With the growing number of options to configure alerts and ongoing concerns about alert fatigue, we needed to understand which OPA trigger would do the best job of eliciting responses. When we needed a new alert to prompt Gynecologic Oncology providers about expected mortality, we built it in 3 different versions randomized by patient and published our analysis of their differences in Applied Clinical Informatics. In 685 patients across 1,786 outpatient encounters, meaningful responses were almost 5x higher for required Open Encounter (94.8% of encounters) and required Navigator (90.1%) alerts than for Storyboard alerts (19.7%) (p<0.001). When responses were given, they were most likely meaningful with Storyboard (98.3% of responses) and least likely with Open Encounter (68.0%) (p<0.001). Clinical responses differed by trigger as well, with Navigator responses least likely to indicate poor prognosis. (Note: This has not been presented at Epic, but it has been proposed as a talk for XGM 2025.)