Jensen, Maribeth

Abstract Pain medicine triage plays a crucial role in ensuring patients receive timely and appropriate care by scheduling them to the most suitable treatment path. However, the absence of standardized triage protocols in pain medicine often leads to inefficiencies, including delay of care and wastage of healthcare resources. This study aims to develop a rule-based automated referral triage system leveraging information from patients' medical notes for scheduling patients to specific procedures in the pain medicine department. The proposed triage system, grounded in the knowledge and expertise of clinical providers, processed referral order comments and referring provider notes by iteratively refining the Natural Language Processing (NLP) rules and post-processing rules through intensively reviewing 76 patients. A post-processing regression model was incorporated to further enhance the accuracy. To ensure alignment with real-world practices, the system was integrated into an electronic health record (EHR) platform for real-time application, streamlining scheduling workflows and enhancing usability in daily clinical settings. After three iterations, the proposed NLP and post-processing rules improved accuracy from 76.3 to 80.3% compared to machine learning (ML) approaches in the preliminary study. The post-processing model further increased accuracy to 84.2%. The implementation accuracy of 200 cases for the first 3 months was consistent with our prediction at 83.5%, which concluded that the improvement over ML models (p-value = 0.018) was statistically significant at 95% significance level. This study demonstrates the feasibility and benefits of a knowledge-driven approach to referral triage in specialized medical fields. It lays a foundation for others in building similar triaging solutions to other specialties.

Introduction: Atrial fibrillation (AF) is the most prevalent arrhythmia, placing a substantial burden on the healthcare system. Patients with AF are cared for in a variety of settings including primary care and cardiology clinics, while some require cardiac electrophysiology (EP) clinics offering specialized drugs or procedures. Appropriate triage of AF referrals to these specific clinics will lead to timely and appropriate care and reduce inefficiency in the health system. Goal: To develop a machine learning (ML) model to triage AF referrals to a large Cardiology center to either the Cardiology or EP clinic based on likelihood of requiring an invasive EP procedure or initiation of antiarrhythmic drug. Methods: We retrospectively analyzed 6,755 patients who were referred to CV with referral diagnosis of AF (mean age 70 ± 11 years, 66% male). We employed multiple input variables to predict the suitability for EP clinic defined as need for initiating an antiarrhythmic drug or an invasive EP procedure. These encompassed patient demographics, relevant comorbidities, medications, prior procedures, medical notes analyzed using natural language processing (NLP), and Holter monitoring data. Results: For internal referrals, XG boost model performed the best with AUC at 79%, accuracy at 72%, and F1 score at 68%. This model produces true positive (TP) at 68% and true negative (TN) at 74%, which was better in comparison with triaging using NLP only (AUC = 69%, TP = 58%, TN = 61%), Holter only (AUC = 52%, TP = 51%, TN = 77%), and NLP/Holter (AUC = 67%, TP = 60%, TN = 60%). When supplementing ML model with Holter readings on average HR ≥ 100 and maximum HR ≥ 150, the results can be improved to TP at 70% and TN at 73%. Conclusions: Our model effectively identifies patients who are more likely to require specialized EP care, enabling the automation and increased efficiency of the AF referral triage process. Prospective validation studies are paramount in determining the true value of this tool.