In the high-pressure environment of modern acute care, the difference between a successful recovery and a cycle of repeated hospitalizations often comes down to timing. For patients living with serious illnesses, the transition from hospital to home can be fraught with complications, leading to a significant number of readmissions. Now, a new collaborative effort between the Mayo Clinic and the clinical intelligence firm Bayesian Health aims to address this challenge by utilizing artificial intelligence to identify patients who would benefit from palliative care earlier in their hospital stay.
This initiative represents a significant shift in how healthcare systems manage complex, life-limiting conditions. By integrating predictive analytics directly into the Electronic Health Record (EHR) workflow, the platform seeks to move beyond reactive care, providing clinicians with actionable data to improve patient outcomes. The project was developed under the Mayo Clinic’s Practice Transformation Ventures (PTV) framework, with the Department of Medicine in Rochester acting as the core validator for the technology.
Addressing a Critical Gap in Serious Illness Care
Industry data indicates that approximately one-third of all hospital readmissions involve patients struggling with serious or chronic illnesses. Despite the documented benefits of early supportive intervention, current clinical workflows often fail to connect these patients with palliative care teams in a timely manner. Frequently, fewer than half of the patients who meet the criteria for a consultation ever receive one during their inpatient stay.
The primary barrier is rarely a lack of clinical intent, but rather the immense complexity of managing data in a fast-paced hospital setting. Clinicians are often overwhelmed by the sheer volume of information in a patient’s longitudinal record. By the time a manual consultation is triggered, the opportunity to prevent aggressive, non-beneficial treatments or prepare for a safer discharge may have already passed. This new AI-powered solution acts as a force multiplier, synthesizing vast amounts of clinical data to provide teams with a clear, hospital-wide view of patient needs.
Validated Clinical Outcomes
The platform’s efficacy is supported by data from a randomized clinical trial conducted internally by the Mayo Clinic’s Department of Medicine. The trial evaluated an earlier version of the software, which demonstrated significant improvements in both clinical processes and patient outcomes. According to the validated findings, the implementation of the tool was associated with a 44% increase in timely palliative care referrals. The data showed a 25% reduction in 60-day readmissions and a 28% reduction in 90-day readmissions.
Jacob J. Strand, M.D., chair of Palliative Care at Mayo Clinic, has noted that the fundamental challenge in palliative medicine has been identifying the underlying patient need early enough to meaningfully alter the course of care. By providing specific, patient-centered signals to both bedside and central teams, the technology allows for more consistent decision-making, effectively cutting through the noise of inpatient complexity.
How the Technology Works
Unlike traditional diagnostic tools that may focus on isolated vital signs, the Bayesian Health platform continuously analyzes the entire longitudinal electronic record of a hospitalized patient. It applies complex clinical reasoning to identify subtle, compounding changes that might suggest a decline in health or an increase in caregiver distress. This approach allows the system to compare real-time data against a patient’s individual baseline rather than relying on generic population averages.
To ensure the tool remains useful for frontline staff, the developers implemented a dual-sided operational dashboard. This design is intended to minimize “notification fatigue,” a common issue with clinical alerting systems. Palliative consultation teams gain a real-time, hospital-wide overview of vulnerable patients, while bedside clinicians receive clear, interpretable guidance directly within their existing EHR workflow. The system also utilizes a continuous reinforcement framework, meaning the underlying models learn from clinician feedback and shifting patient trends, theoretically improving in accuracy over time.
Looking Ahead
As healthcare systems globally continue to grapple with the rising costs and clinical complexities of serious illness management, the integration of AI-driven clinical intelligence is becoming increasingly vital. The collaboration between the Mayo Clinic and Bayesian Health highlights a proactive approach to patient care, focusing on the early identification of needs to ensure that supportive interventions are delivered when they are most effective.
The Mayo Clinic has not yet announced a widespread rollout schedule or additional trial phases for the current iteration of the platform. For those interested in the ongoing evolution of medical technology and the future of palliative care, updates regarding institutional innovations and clinical research developments are typically shared through official Mayo Clinic communication channels. We encourage our readers to share their thoughts on the integration of AI in clinical settings in the comments section below.
Keep reading