Study design: Retrospective review.
Objective: Determine if factors under surgeon control (anchor density or pedicle screw density) or those not under surgeon control (curve magnitude, levels requiring fusion, and curve flexibility) correlate with standard, short-term quality and outcome measures for adolescent idiopathic scoliosis.
Summary of background data: Pedicle screw fixation has revolutionized posterior spinal instrumentation and fusion for adolescent idiopathic scoliosis and seems to provide greater radiographical coronal plane curve correction than less expensive constructs. Other clinically relevant improvements in outcome have been difficult to demonstrate.
Methods: Retrospective review of 119 posterior spinal instrumentation and fusion cases for adolescent idiopathic scoliosis by 4 surgeons at 1 institution. Average follow-up was 586.7 days. Outcome measures were main thoracic curve correction, complications, reoperations, infection, intensive care unit days, length of stay, estimated blood loss, transfusion, procedure time, implant charges, and total hospital charges. "Surgeon-dependent" variables were implant density (fixation/instrumented level) and pedicle coefficient (implant density × percentage of anchors that are pedicle screws). "Surgeon-independent" variables were main thoracic curve magnitude, main thoracic curve flexibility, and levels fused. Correlations were estimated using Pearson correlation coefficients. One-way analysis of variance was used to estimate the effect of "type of surgeon" or "surgeon" on surgeon-dependent variables.
Results: Complications, reoperations, and infections did not correlate with surgeon-dependent or surgeon-independent variables. Main thoracic curve correction correlated strongly with curve flexibility (correlation coefficient [cc] = 0.4089, P < 0.0001). Surgeon-independent variables were levels fused correlated significantly with procedure time (cc = 0.610, P < 0.001), hospital charges (cc = 0.309, P < 0.001), hospital length of stay (cc = 0.366 [P < 0.001]), implant charges (cc = 0.199, P < 0.047), and estimated blood loss (cc = 0.243, P < 0.013). Surgeon-dependent variables were implant density significantly correlated with implant charges (cc = 0.243, P < 0.015) and inversely with length of stay (cc = -0.236, P < 0.015). Pedicle coefficient was not significantly correlated with any outcome measure.
Conclusion: Levels fused, a surgeon-independent variable, had the most consistently strong correlations with standard short-term quality indicators. With physician grading by payers largely dependent on easily measured outcomes from medical records, hospital and billing records, physicians need to be aware of the surgeon-dependent and surgeon-independent variables that may affect their outcomes and cost-effectiveness profile.
Level of evidence: 3.