Artificial Intelligence-guided Resistance Training Maintained Improved Patient-reported Outcomes In Subjects With Low Back Pain

Lower back pain (LBP) is a prevalent issue, with diverse intervention options targeted at alleviating chronic LBP. Previous studies have demonstrated artificial intelligence (AI)-guided resistance training increased functional outcomes and patient satisfaction in individuals with LBP after an eight-week AI-guided core resistance training intervention. PURPOSE: Determine the retention of improvement in LBP patient-reported outcomes (PROs) after an AI-guided rehabilitation a year after intervention. METHODS: Accrued subjects were randomized into five groups (CTRL, TRAIN, CLIN, COMBO, HOME) and completed an 8 week training intervention. The CTRL group (N = 9) received no treatment, TRAIN (N = 13) received supervised AI-guided core-focused resistance training in the clinic, CLIN (N = 27) received standard clinical care, COMBO (N = 20) received both clinical care and AI-guided training, and HOME (N = 20) received AI-guided training in their homes. Subjects completed patient reported outcomes measurement information system (PROMIS) pain interference and physical function and Tampa Scale of Kinesiophobia (TSK) questionnaires at baseline and again at one year after the training intervention period. Individual subject differences between baseline and follow-up scores were calculated. Differences in scores between timepoints per group were determined via Wilcoxon Rank Tests (p < 0.05). RESULTS: There were no significant differences between baseline and follow-up for any group on the TSK questionnaire (p ≥ 0.37). HOME reported significantly lower median PROMIS pain interference score at follow-up compared to baseline (48 vs. 57, p = 0.01) and higher PROMIS physical function score at follow-up compared to baseline (53 vs. 47, p = 0.00). CONCLUSION: These results suggest there was retention in decreased pain and increased physical function measured by PROs a year after the training intervention for subjects with access to continue AI-guided training at home. Although improvements were seen at 8 weeks in the COMBO and TRAIN groups, there were no differences in PROs compared to baseline one year after the intervention. AI-guided interventions, particularly when accessible from home, may offer a viable supplement to standard clinical care for LBP and to encourage continued exercise and training to alleviate LBP.