Predicting Likelihood of Surgery Before First Visit in Patients With Back and Lower Extremity Symptoms: A Simple Mathematical Model Based on More Than 8, 000 Patients. Issue 18 (15th September 2018)

• Record Type: Journal Article
• Authors: Boden, Lauren M.; Boden, Stephanie A.; Premkumar, Ajay; Gottschalk, Michael B.; Boden, Scott D.
• Is Part Of: Spine. Volume 43:Issue 18(2018)
• Publication Date: 2018-09-15
• Subjects: clinical appropriateness -- clinical triage -- low back pain -- multiple regression -- patient allocation -- predictive model -- radicular lower extremity symptoms -- sciatica -- scoring system -- spine center -- spine surgery; Spine -- Abnormalities -- Periodicals; Spine -- Diseases -- Periodicals; Spine -- Surgery -- Periodicals; 616.73005
• DOI: 10.1097/BRS.0000000000002603
• Languages: English
• ISSNs: 0362-2436

Abstract

Abstract : Study Design: Retrospective analysis of prospectively collected data. Objective: To create a data-driven triage system stratifying patients by likelihood of undergoing spinal surgery within 1 year of presentation. Summary of Background Data: Low back pain (LBP) and radicular lower extremity (LE) symptoms are common musculoskeletal problems. There is currently no standard data-derived triage process based on information that can be obtained before the initial physician-patient encounter to direct patients to the optimal physician type. Methods: We analyzed patient-reported data from 8006 patients with a chief complaint of low back pain and/or LE radicular symptoms who presented to surgeons at a large multidisciplinary spine center between September 1, 2005 and June 30, 2016. Univariate and multivariate analysis identified independent risk factors for undergoing spinal surgery within 1 year of initial visit. A model incorporating these risk factors was created using a random sample of 80% of the total patients in our cohort, and validated on the remaining 20%. Results: The baseline 1-year surgery rate within our cohort was 39% for all patients and 42% for patients with LE symptoms. Those identified as high likelihood by the center's existing triage process had a surgery rate of 45%. The new triage scoring system proposed in this study was able to identify a high likelihood group in which 58% underwent surgery, which is a 46% higher surgery rate than in nontriaged patients and a 29% improvement from our institution's existing triage system. Conclusion: The data-driven triage model and scoring system derived and validated in this study (Spine Surgery Likelihood-11), significantly improved existing processes in predicting the likelihood of undergoing spinal surgery within 1 year of initial presentation. This triage system will allow centers to more selectively screen for surgical candidates and more effectively direct patients to surgeons or nonoperative spine specialists. Level of Evidence: 4