Introduction: The relative paucity of pediatric neurosurgical conditions necessitates tremendous population sizes in order to robustly assess the quality and outcomes of present care. Although intrinsically limited by their coarse resolution, national databases are becoming increasingly powerful and validated to the point of enabling the regimented analysis of outcomes and their predictors.
Methods: Statewide inpatient datasets comprising 100% of hospital discharges in the 3 most populous states with longitudinal identifiers (CA, FL, NY), which track patients across multiple admissions, institutions, and years, were examined for 2006-2010. Algorithms constructed from ICD-9 codes and validated with chart review were used to identify all-cause craniotomies and craniectomies, prospectively-designated predictors and outcomes. Initial operations in 2006 or 2010 were excluded to ensure at least 1yr of follow-up, age >21 and prior craniotomy were also excluded. Primary outcomes, secondary outcomes, and predictors (including reason for operation, neurological sequelae, hospital volume, Charlson comorbidity index, and demographics) were analyzed with logistic regression.
Results: CA, FL, and NY together comprise 24% of the US population. There were a total of 11,135 all-cause craniotomies for 10,159 kids over the 5-year sample, with 6,027 patients remaining after exclusion, at a median inflation-adjusted admission cost of $28,542 (IQR:16,498-54,623). Postoperatively, 9% of patients were readmitted for re-operation (median time to readmission: 139days, IQR:42-371) and 34% were readmitted with new-onset primary outcomes including convulsions (8%; median 174days, IQR:50-366), hydrocephalus (6%; median 105days, IQR:36-296). , and hemorrhage (4%; median 84days, IQR:26-239). Additionally, there were readmissions with sepsis (4%; median 148days, IQR:59-301), meningitis (3%; median 42days, IQR:14-312), and VFib (0.3%; median 272days, IQR:112-549). Notably, 4% of children died during the initial admission, at a median of 4days after craniotomy (IQR:2-9), but only a 0.3% mortality on the day of surgery.
Conclusions: Children were most frequently readmitted for outcomes associated with neurological condition and comorbidities at presentation. These findings reinforce the need for extended vigilance and careful follow-up of children, in which many neurosurgical sequelae present beyond the traditional postoperative periods of 30-180days.
Patient Care: These findings reinforce the need for extended vigilance and careful follow-up of children, in which many neurosurgical sequelae present beyond the traditional postoperative periods of 30-180days.
Learning Objectives: Participants should be able to better understand the incidence of 1) pediatric craniotomies, 2) associated presenting conditions, neurological sequelae, and comorbidities, and 3) costs associated with resulting outcomes.