Effects Of Artificial Intelligence-guided Resistance Training On 52-week Clinical Visits In Low Back Pain Subjects

Lower back pain (LBP) is a common cause of disability in US adults and causes significant impact on healthcare costs. A variety of treatment plans are used to target chronic LBP. PURPOSE: Determine changes in clinical treatment one-year after intervention for LBP. METHODS: 67 subjects randomized into four Groups (CTRL, TRAIN, CLIN, COMBO) and completed an 8-week intervention. CTRL = no treatment, TRAIN = supervised AI-guided core-focused resistance training, CLIN = self-selected clinical care, and COMBO = clinical care + AI-guided training. Post-intervention, all subjects pursued clinical care at their discretion. Clinical visits were tracked for 1-year post-enrollment. Generalized quantile regression (Χ2 < 0.05) with factors of Time (week 0, 16, 26, 39, & 52) and Group were used to assess significance. Survival Analysis assessed when subjects increased their clinical care visits normalized to the Group average at week 0 (Χ2 < 0.05). RESULTS:CLIN & COMBO groups received more clinical care visits than the CTRL & TRAIN groups at weeks 0 & 16 (Χ2 < 0.01). No Group differences were observed from week 26 - week 52. Mean clinical visits for CLIN decreased from week 0 (3.0) to week 16 (1.0); and for COMBO from week 0 (4.0), to week 16 (2.0), to week 26 (0.0). All other Group*Time interactions exhibited mean clinical visits = 0.0. Least squares mean estimate of clinically-prescribed back restrictions at week 0 was CLIN = 8.0 and COMBO = 8.0. These regressed to 0.0 by week 16 for CLIN and by week 26 for COMBO. There were no significant between group differences for increased clinical visits (Χ2 = 0.65). 52-week survival relative to increased clinical visits was CTRL = 77%, TRAIN = 75%, CLIN = 62%, COMBO = 60%. All increased visit frequency occurred by week 26. CONCLUSIONS: For clinical LBP patients, these results indicate decrease in clinical treatment schedules between 16 and 26 weeks. Incorporation of AI-guided resistance training neither aided nor impeded progress from clinical treatment. Previously, we have shown significant improvement in biomechanical and clinical variable performance by 8 weeks in the TRAIN, CLIN, and COMBO groups. These improvements may contribute the overall decrease in clinical visits.