154 Field Testing a Quality Measure for Treatment of Status Epilepticus in the Pediatric Emergency Department

Abstract

Status epilepticus (SE) in children is a neurologic emergency. Although short-term mortality rates are low (3%), SE can lead to poor long term outcomes including disrupted neurodevelopment, increased risk of SE recurrence, and increased long term mortality. SE is “refractory” if seizures continue after first- and second-line therapy. Refractory status epilepticus (RSE) is associated with mortality rates as high as 32%. Expedient treatment of SE is essential for achieving good outcomes, however the average time to first-, second-, and third-line treatments are longer than guidelines recommend at US tertiary care hospitals. To address this issue, the American Academy of Neurology proposed as a quality measure, the proportion of pediatric patients receiving third-line treatment for convulsive RSE within 60 minutes. We field-tested this quality measure with the goal of determining whether it is feasible to measure and whether it actually reflects quality of care. Using the statistical programming language R, we developed a tool to automatically identify patients in RSE and calculate time to third-line treatment using data from electronic medical records. We tested the fidelity of our tool on a large cohort of patients using manual chart review as a gold standard. The cohort included all pediatric patients (age 0-21) admitted to both the pediatric ED and pediatric ICU at New York Presbyterian Cornell Hospital between 2012 and 2017 with ICD-9 and ICD-10 codes corresponding to epilepsy and recurrent seizures as well as convulsions not elsewhere specified and general symptoms. The tool identifies cases by searching for the term “status epilepticus” in radiation reports, EEG notes, and discharge summaries, as well as checking for ICD codes that correspond to status epilepticus. Any case that meets at least one of those conditions is next checked to see if a 2nd or 3rd line medication for SE, generally an IV antiepileptic drug (AED), was given while in the ED or within 3 hours of being transferred to the pediatric ICU. By manual chart review, we identified 62 episodes of convulsive SE out of a total of 665 patient visits. Of the 62 episodes of convulsive SE, 25 responded to first line treatment with benzodiazepines and 16 responded to second line treatment with an IV AED. In 6 cases, the patient was intubated and given paralytics after second line treatment - 2 were subsequently given a third line medication. Finally, there were 14 cases of convulsive RSE that required third line treatment for continued seizures. Our tool identified 15 of the 16 cases that received a 3rd line medication giving a sensitivity of 93.8% and a specificity of 94.9% for detection of convulsive RSE. We have designed a computable phenotype that identifies cases of convulsive RSE with a high degree of sensitivity and specificity. We’ve demonstrated that it is feasible to automate case detection for the quality measure: proportion of pediatric patients receiving third line treatment for convulsive RSE within 60 minutes. This will allow informed quality improvement as well as research into factors that influence quality care for convulsive RSE. Future directions of this work include validating the tool with data from other institutions, automating calculation of the quality measure, and ensuring the quality measure actually reflects quality of care.