Evaluation of an Algorithmic Approach to Pediatric Back Pain

Abstract

Pediatric patients require a systematic approach to treating back pain that minimizes the number of diagnostic studies without missing specific diagnoses. This study reviews an algorithm for the evaluation of pediatric back pain and assesses critical factors in the history and physical examination that are predictive of specific diagnoses. Eighty-seven pediatric patients with thoracic and/or lumbar back pain were treated utilizing after this algorithm. If initial plain radiographs were positive, patients were considered to have a specific diagnosis. If negative, patients with constant pain, night pain, radicular pain, and/or an abnormal neurological examination obtained a follow-up magnetic resonance imaging. Patients with negative radiographs and intermittent pain were diagnosed with nonspecific back pain. Twenty-one (24%) of 87 patients had positive radiographs and were treated for their specific diagnoses. Nineteen (29%) of 66 patients with negative radiographs had constant pain, night pain, radicular pain, and/or an abnormal neurological examination. Ten of these 19 patients had a specific diagnosis determined by magnetic resonance imaging. Therefore, 31 (36%) of 87 patients had a specific diagnosis. Back pain of other 56 patients was of a nonspecific nature. No specific diagnoses were missed at latest follow-up. Specificity for determining a specific diagnosis was very high for radicular pain (100%), abnormal neurological examination (100%), and night pain (95%). Radicular pain and an abnormal neurological examination also had high positive predictive value (100%). Lumbar pain was the most sensitive (67%) and had the highest negative predictive value (75%). This algorithm seems to be an effective tool for diagnosing pediatric back pain, and this should help to reduce costs and patient/family anxiety and to avoid unnecessary radiation exposure.