• Aaron J Buckland, Gregory Poorman, Robert Freitag, Cyrus Jalai, Eric O Klineberg, Michael Kelly, and Peter G Passias.
• *Division of Spine Surgery, Department of Orthopaedic Surgery, Hospital for Joint Diseases, New York University Langone Medical Center, c/- Spine Research Center, New York, NY †Division of Orthopaedic Surgery, University of California Davis, Sacramento, CA ‡Department of Orthopaedic Surgery, Washington University, St. Louis, MO §New York Spine Institute/Division of Spine Surgery, Hospital for Joint Diseases, New York University Langone Medical Center, New York, NY.
• Spine. 2017 Jan 6.

Study DesignComparison between national administrative databases and a prospective multicenter physician managed database.ObjectivesThis study aims to assess the applicability of National Administrative Databases (NADs) in Adult Spinal Deformity (ASD). Our hypothesis is that NADs do not include comparable patients as in a physician-managed database (PMD) for surgical outcomes in adult spinal deformity.Summary Of Background DataNADs such as National Inpatient Sample (NIS) and National Surgical Quality Improvement Program (NSQIP) provide large numbers of publications owing to ease of data access and lack of IRB approval requirement. These databases utilize billing codes, not clinical inclusion criteria, and have not been validated against PMDs in ASD surgery.MethodsThe NIS was searched for years 2002-2012 and NSQIP for years 2006 - 2013 using validated spinal deformity diagnostic codes. Procedural codes (ICD-9 and CPT) were then applied to each database. A multicenter PMD including years 2008-2015 was used for comparison. Databases were assessed for levels fused, osteotomies, decompressed levels, and invasiveness. Database comparisons for surgical details were made in all patients, and also for patients with ≥ 5 level spinal fusions.Results37,368 NIS, 1,291 NSQIP and 737 PMD patients were identified. NADs showed an increased use of deformity billing codes over the study period (NIS doubled, 68x NSQIP, p < 0.001), but ASD remained stable in the PMD.Surgical invasiveness, levels fused and use of 3-column osteotomy (3-CO) were significantly lower for all patients in the NIS (11.4-13.7) and NSQIP databases (6.4-12.7) compared to PMD (27.5-32.3). When limited to patients with ≥5 levels, invasiveness, levels fused, and use of 3-CO remained significantly higher in the PMD compared with NADs (p < 0.001).ConclusionsNational databases NIS and NSQIP do not capture the same patient population as is captured in PMDs in ASD. Physicians should remain cautious in interpreting conclusions drawn from these databases.Level Of Evidence4.

10 keywords...

hide…