EFFECT OF VIRTUAL REALITY–BASED ARTIFICIAL INTELLIGENCE REHABILITATION ON PAIN AND FUNCTIONAL OUTCOMES IN PATIENTS WITH CHRONIC MUSCULOSKELETAL PAIN: A RANDOMIZED CONTROLLED TRIAL

Abstract

Background: Chronic musculoskeletal pain (CMP) is one of the major causes of disability across the globe and has complicated biopsychosocial treatment issues. VR AI through immersion in an artificial virtual reality and artificial intelligence (AI) instructions (herein VR AI) would potentially support engagement, customization of rehabilitation and pain processing via multisensory immersion, motor retraining, and cognitive mechanisms. Nevertheless, the presence of high-quality randomized evidence addressing pain and functional outcomes with a high quality of effect estimates is relatively low.

Objective: The study aims to find out whether a 12 week, home based VR AI rehabilitation intervention plus usual care is better at reducing the intensity of pain and enhancing the function of adults with CMP than usual care alone.

Methodology:

This study was a single-blind (outcome assessor), parallel-group randomized controlled trial conducted to evaluate the effectiveness of a virtual reality–based artificial intelligence (VR-AI) rehabilitation program on pain and functional outcomes in individuals with chronic musculoskeletal pain. Participants were eligible if they were aged between 18 and 75 years, had chronic musculoskeletal pain lasting ≥3 months (e.g., low back pain, neck pain, shoulder or knee pain), reported an average pain intensity ≥4 on a 0–10 Visual Analog Scale (VAS), were able to understand and follow instructions and had access to stable internet connectivity at home. Those with progressive neurological disorders or inflammatory rheumatological conditions, recent musculoskeletal surgery (<6 months), uncontrolled cardiovascular, vestibular, or severe psychiatric conditions, epilepsy or severe motion sickness contraindicating VR use and prior regular exposure to VR-based rehabilitation within the past year were excluded from the study.

Intervention:

VR AI group was treated to a structured AI adaptive VR rehabilitation (45min sessions, 3 times/week, 12 weeks) comprising of graded movement activities, biofeedback, computer-generated distraction modules, and automated difficulty adjustment through reinforcement learning. Control group Underwent conventional care (conventional physiotherapy recommendations and exercise program) equivalent in terms of contact frequency. Main outcome measures: The main outcome measure was pain intensity in Visual Analogue Scale in 12 weeks. Secondary outcomes were that of functional disability (Oswestry Disability Index or regional equivalent) pain catastrophizing, Kinesio phobia, and health related quality of life.

Results:

 VR AI group had lower mean pain VAS (mean 3.53, SD 1.44) at 12 weeks as compared to controls (mean 4.77, SD 1.64). Mean difference (VR AI - Control): = -1.24 (95% CI -1.80 -0.68), p=0.001; Cohen d=-0.80 (large effect). Functional disability showed better improvement in the VR AI group with the mean difference of -6.8 (95% CI -9.6 -4.0), p = 0.001; Cohen d = -0.72. The VR AI had clinically significant and statistically significant improvements in secondary outcomes (catastrophizing, Kinesio phobia, and EQ 5D index). The rate of adherence (median 86) was also high and adverse events were mild and rare.

Conclusions:

This RCT demonstrated that a 12 week home based VR AI rehabilitation intervention had significant effects of reducing pain intensity and increasing functional capacity in adults with CMP compared to usual care with moderate to large effect sizes. VR AI is a potentially promising adjunct to standard rehabilitation and has to undergo additional translational research, cost effectiveness and longitudinal follow up.