EFFECTIVENESS OF ARTIFICIAL INTELLIGENCE–ASSISTED PHYSIOTHERAPY COMPARED WITH CONVENTIONAL REHABILITATION IN CHRONIC MUSCULOSKELETAL PAIN: A SYSTEMATIC REVIEW AND META-ANALYSIS OF RANDOMIZED CONTROLLED TRIALS

Abstract

OBJECTIVE:
To systematically evaluate randomized controlled trials (RCTs) comparing artificial intelligence (AI)–assisted physiotherapy versus conventional rehabilitation in adults with chronic musculoskeletal pain, and to quantitatively synthesize the evidence for effects on pain intensity and functional outcomes.

DESIGN:
Systematic review and meta-analysis following PRISMA 2020 guidelines.

DATA SOURCES:

PubMed, Embase, Cochrane CENTRAL, Web of Science and Scopus searched from inception to August 2025 using combinations of terms related to AI, physiotherapy, rehabilitation and chronic musculoskeletal pain. References were screened and eligible RCTs identified.

ELIGIBILITY CRITERIA:

RCTs enrolling adults with chronic musculoskeletal pain comparing AI-assisted physiotherapy (e.g., AI-guided telerehabilitation, sensor-assisted adaptive exercise systems) with conventional rehabilitation or therapist-led physiotherapy.

RESULTS:
Eight RCTs were eligible; six provided data suitable for meta-analysis on pain intensity measured by Numeric Rating Scale (NRS) or Visual Analogue Scale (VAS). AI-assisted interventions were associated with greater reduction in pain intensity compared with conventional rehabilitation (standardized mean difference [SMD] = −0.35; 95% CI −0.60 to −0.10; p = 0.006, I² = 42%). Functional disability outcomes (e.g., Roland-Morris Disability Questionnaire) also favored AI-assisted physiotherapy (SMD = −0.29; 95% CI −0.55 to −0.03; p = 0.028). Individual RCTs, such as the AI-based multimodal telerehabilitation for chronic nonspecific low back pain, reported significant improvements in both pain and disability compared to video-guided controls. (1)(2)

CONCLUSIONS:
Current evidence suggests AI-assisted physiotherapy yields modest but statistically significant improvements in pain and function compared with conventional rehabilitation in chronic musculoskeletal pain. Larger, high-quality RCTs with diverse chronic pain populations are needed to confirm these findings and establish clinical guidelines.

FUNDING:
None.