Background: Improving the quality of End of Life (EOL) care continues to be a challenge. Enhanced prognostic awareness is critical for all members of the clinical team. In December 2020, The McKesson Advance Care Planning Enrollment eXtended (APEX) mortality risk predictive analytics model was implemented to improve prognostic awareness in OCM population and improve the timing of initiation of end of life care. (See ASCO 2021 abstract #1560). Methods: The APEX tool was provided in collaboration with the McKesson/US Oncology network analytics team. A process was established for dissemination of the report information. In the pilot, 12 practice locations with varying community landscapes, socio-cultural dynamics, and site clinical personnel resources were selected. At each site clinical leads and physician champions were selected. Education was provided on the tool, prognostic variables, and appropriate interventions. Biweekly, each site was provided a list of stratified patients based on their risk of mortality within the next 90 days. Patients that were identified as “very high” or “high” risk were reviewed by the clinical teams and discussed in routine huddles. Physicians and teams reported their planned interventions before and after mortality risk identification. Results: In the pilot, 105 patients were identified as very high or high risk. Reported interventions included the option to continue treatment, ACP Discussion, hospice referral/enrollment, palliative care referral, or continue close monitoring. Prior to the report, 14 identified patients were admitted to hospice and 30 patients had 1 or more advance directives documented. For 26 patients, treatment changes occurred including hospice enrollment, reduction in chemotherapy dosage, change in regimen, or initiating intensive monitoring. 23 patients indicated on the report expired in the interim between generation of the report and receipt by the clinic. No changes in treatment were made in 22 patients. There was physician reported disagreement with the mortality risk assessment in 4 patients. Conclusions: We describe implementation of a mortality predictive model in our practice. The care teams found the tool useful to identify patients at high risk of mortality. Interventions were varied and we will track the outcomes based on intervention. We are using the information from the pilot to continue refining the tool and implementation.