    • Recruiting
    • Shuang Ho Hospital
    • Contact: Hanyun Hsiao, master; Phone Number: 1624 +88622490088; Email: 10252@s.tmu.edu.tw

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

Musculoskeletal disorders that commonly cause shoulder pain in the clinic include Adhesive Capsulitis of shoulder (AC or frozen shoulder), Rotator Cuff Tear or Rupture (RCT), and Shoulder Impingement Syndrome (SIS). The incidence of AC in the general population is approximately 2-5%, most commonly occurring in women aged 40-60 years; the incidence of RCT is 20.7% and increases with age, most commonly associated with SIS is the most frequent cause of shoulder pain, accounting for about 44-65% of cases, usually affecting people over the age of 40.

Description

Inclusion Criteria:

1. The International Classification of Diseases, 10th revision (ICD-10) codes were selected before the study started and included the ICD-10 codes M75 (Shoulder lesions), S42 (Fracture of shoulder and upper arm), S43 (Dislocation and sprain of joints and ligaments of shoulder girdle), and S46 (Injury of muscle, fascia and tendon at shoulder and upper arm level)
2. Patients who need rehabilitation after undergoing surgical procedure and are able to perform stretch, active assistive range of motion (AAROM) or supervised active range of motion (AROM)
3. between 20-80 years old
4. Are able to follow motor commands

Exclusion Criteria:

1. Patients with central and peripheral nervous system disease, such as cerebrovascular accident (CVA), Parkinson's disease (PD), myasthenia gravis (MG), poliomyelitis
2. Patients who had shoulder contusion, vascular injury, severe crush injury and amputation

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
shoulder musculoskeletal group; The International Classification of Diseases, 10th revision (ICD-10) codes were selected before the study started and included the ICD-10 codes M75 (Shoulder lesions), S42 (Fracture of shoulder and upper arm), S43 (Dislocation and sprain of joints and ligaments of shoulder girdle), and S46 (Injury of muscle, fascia and tendon at shoulder and upper arm level)	Other: usual care; usual care(rehabilitation program)

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Accuracy	To explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders	Change from Baseline at 2 months
Precision	To explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders	Change from Baseline at 2 months
Recall	To explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders	Change from Baseline at 2 months
F-1 score	To explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders	Change from Baseline at 2 months
AUC	To explore the possibilities of using supervised machine learning approach to predict rehabilitation programs for shoulder musculoskeletal disorders	Change from Baseline at 2 months

Collaborators and Investigators

• Sponsor: Taipei Medical University Shuang Ho Hospital

Publications and helpful links

General Publications

• Burns DM, Leung N, Hardisty M, Whyne CM, Henry P, McLachlin S. Shoulder physiotherapy exercise recognition: machine learning the inertial signals from a smartwatch. Physiol Meas. 2018 Jul 23;39(7):075007. doi: 10.1088/1361-6579/aacfd9.
• Challoumas D, Biddle M, McLean M, Millar NL. Comparison of Treatments for Frozen Shoulder: A Systematic Review and Meta-analysis. JAMA Netw Open. 2020 Dec 1;3(12):e2029581. doi: 10.1001/jamanetworkopen.2020.29581.
• Linsell L, Dawson J, Zondervan K, Rose P, Randall T, Fitzpatrick R, Carr A. Prevalence and incidence of adults consulting for shoulder conditions in UK primary care; patterns of diagnosis and referral. Rheumatology (Oxford). 2006 Feb;45(2):215-21. doi: 10.1093/rheumatology/kei139. Epub 2005 Nov 1.
• Oude Nijeweme-d'Hollosy W, van Velsen L, Poel M, Groothuis-Oudshoorn CGM, Soer R, Hermens H. Evaluation of three machine learning models for self-referral decision support on low back pain in primary care. Int J Med Inform. 2018 Feb;110:31-41. doi: 10.1016/j.ijmedinf.2017.11.010. Epub 2017 Nov 23.
• Gupta R, Srivastava D, Sahu M, Tiwari S, Ambasta RK, Kumar P. Artificial intelligence to deep learning: machine intelligence approach for drug discovery. Mol Divers. 2021 Aug;25(3):1315-1360. doi: 10.1007/s11030-021-10217-3. Epub 2021 Apr 12.

Study record dates

Study Major Dates

• Study Start (Actual): July 11, 2022
• Primary Completion (Anticipated): April 30, 2024
• Study Completion (Anticipated): April 30, 2024

Study Registration Dates

• First Submitted: May 5, 2023
• First Submitted That Met QC Criteria: May 5, 2023
• First Posted (Actual): May 15, 2023

Study Record Updates

• Last Update Posted (Actual): May 15, 2023
• Last Update Submitted That Met QC Criteria: May 5, 2023
• Last Verified: June 1, 2022

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Musculoskeletal Diseases
• Other Study ID Numbers: N202206013

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No