Incidence and Risk Factors for 30-Day Unplanned Readmissions After Elective Posterior Lumbar Fusion
Retrospective study of prospectively collected data.Objective.
To perform a multiinstitutional assessment on the incidence and risk factors for unplanned readmissions following elective posterior lumbar fusion (PLF) surgery.Summary of Background Data.
Understanding what may drive rehospitalizations is a necessary step toward higher quality care. Identifying risk factors for unplanned readmission is especially important for elective PLF, which is a common procedure that is known to be associated with significant adverse events.Methods.
Adult patients undergoing PLF were identified using current procedure terminology (CPT) from the American College of Surgeons National Surgical Quality Improvement Program. Both descriptive and comparative statistics were performed for patient characteristics, clinical factors, and postoperative complications. Subsequently, a step-wise multivariate logistic regression was employed.Results.
Of the 2301 patients who met inclusion criteria for this study, 117 were unplanned readmissions (5.1%). These occurred at a mean of 15.9 days (range: 3–30 days) after surgery. The risk-adjusted analysis revealed that bleeding disorder (odds ratio, OR = 2.8, confidence intervals, CI = 1.0–7.6, P = 0.043), insulin dependent diabetes (OR = 2.5, CI = 1.4–4.4, P = 0.004), and total length of stay > 5 days (OR = 1.8, CI = 1.2–2.8, P = 0.009) were independent predictors for unplanned readmission. Significant postoperative complications included wound complications (OR = 27.6, CI = 13.9–54.8, P < 0.0001), pulmonary embolism and/or deep vein thrombosis/thrombophlebitis (OR = 11.9, CI = 5.0–28.5, P < 0.0001), sepsis (OR = 8.5, CI = 2.3–32.1, P = 0.002), and urinary tract infections (OR = 2.4, CI = 0.9–6.9, P = 0.094).Conclusion.
The unplanned readmission rate for patients undergoing PLF was low, but this study's findings of potentially modifiable risk factors suggest that substantial improvement with this quality metric is possible.Conclusion.
Level of Evidence:N/A