Abstracts

Rationale: Seizures are one of the top 10 reasons for hospitalization in children but the incidence of seizures in children with epilepsy hospitalized for other reasons is unknown. Breakthrough seizures in children with epilepsy can progress to status epilepticus as well as being disturbing to patients, parents and medical staff. Factors in addition to baseline seizure control may determine whether seizures occur in hospitalized children with epilepsy. Children with underlying neurologic problems are more likely to experience medication errors, more likely to have difficult to control seizures, and more likely to be hospitalized. Study of this topic would benefit from a conceptual model and the development of a data acquisition tool based on the model Methods: Review of the literature and interviews with stakeholders, including parents of children with epilepsy, nurses, pediatricians, pediatric neurologists, epileptologists, quality improvement experts,and hospital administrators were undertaken to assess the potential causes and significance of breakthrough seizures. Results: Key informants and the medical literature confirmed that this topic area has been under studied, and is important. Three domains of risk factors for breakthrough seizures emerged: Baseline Patient Risk Factors, i.e. underlying neurologic and systemic problems; Acute Patient Risk Factors, i.e. concurrent systemic illness; and Systems Factors, including medication errors and changes in medication timing. Using MedQuest - Clinical Data Collection Design System (v.720), a medical record abstraction tool was designed to capture these risk factors for each domain together with the outcome of interest: breakthrough seizures and their consequences. The tool also requires all anticonvulsant medication orders and administration times to be captured systematically. This feature avoids requiring subjective decision making by abstractors related to the presence or absence of a medication error. Pilot testing of the tool to abstract data from 30 charts of children with epilepsy admitted for reasons other than epilepsy was successful: salient features of breakthrough seizures documented in the medical record were captured by the tool. Risk factors modeled by the three domains above, including anticonvulsant medication errors, when present in the charts, could be tracked with the tool. Conclusions: Breakthrough seizures in children with epilepsy admitted to the hospital for reasons other than epilepsy may in part be preventable, but little is known about the scope of this problem, or the factors associated with it. Using a conceptual model as a framework, we have developed a structured medical record abstraction instrument that will permit the collection of data within each of the risk factor domains. The data obtained using this instrument will provide information on the incidence, associated risk factors including medication errors, and consequences of breakthrough seizures.