ChatGPT and DeepSeek-Assisted Rehabilitation in Subacromial Pain

NCT07148687 | Completed

• 1 other identifier: 2025/874
• Study Type: interventional
• Target: 75
• Locations: 1 country
• Sites: 1

Brief Summary

This study aims to evaluate the effects of rehabilitation programs supported by artificial intelligence-based large language models (LLMs), specifically ChatGPT and DeepSeek, in individuals with subacromial pain syndrome (SAPS). Participants will be randomly assigned to one of three groups: standard physiotherapy, standard physiotherapy plus ChatGPT-guided exercises, and standard physiotherapy plus DeepSeek-guided exercises. Outcomes including pain, disability, quality of life, muscle strength, and joint range of motion will be assessed before and after a 6-week intervention. The results will help determine whether AI-generated exercise recommendations can enhance clinical outcomes when combined with conventional physiotherapy.

Conditions

Subacromial Pain Syndrome

Keywords

ChatGPT; DeepSeek; Large Language Models; Shoulder Rehabilitation; Subacromial Pain; Artificial Intelligence

Outcome Measures

Primary Outcomes (1)

• Change in Disability (Shoulder Pain and Disability Index, SPADI)

  Disability related to shoulder pain will be measured.Shoulder Pain and Disability Index (0-100; higher scores indicate worse disability).

  Baseline and after 6 weeks of intervention

Secondary Outcomes (10)

• Change in upper extremity function (Quick DASH)

  Baseline and after 6 weeks

• Change in shoulder range of motion (ROM)

  Baseline and after 6 weeks

• Change in Quality of Life (Short Form-12, SF-12)

  Baseline and after 6 weeks of intervention

• Change in Pain Intensity (Numeric Pain Rating Scale, NPRS)

  Baseline and after 6 weeks of intervention

• Change in shoulder and elbow muscle strength (K-Force)

  Baseline and after 6 weeks

Study Arms (3)

Standard Exercise Group

ACTIVE COMPARATOR

Participants receive standard physiotherapy combined with a conventional exercise program for subacromial pain syndrome based on evidence-based clinical guidelines. Each group included 23 participants who completed the 6-week intervention and were included in the final analysis.

Behavioral: Conventional Exercise Program

ChatGPT-Guided Exercise Group

EXPERIMENTAL

Participants receive standard physiotherapy combined with exercises generated by the ChatGPT-4o large language model. Each group included 23 participants who completed the 6-week intervention and were included in the final analysis.

Behavioral: ChatGPT-Guided Rehabilitation

DeepSeek-Guided Exercise Group

EXPERIMENTAL

Participants receive standard physiotherapy combined with exercises generated by the DeepSeek-R1 large language model. Each group included 23 participants who completed the 6-week intervention and were included in the final analysis.

Behavioral: DeepSeek-Guided Rehabilitation

Interventions

ChatGPT-Guided Rehabilitation BEHAVIORAL

AI-generated exercise recommendations from ChatGPT-4 tailored to the patient's profile and applied under physiotherapist supervision.

ChatGPT-Guided Exercise Group

DeepSeek-Guided Rehabilitation BEHAVIORAL

Personalized rehabilitation exercises generated by DeepSeek-R1 and implemented under physiotherapist guidance.

DeepSeek-Guided Exercise Group

Conventional Exercise Program BEHAVIORAL

Exercise protocol targeting subacromial pain syndrome, including shoulder mobility, strengthening, and posture correction exercises based on established clinical guidelines.

Standard Exercise Group

Eligibility Criteria

• Age: 18 Years - 65 Years
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• Being between the ages of 18 and 65
• Having been diagnosed with Subacromial Pain Syndrome (SAPS) for at least four weeks, diagnosed through clinical and physical examination
• Actively maintaining shoulder range of motion
• Being cognitively and physically capable of participating in an exercise program
• Agreeing to participate voluntarily in the research process and providing written consent

You may not qualify if:

• Presence of other shoulder pathologies such as rotator cuff tear, frozen shoulder, glenohumeral instability, or labrum tear
• Presence of conditions included in the differential diagnosis such as cervical radiculopathy, neurological disorders, rheumatological diseases, or fibromyalgia
• History of shoulder surgery within the last 6 months
• Pregnancy
• History of systemic or neuromuscular disease
• Refusal to consent to voluntary participation

Sponsors & Collaborators

• Kirsehir Ahi Evran Universitesi: lead

Study Sites (1)

Kırşehir Ahi Evran University Training and Research Hospital, Physical Medicine and Rehabilitation Department

Kırşehir, Merkez, 40100, Turkey (Türkiye)

Location

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: TRIPLE
• Who Masked: PARTICIPANT, INVESTIGATOR, OUTCOMES ASSESSOR
• Masking Details: Participants were informed that they would receive one of several supervised exercise-based rehabilitation programmes for subacromial pain syndrome, but were not informed of their specific group assignment or the use of large language models. The physician establishing the diagnosis and the physiotherapist performing all outcome assessments are blinded to group allocation. Treating physiotherapists are aware of group assignment but are not involved in outcome assessments or data analysis.
• Purpose: TREATMENT
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: PhD

Model Details: Participants will be randomly assigned to three parallel groups: 1. standard physiotherapy combined with a conventional exercise program for subacromial pain syndrome based on current clinical guidelines, 2. standard physiotherapy combined with ChatGPT-guided exercises, and 3. standard physiotherapy combined with DeepSeek-guided exercises.

Study Record Dates

• First Submitted: August 3, 2025
• First Posted: August 29, 2025
• Study Start: September 25, 2025
• Primary Completion: December 1, 2025
• Study Completion: December 20, 2025
• Last Updated: February 27, 2026

Record last verified: 2026-02

Data Sharing

• IPD Sharing: Will not share

Locations

TU (1)