By integrating an automated referral system into electronic health records, Winship researcher, Ravi Parikh, MD, and colleagues have demonstrated a scalable way to increase palliative care access and optimize treatment decisions for patients with cancer.
Ravi B. Parikh, MD, MPP
A new study published in JAMA Network Open demonstrates that embedding an algorithm-driven referral system into electronic health records (EHRs) significantly increases palliative care utilization for patients with advanced cancer. Palliative care is a specialized medical approach focused on improving quality of life for people with serious illnesses, including advanced cancer, by managing symptoms and providing emotional support. The Be-A-PAL trial, a pragmatic randomized controlled study, tested an automated default referral system for outpatient palliative care across 15 clinics in the Tennessee Oncology network, a major community oncology provider.
Ravi B. Parikh, MD, MPP, medical director of the Data and Technology Applications Shared Resource at Winship Cancer Institute of Emory University, helped lead in designing and conducting the study while previously based at the University of Pennsylvania. Partnering with Tennessee Oncology, the research team demonstrated how collaborations between academic centers and community oncologists can drive scalable solutions to enhance cancer care delivery.
Optimizing Palliative Care Access
The study’s findings revealed a substantial increase in palliative care access among patients in the intervention group, from 8.3% in typical care settings to 43.9% with the algorithm-based referral system. Additionally, the intervention reduced the use of chemotherapy in the last 14 days of life from 16.1% to 6.5%, aligning with national guidelines promoting value-based, patient-centered care.
“There are two primary findings of our study,” says Parikh, who also serves as associate professor, Department of Hematology and Medical Oncology at Emory University School of Medicine. “First, automated palliative care triggers lead to a threefold increase in referrals to specialty palliative care, ensuring that high-risk patients receive greater and timelier access to high-quality care to address their symptoms and goals of care. Second, oncologists and patients were remarkably amenable to automated palliative care referrals, with low rates of opt-out and refusal. This suggests that patients find it acceptable to consider specialty palliative care as a standard part of their oncology care.”
The Be-A-PAL trial provides a scalable model for increasing palliative care access in community oncology settings, where many patients with cancer receive treatment. Researchers note that a broader and more sustained implementation of this approach could further enhance patient outcomes and quality of life metrics.
The study also challenges the prevailing viewpoint that early palliative care shortly after diagnosis is the only way to achieve meaningful impacts. Instead, it demonstrates that a “precision palliative care” approach using automated EHR triggers can also have significant benefits.
“This study represents an important step in optimizing palliative care delivery through EHR-driven interventions,” Parikh adds. “With further refinement and broader implementation, these strategies could contribute to measurable improvements in patient care and quality of life over time.”
The study also highlights the impact of default referral nudges in clinical practice. Rather than requiring oncologists to manually identify and refer patients for palliative care, the algorithm automatically placed referral orders, requiring oncologists to actively opt out if they felt a referral was unnecessary. This approach significantly increased palliative care utilization while maintaining physician discretion.
Algorithm-Based Interventions in Cancer Care
“The Be-A-PAL trial is a powerful example of how algorithm-based interventions can improve cancer care,” says Sandhya Mudumbi, MD, medical director of supportive care (palliative care, psychosocial oncology and integrative oncology) at Tennessee Oncology. “In this pragmatic randomized controlled trial conducted at Tennessee Oncology in partnership with Dr. Ravi Parikh, we tested an automated EHR algorithm that prompted default outpatient palliative care referrals, with an opt-out nudge to oncologists. Compared to usual care—where oncologists determined referrals—we saw a significant increase in palliative care access and improvements in key outcomes. We hope this is just the beginning of many more collaborations to advance cancer care across academic and community oncology settings,” Mudumbi concludes.
Beyond increasing palliative care access, the Be-A-PAL trial underscores the potential for technology-driven interventions to enhance efficiency in cancer care delivery. By automating referral decisions, the algorithm not only reduced the burden on oncologists but also created a streamlined and equitable process for identifying patients in need of supportive care. The reduction in late-stage chemotherapy use further suggests that these interventions may contribute to more appropriate treatment pathways, potentially leading to cost savings and improved patient-centered outcomes.
Moving forward, Parikh and colleagues plan to enhance algorithmic identification of palliative care-eligible patients by incorporating more detailed EHR data and machine learning methods. Additional research will also explore ways to expand this intervention across broader health care settings and investigate behavioral strategies that encourage greater oncologist buy-in for palliative care referrals.
As oncology care continues to evolve, the study’s findings support ongoing efforts to integrate evidence-based digital solutions into routine practice, ensuring that palliative care services reach patients earlier and more effectively.
This study was supported with funding from The Emerson Collective Digital Oncology Care Award.