State University of New York Upstate Medical University, Syracuse, NY
Ajeet Gajra , Yolaine Jeune-Smith , Alexandrina Balanean , Amy R. Ellis , Kelly A. Miller , John Showalter , Danielle Bergman , Sibel Blau
Background: Timely integration of palliative care (PC) and hospice management for patients with advanced cancer requires informed clinical decision-making and expectation-setting to help patients realize their end-of-life (EOL) goals. Longer stay with hospice is a quality indicator in oncology care that requires earlier referral to PC and hospice. We have previously demonstrated that an augmented intelligence (AI) tool used to predict 30-day mortality can assist with an increase in referrals to PC and hospice. In this secondary analysis, we report on the impact of the AI tool on the timeliness of referral to PC and hospice prior to death. However, calculating days with hospice across multiple hospice providers and geographies can be challenging. Thus, we used a real-world (RW) measure of 14 to 90 days prior to death as a surrogate for timeliness for hospice and 90 to 180 days prior to death for a PC referral. Methods: Medical records of patients at a large community-based hematology/oncology practice in the Pacific Northwest who experienced a mortality event pre-deployment (January 2017 to April 2018) or post-deployment (May 2018 to June 2021) of the AI tool were electronically reviewed for evidence of a PC or hospice referral. Patients were included if the referral was between 14 to 90 days of the mortality event for hospice care or between 90 to 180 days prior to death for PC. Outcomes for additional timepoints (1-3 days and 4-13 days and >90 days) will be provided at the final presentation. Data were analyzed using a statistical process control chart. Results: Of the patients who experienced a mortality event, the following percentages had been referred pre- and post-deployment of the AI tool: PC 7.1% pre- and 15.0% post-deployment; hospice 11.5% pre- and 32.1% post-deployment. A system shift (≥6 points in a row steadily increasing or decreasing) occurred early after deployment, in June 2018. The overall improvements were 111.0% in PC referrals and 179.1% in hospice referrals within the respective timeframes. Conclusions: Deployment of an AI tool at a hematology/oncology practice substantially increased the proportions of patients referred to PC 90-180 days prior to death and hospice between 14 -90 days prior to death, suggesting a favorable impact on timeliness of referrals. If confirmed in additional studies, the AI-based tool can be utilized to integrate PC early in the management of patients with advanced cancer.
Referral type | Time frame | Pre-deployment (January 2017 to April 2018) | Post-deployment (May 2018 to June 2021) | Overall Improvement |
---|---|---|---|---|
PC | 90 to 180 days before death | 7.1% | 15.0% | 111.0% |
Hospice | 14 to 90 days before death | 11.5% | 32.1% | 179.1% |
Disclaimer
This material on this page is ©2024 American Society of Clinical Oncology, all rights reserved. Licensing available upon request. For more information, please contact licensing@asco.org
Abstract Disclosures
2023 ASCO Annual Meeting
First Author: Christine Ann Garcia
2022 ASCO Quality Care Symposium
First Author: Megan C. Edmonds
2024 ASCO Genitourinary Cancers Symposium
First Author: Karan Jatwani
2023 ASCO Quality Care Symposium
First Author: Aynharan Sinnarajah