Background: Low back pain is a common cause of disability in the US with increasing financial burden on healthcare. A variety of treatment options exist to combat LBP. Home-based therapy is a low-cost option, but there is a lack of data on how it compares to therapy in clinical settings. It was hypothesized that when using artificial intelligence-guided therapy, supervised in-clinic interventions would have a greater influence on patient-reported outcomes and strength than unsupervised, home interventions.
Methods: This is a non-randomized controlled trial of 51 patients (28 female, 23 male). The investigation compared an 8-week, core-focused exercise intervention in a Clinic (supervised) versus Home (unsupervised) setting. Outcome variables included measures of strength, performance, and patient-reported outcomes related to function. Generalized linear regression (p < 0.05) was used to evaluate outcomes were evaluated with respect to sex, intervention setting, and time.
Findings: Male subjects exhibited greater strength (p ≤ 0.02) but not greater patient-reported outcomes (p ≥ 0.30) than females. The Clinic group exhibited slightly greater lateral pull-down strength (p = 0.002), greater eccentric phase range of motion during overhead press (p < 0.01), and shorter concentric phase duration during bench press (p < 0.01) than the Home group. Significance between groups was not observed in any other strength, performance, or patient-reported outcome (p ≥ 0.11).
Interpretation: A lack of consistent significance indicated that the hypothesis was not supported. AI-guided, telehealth exercise produced comparable outcomes in both home and clinical settings. Telehealth options may offer a lower-cost alternative to clinic-based exercise therapy for patients with nonspecific lower back pain.
Keywords: Artificial intelligence; Exercise is medicine; Low back pain; Moderate intensity exercise; Telemedicine.
Copyright © 2023 Elsevier Ltd. All rights reserved.