AI for Clinicians

Home / Oncology

Use caseIntermediate5 min read

Preparing for Palliative Care and Goals of Care Conversations

How an AI tool can help you structure and prepare for difficult goals of care conversations — without scripting what should always remain a deeply human exchange.

The problem

Goals of care conversations are among the most demanding consultations in oncology. They require precise, honest communication — and simultaneously, great sensitivity to what the patient is ready to hear, what the family is carrying, and what has already been said in prior appointments.

No two of these conversations are the same. But you have them frequently, often back-to-back with other demanding consultations. There are days when you walk in knowing the clinical picture clearly but without having had a moment to think through how to frame it for this patient.

How AI helps

An AI tool cannot have this conversation for you. It cannot read the patient. It cannot hold the silence when someone receives very difficult news. What it can do is help you prepare.

You can ask the tool to help you anticipate what the patient and family are likely to ask. You can ask it to suggest a framework for structuring the conversation. You can ask it to draft the language you might use for a specific part of the discussion — such as explaining what palliative care means and what it does not mean — and then refine that language into something that sounds like you, not like a policy document.

Used carefully, this preparation makes you more present in the room. You have already thought through the hard parts. You are not searching for words at the moment when words matter most.

A real example

Dr Ingrid is a consultant medical oncologist. She is about to see a patient in her early seventies who has been through two lines of treatment for advanced non-small cell lung cancer (NSCLC). The patient's condition has deteriorated. The family — two adult children who attended the last appointment — have been asking directly about what further treatment options exist.

Before the appointment, Dr Ingrid opens an AI tool and types:

Try it yourself
I am preparing for a goals of care conversation with a patient in her early 
seventies with advanced lung cancer who has had two lines of treatment. 
Her condition has deteriorated. Her adult children attended the last appointment 
and have been asking about further treatment options.

Please help me prepare by:
1. Listing the questions the patient and family are most likely to raise
2. Suggesting a structure for this conversation — what to cover and in what order
3. Drafting plain-language wording for explaining what palliative care focuses on, 
   in a way that does not feel like giving up

Do not include clinical recommendations or any language about prognosis.

The tool returns a structured preparation document. Dr Ingrid reads it. She adjusts the suggested opening — the tool's phrasing is slightly formal for how she normally speaks. She keeps the list of likely family questions, which includes two she had not thought to prepare for: one about whether the patient will be in pain, and one about whether she will be able to stay at home.

She walks into the consultation prepared for both.

Try it yourself

Prompt

Things to watch for

The tool's language may sound too formal or too clinical. Any phrasing the tool suggests for use in the consultation itself — sentences you might actually say — needs to be rewritten in your own voice before you use it. Patients hear the difference between a doctor speaking naturally and a doctor reading from something prepared.

It cannot account for what has already been said. The tool does not know the history of this patient's conversations with your team, what the patient has explicitly asked to know or not know, or what informal things a nurse or colleague may have already communicated. You supply that context.

It may produce content that is not appropriate for all care settings. The concept of "palliative care" is framed differently in different health systems and cultural contexts. Review any explanatory language carefully before using it with a patient from a background where the terminology may carry different meaning.

Emotional complexity cannot be outsourced. Preparation helps. The conversation itself requires you to be fully present, reading the patient, adjusting in real time. The tool helps you get ready. It cannot substitute for the clinical skill and human presence that make these conversations genuinely supportive.

Remember: AI is a helpful assistant, not a clinician. You make the call.

Was this lesson helpful?

Related lessons