AI-driven solutions for low back pain: A pilot study on diagnosis and treatment planning

Abstract

Low back pain (LBP) mainly affects the working-age population, and few specific causes can be identified, making diagnosis difficult and rendering them nonspecific. Artificial intelligence (AI) can be a great ally for prognosis, diagnosis, and treatment plans in healthcare. To describe the development of software aimed at providing prognoses, diagnoses, and treatment suggestions for LBP with AI support, as well as to report the functionality and initial limitations through a pilot study. Fifty assessment records from a database of patients at the Physiotherapy School Clinic of the University of Gurupi-UnirG, who were treated for LBP, were analyzed. Using data mining, including information described by patients and post-processing of discovered anamnesis patterns (rules), it was possible to develop software for evaluation and intervention in this patient group. Subsequently, a pilot study was initiated with 34 patients residing in the city of Gurupi-TO to test the application’s functionality. The software enabled more accurate treatments, diagnoses, and prognoses during the pilot study, directing the patient towards physiotherapeutic intervention based on the presented condition.