Predicting Failure of Ventricular Shunts in the Emergency Department: The SMaRT (Shunt Malfunction and Revision Triage) Score.

Abstract

To create a simple scoring system to stratify likelihood of shunt failure in the emergency department. We conducted a retrospective cohort study of all children presenting to the pediatric emergency department with concern for ventricular shunt failure from 2017 to 2022. A separate dataset from 2022 to 2023 was used for validation. The main predictors of shunt revision were determined using univariate and multivariable regression models, and a scoring system utilizing the top predictors of ventricular shunt malfunction was created. There were 1167 neurosurgical consultations for shunt malfunction during the study period which resulted in 285 shunt revisions. Utilizing the 10 variables most associated with revision (increased ventricle size, abnormal shunt series, bradycardia, lethargy, altered mental status, vomiting, headache, shorter time interval from last shunt surgery, seizure, and the presence of upper respiratory symptoms), a SMaRT score was developed. For Shunt Malfunction and Revision Triage (SMaRT) score <4, we achieved a receiver-operating characteristic area under the curve of 0.87 with sensitivity of 0.93, specificity of 0.58, positive predictive value of 0.42, and negative predictive value of 0.96. This was confirmed in a separate validation set of 119 patients. A similar analysis excluding imaging data yielded a significant decrease in the discrimination of the scoring system, dropping the receiver-operating characteristic area under the curve to 0.71. The SMaRT score provides a data-driven tool to determine the utility of a neurosurgical consult for shunt malfunction with a high negative predictive value for shunt malfunction in its lowest tier. Use of this tool has the potential to improve triage and resource utilization in the emergency department.