Risk factors for unplanned readmission within 30 days after pediatric neurosurgery: a nationwide analysis of 9799 procedures from the American College of Surgeons National Surgical Quality Improvement Program

Abstract

OBJECTIVE Hospital readmission rate is increasingly used as a quality outcome measure after surgery. The purpose of this study was to establish, using a national database, the baseline readmission rates and risk factors for patient readmission after pediatric neurosurgical procedures. METHODS The American College of Surgeons National Surgical Quality Improvement Program–Pediatric database was queried for pediatric patients treated by a neurosurgeon between 2012 and 2013. Procedures were categorized by current procedural terminology (CPT) code. Patient demographics, comorbidities, preoperative laboratory values, operative variables, and postoperative complications were analyzed via univariate and multivariate techniques to find associations with unplanned readmissions within 30 days of the primary procedure. RESULTS A total of 9799 cases met the inclusion criteria, 1098 (11.2%) of which had an unplanned readmission within 30 days. Readmission occurred 14.0 ± 7.7 days postoperatively (mean ± standard deviation). The 4 procedures with the highest unplanned readmission rates were CSF shunt revision (17.3%; CPT codes 62225 and 62230), repair of myelomeningocele > 5 cm in diameter (15.4%), CSF shunt creation (14.1%), and craniectomy for infratentorial tumor excision (13.9%). The lowest unplanned readmission rates were for spine (6.5%), craniotomy for craniosynostosis (2.1%), and skin lesion (1.0%) procedures. On multivariate regression analysis, the odds of readmission were greatest in patients experiencing postoperative surgical site infection (SSI; deep, organ/space, superficial SSI, and wound disruption: OR > 12 and p < 0.001 for each). Postoperative pneumonia (OR 4.294, p < 0.001), urinary tract infection (OR 4.262, p < 0.001), and sepsis (OR 2.616, p = 0.006) also independently increased the readmission risk. Independent patient risk factors for unplanned readmission included Native American race (OR 2.363, p = 0.019), steroid use > 10 days (OR 1.411, p = 0.010), oxygen supplementation (OR 1.645, p = 0.010), nutritional support (OR 1.403, p = 0.009), seizure disorder (OR 1.250, p = 0.021), and longer operative time (per hour increase, OR 1.059, p = 0.029). CONCLUSIONS This study may aid in identifying patients at risk for unplanned readmission following pediatric neurosurgery, potentially helping to focus efforts at lowering readmission rates, minimizing patient risk, and lowering costs for health care systems.