Artificial Intelligence-Based Physical Therapy Interventions for Non-Specific Low Back Pain: A Systematic Review and...

Background/Objectives: Non-specific low back pain (NSLBP) is the leading cause of disability worldwide. Artificial intelligence (AI) technologies are increasingly being integrated into healthcare interventions for NSLBP, yet their effectiveness remains uncertain. This systematic review and meta-analysis aimed to evaluate the effectiveness of AI-based Physical therapy (PT) interventions on pain intensity and disability outcomes in patients with NSLBP. Methods: We conducted a comprehensive search across six electronic databases. Randomised controlled trials (RCTs) evaluating AI-based interventions for NSLBP were only included. Mean differences (MD) with 95% confidence intervals (CIs) were calculated using random-effects models. Heterogeneity was assessed using I² statistics and Cochran's Q test. Results: Five RCTs (n = 1939) met the inclusion criteria for systematic review. Three RCTs (n = 594 participants) provided data for meta-analysis. AI-based interventions significantly reduced pain (pooled MD -0.721, 95% CI -1.047 to -0.395; z = -4.34, p < 0.001; I² = 9.5%). Disability also significantly improved (pooled MD -1.031, 95% CI -2.020 to -0.042; t(2) = -4.48, p = 0.046; I² = 0%). Neither effect reached the minimal clinically important difference (1.0 for pain, 2-4 for disability). No serious adverse events were reported. Conclusions: AI-based PT interventions produce statistically significant but clinically small improvements in pain and disability for NSLBP. Certainty of evidence is low due to risk of bias and imprecision. Larger, blinded RCTs with standardised outcomes are needed.