202. A Population-Based Study of the Durability of Treatment for Childhood Hydrocephalus

Authors: Joseph H. Piatt, Jr., MD, FAANS; Allison LeHanka (Wilmington, DE)


The adoption of survival analysis in the late 1980s gave the study of the management of hydrocephalus a sound scientific basis, but from that time observations have been limited to specialized settings - major children's hospitals, quality improvement programs, and clinical research networks.  We report here the first analysis of population-based data in the United States. 


The Nationwide Readmission Database is compiled by the Healthcare Utilization Project of the Agency for Healthcare Research and Quality in collaboration with the 27 participating State Inpatient Databases.  It encompasses all discharges, and individual patients are assigned a code that permits longitudinal analysis through the calendar year.  Cases were identified by ICD-9 and -10 diagnosis and procedure codes in the 2015 data set.  Only the first encounter for each patient was analyzed for time to any readmission or readmission for surgery.  Rates were estimated by life tables.  Comparative treatment durability was analyzed by Cox proportional hazards regression.


6818 distinct patients 0 to 18 years of age were discharged with a diagnosis of hydrocephalus. Estimated readmission rates at 30 days and 1 year were 15.8% and 51%, respectively.    3153 patients were discharged after surgery for hydrocephalus. Estimated surgical readmission rates at 30 days and 1 year were 10.9% and 33%, respectively.  The 30-day surgical readmission rate after shunt surgery was 10.7%.  Infants were at greater risk of surgical readmission than older children (HR=1.34, p=0.0030). 

infant_v_olderPatients undergoing shunt revision were at greater risk for surgical readmission than patients undergoing initial insertion (HR=1.40, p=0.0024). 



Patients with hydrocephalus are medically fragile, and the durability of treatments for hydrocephalus is less than satisfactory.  These data reflect the experiences of patients outside of research settings, so they can provide a context for quality assessment.