Automating the identification of patients admitted to tertiary care hospital with immunotherapy related adverse events (irAEs).

Abstract

566 Background: Immune checkpoint inhibitors (ICIs) have revolutionized oncology by offering prolonged survival in advanced cancers that were previously treatment-refractory. However, the effectiveness of these agents is accompanied by the potential for autoimmune toxic effects known as immune-related adverse events (irAEs). These adverse effects are difficult to predict, can be long-lasting, and occasionally fatal. Early identification and treatment of irAEs can reduce morbidity and enable prompt resumption of anti-cancer therapy, ultimately improving patient outcomes. Methods: We manually reviewed 658 patient charts of individuals who received ICIs and were admitted to a tertiary care hospital over a three-month period. This process identified 750 phrases indicative of irAEs. Subsequently, an in-house Epic reporting team developed a program to search for these specific phrases within the charts of patients admitted to the hospital who had received ICIs in the 24 months prior to their admission. During the program's development, its performance was compared to the previous manual chart review process. Results: A comparative analysis was conducted between the new program and the current manual chart review over a three-month period. Implementing the new program resulted in a 68% reduction in the average number of charts requiring manual review (69 vs. 219 per month). Furthermore, the program identified an additional 17% of cases (21 vs. 18 per month). The positive predictive value increased by 400% (32% vs. 8% per month), and the time required to complete the task decreased by 70% (4.6 vs. 15 hours per month). Conclusions: Utilizing an EMR-based program that automatically searches for patients admitted with irAEs can significantly shorten the time required for identification and subsequent management of these complex patients.