For Deborah Miller, a 72 year old resident of Pittsboro, the hardest part of her cancer journey is not the treatment itself but the silence that often follows diagnosis. Living with metastatic breast cancer since 2023, she finds that honest conversations about the future are the most important part of her care. Yet she admits it is hard to never be able to ring the bell, a metaphor for the uncertainty that hangs over every appointment. Her experience highlights a gap in oncology care that researchers at the University of North Carolina are trying to close with a new decision support tool.
The tool is designed to help oncologists recognize when patients with metastatic breast cancer may need discussions about end of life planning and supportive care. It does not rely on artificial intelligence, a deliberate choice by the development team to ensure transparency. Instead, it uses routine information already in a patient’s medical record to estimate risk. This approach aims to give doctors a clear, understandable nudge to start difficult conversations sooner, rather than waiting for a crisis to force the issue.
Reading the Data, Not Guessing
Emily Ray, a medical oncologist at UNC Lineberger and associate professor in the UNC School of Medicine, helped develop the tool. She explained that doctors often struggle to tell patients that their time may be limited. The emotional weight of these conversations is heavy, and the lack of a clear signal can lead to hesitation. The new tool provides that signal by analyzing existing data points such as lab results, vital signs, cancer subtype, and current medications.
The system estimates a patient's risk of death within 30 or 90 days. When that risk is high, the oncologist receives a simple notification flagging the patient. This is not a prediction of fate, but a clinical cue to reassess the care plan. Ray emphasized that the team intentionally avoided using artificial intelligence so that doctors could better understand how the tool reaches its conclusions. In a field where trust is paramount, explainability is a feature, not a bug.

The Case for Early Palliative Care
The ultimate goal of the tool is not just to have a hard conversation, but to lead to earlier use of palliative care. Palliative care focuses on comfort and support for people with serious illnesses. It is often misunderstood as being only for the very end of life, but Ray clarified that it can be provided alongside cancer treatment. In fact, it has been shown to improve quality of life by managing symptoms and supporting emotional well being.
By flagging patients who are at higher risk, the tool aims to bridge the gap between aggressive treatment and comfort care. This allows for a smoother transition where the patient’s goals are central. For someone like Miller, who values being listened to by her care team, this means getting the support she needs earlier, rather than in a rush at the very end. The tool supports this by making the need for such support visible and actionable for the clinician.

A Scale Problem
The need for better end of life planning is not a niche issue. According to the National Cancer Institute, more than 150,000 women in the United States are living with metastatic breast cancer. Each of these patients and their families faces the challenge of navigating a complex medical landscape. The lack of curative options for metastatic disease makes these conversations even more critical, as the focus shifts from cure to control and comfort.
The new tool is a response to this scale. It provides a standardized way to identify when these conversations are most needed, reducing the variability in clinical practice. This consistency can lead to better outcomes and less distress for patients and families. It is a practical solution to a human problem, grounded in data but driven by compassion.

What Comes Next
The study detailing this tool was published in JCO Oncology Practice in September. This publication marks a significant step in the validation of the approach. It provides evidence that such a tool can be effective and useful in a clinical setting. The research team is now moving from study to practice.
UNC plans to pilot the tool in its clinics next. This phase will test whether the tool works in real time and whether oncologists find it useful in their daily workflow. The feedback from this pilot will be crucial in refining the tool and potentially expanding its use to other cancer types or settings. For now, it represents a promising new chapter in how we approach end of life care.
Frequently asked questions


Comments
0