How to Talk to Cancer Patients About End-of-Life Care | Megha Shah, DNP, FNP, OCN

Nobody enters oncology practice to avoid hard conversations. But even experienced clinicians feel uncertain when it comes to talking with a patient about the end of life. There is a particular weight to those conversations, a feeling that saying the wrong thing might take something away from a person who has already lost so much.

What I have come to believe, after years of clinical practice as a nurse practitioner in an outpatient oncology infusion center, is that the discomfort clinicians feel in these moments often has less to do with the patient than with us. Patients, more often than not, want to talk. They want someone to sit down, make eye contact, and not change the subject. What they need is a clinician who knows how to be present for that conversation.

This is what I have learned about how to do that.

The Role of the Oncology Nurse Practitioner in Goals-of-Care Conversations

End-of-life discussions have historically been framed as physician conversations. Goals-of-care meetings, prognostic disclosures, code status discussions: these have traditionally centered on the attending oncologist. That framing is incomplete, and in the outpatient infusion setting, it is often clinically impractical.

As a doctorate-prepared nurse practitioner, I work in partnership with the oncology team to manage the full spectrum of a patient’s care. That includes prescribing and adjusting medications for symptom management, interpreting labs and imaging in the context of treatment response, and participating in shared decision-making conversations with patients and families. When a patient’s disease is progressing or treatment options are narrowing, I am often the clinician they see most frequently. I am in the room when the questions surface that patients have not yet found a way to ask the oncologist.

That positioning is not incidental. It is a clinical responsibility. Patients frequently share things with oncology nurse practitioners that they have not voiced elsewhere, fears they are embarrassed to raise, questions they consider too basic to ask, concerns they have not yet been able to name. When a patient says “I just don’t want to suffer” during a symptom check or “I keep wondering what happens if this stops working” while we are reviewing their labs, that is not a passing comment. That is an opening, and responding to it is part of my clinical role.

Nurse practitioners in oncology are not adjuncts to goals-of-care conversations. We are active participants, and in many practice settings, we are the ones who initiate them.

Getting the Timing Right

There is no universally right time to raise end-of-life topics, but there are clinical patterns that make these conversations both appropriate and necessary. Goals-of-care discussions become more relevant when treatment options are narrowing, when disease is progressing on the current regimen, when a patient’s functional status is declining between visits, or when hospitalizations are becoming more frequent.

In my own practice, I pay attention to trajectory. A patient whose Eastern Cooperative Oncology Group (ECOG) performance status has dropped from 1 to 3 over the course of two months is telling me something about where their illness is going, even if the conversation about that change has not yet happened explicitly. That clinical observation creates an obligation to initiate a conversation, not to deliver a verdict, but to make space for the patient to tell me what matters to them right now.

Some patients raise these topics themselves. A patient who says “I don’t know how much more of this I can handle” or “I want to make sure my family knows what I want” is signaling readiness. Do not let that moment pass without acknowledging it directly.

Other patients will not raise it, and the silence should not be interpreted as indifference or denial. Some patients are waiting for permission. Some are managing their family’s anxiety more than their own. Some are processing information on a timeline that does not match the clinical one. Asking whether someone would like to talk about their wishes is not the same as telling them they are dying, and understanding that distinction may help clinicians initiate these conversations with more confidence.

Language That Helps

The language used in these conversations carries real clinical weight. Certain phrases close the conversation down. Others open it.

Phrases that invite:

• “Some of my patients find it helpful to talk about what they would want if things got harder. Is that something you have been thinking about?”
• “You mentioned feeling more fatigued lately. As we think about your treatment plan together, I want to make sure we understand what matters most to you.”
• “Is there anything you have been worried about that you have not had a chance to bring up?”

Phrases to avoid:

• “There is nothing more we can do.” This is almost never clinically accurate. There is always something that can be done for comfort, symptom control, dignity, and quality of life. This phrasing forecloses hope in a way that misrepresents the full scope of palliative and supportive care.
• “We should talk about what happens when…” The word “when” implies a certainty about timing that rarely exists. “If things were to change” gives the patient more agency and more room.
• “Are you okay with a DNR?” This question drops a complex clinical and ethical decision on a patient without adequate context or preparation. Goals-of-care conversations build gradually toward code status discussions. They do not begin there.

Language that centers what the patient values tends to be more productive than language that centers medical options. “What are you hoping treatment will help you do or feel?” gives far more clinically useful information than “Do you want to continue treatment?” and it puts the patient’s lived experience at the center of the decision-making process, which is where it belongs.

Managing the Clinical Complexity of Symptom Control at End of Life

One area where the DNP’s prescriptive authority becomes directly relevant to end-of-life care is symptom management. Patients approaching the end of life frequently experience undertreated pain, dyspnea, anxiety, and nausea. Addressing these symptoms is not optional, and it is not separable from the goals-of-care conversation.

