Optimal Methods for Notifying Clinicians about Epilepsy Surgery Patients

点击翻译按钮获取中文摘要

Project Summary Epilepsy is one of the leading neurological disorders in the United States, affecting more than 479,000 children and over 2 million adults. Approximately 30% of epileptic patients have poor seizure control despite antiepileptic medications and are potential candidates for neurosurgical intervention. Early identification and referral of children who are potential surgical candidates is complex and while relevant guidelines exist, there is no standard process to efficiently identify those patients meeting criteria for neurosurgical intervention. Given the large corpus of note-based data available in the electronic health record (EHR), it is challenging for providers to efficiently retain and process all the pertinent patient information. Natural Language Processing (NLP) and machine learning techniques have been successfully used to evaluate clinical notes and make recommendations in the research setting. However, NLP techniques are rarely integrated into practice to provide real-time clinical decision support. We developed and retrospectively evaluated a NLP system to help identify those patients who meet neurosurgical criteria and therefore enable surgical consults and evaluations to occur sooner. Knowing clinical decision support can improve outcomes of care, our proposed research will implement NLP into clinical practice and develop a decision support mechanism to improve the time to surgery for eligible patients. The objective of this project is to implement NLP directly into clinical care and determine the most effective decision support mechanism for provider adherence to epilepsy surgical consult recommendations. The long- term goal of this project is to reduce the time to initial surgery evaluation for patients with intractable epilepsy by integrating NLP-classification criteria into clinical practice. This project is one of the first in the field to study the integration of NLP recommendations into clinical care. We will use a human factors engineering framework to design and to analyze two different alerting methodologies for the best-fit for clinical workflow to produce the optimum provider adherence while reducing alert fatigue. Epilepsy progress notes can be classified across hospitals, and if successful, the system will be implemented in additional pediatric institutions around the United States.