In my practice, I work in close collaboration with the oncologist to adjust analgesic regimens, prescribe anxiolytics, initiate or titrate opioids for dyspnea and pain, and coordinate with palliative care when symptom burden requires a level of specialization beyond what we are managing in the infusion setting. Having that prescriptive authority matters because undertreated symptoms are one of the primary drivers of patient distress at end of life, and distress makes every other aspect of the conversation harder.

When patients know their symptoms will be managed, that their clinician has both the authority and the clinical tools to keep them comfortable, the goals-of-care conversation becomes less frightening. The fear of suffering is often more paralyzing than the reality of dying. Addressing it directly is part of what makes these conversations possible.

When Families Are in the Room

Family dynamics in oncology are rarely straightforward. Some families are closely aligned with the patient’s stated wishes and prepared to support whatever decisions the patient makes. Others are managing their own grief in ways that can, unintentionally, conflict with what the patient actually wants.

A family member who redirects the conversation every time a patient tries to express fear, or who insists that certain topics should not be raised, is not a barrier to work around. That person is also struggling, and they deserve to be acknowledged. But acknowledging family distress and centering family preferences are different things, and it is worth being clear about that distinction.

Patients and families are frequently at different points in their understanding of the illness and its trajectory. When they are not aligned, a single unified conversation rarely helps. Meeting with the patient and family separately when appropriate, validating each person’s experience, and clearly communicating that your clinical obligation is to the patient’s expressed wishes while supporting the family through the process: this approach is slower but more effective.

It also requires the clinician to be clear about whose voice governs. Unless a patient has lost decision-making capacity, their wishes take precedence. Families are important, and their grief is real. They are not the patient.

When Patients Are Not Ready

Some patients will close a conversation about the end of life before it begins. They will change the subject, say they are not there yet, or become visibly distressed. This is their right, and it should be respected without abandoning the clinical responsibility to revisit it.

Not ready today does not mean not ready ever. A brief acknowledgment: “I understand, and I want you to know this is something we can come back to whenever you are ready” plants something without forcing anything. It signals that the door is open, that the conversation will not be avoided indefinitely, and that it will happen on the patient’s terms when possible.

Patients who are not prepared to discuss prognosis directly are often willing to talk about values. “What does a good day look like for you right now?” or “What are you most hoping to hold onto over the next few months?” can give clinically meaningful information about what matters to a patient without requiring them to accept a prognosis they are not yet ready to hear. That information shapes clinical decision-making even when an explicit goals-of-care conversation has not yet taken place.

Supporting Patients Through These Conversations

The most important thing a clinician can do in an end-of-life conversation is not find the right words. It is stay in the room.

Sitting down signals that you are not rushed. Eye contact signals that you are present. Tolerating silence is harder than it sounds, but silence is often where the most important things get said. The impulse to fill it with reassurance or redirect to a clinical task is understandable and almost always counterproductive.

There are no answers that fix what these patients are facing. The clinician’s job is not to resolve the patient’s fear of dying or undo their grief about what they will lose. The job is to be a steady, knowledgeable, honest presence in the middle of it, to witness what the patient is carrying without looking away, and to make clear that the care continues regardless of where the illness goes.

That kind of presence is a clinical competency. It is learned and developed deliberately, the same way procedural and pharmacological skills are.

Documentation, Team Communication, and Continuity

Goals-of-care conversations are clinical encounters and should be documented with the same rigor applied to any other clinical event. Document what was discussed, what the patient expressed, who was present, what decisions were reached or deferred, and any follow-up required. When advance care planning documents are completed or updated, ensure the entire team has access to that information and that it is reflected accurately in the chart.

Team communication is essential. In a collaborative practice model, the oncology NP and the oncologist are managing a shared patient, and both need to be operating from the same understanding of where that patient stands on goals of care. If a patient has told me during an infusion visit that they do not want further aggressive intervention, that information belongs in the chart and in my next communication with the oncologist. It should not stay in the room.

Handoff matters too. A patient who has just learned that further disease-directed treatment is not recommended may be seen by a different member of the team at their next visit. Passing along the clinical and emotional context of what that patient has been told and how they responded is continuity of care. It directly affects the quality of what that patient receives next.

End-of-life conversations are among the most demanding clinical work in oncology. They are also among the most meaningful. The patients who have allowed me into those conversations, the ones who got to say what they were afraid of, what they were proud of, what they still wanted to do, they stay with me. Not because I said the right things, but because I showed up for a conversation that mattered and did not look for a way out of it.

That is the standard I hold myself to as a clinician. It is one worth holding.