Preparing for Medication Adherence Conversations

The Problem

Medication adherence in long-term endocrine conditions is a persistent clinical challenge. A patient on levothyroxine for twenty years who is inconsistently taking it. A patient with osteoporosis who has quietly stopped their bisphosphonate because of side-effect concerns they have not mentioned. A patient with type 2 diabetes who stopped their medication after reading something alarming online.

Having the conversation well takes preparation. You need language that is non-judgmental, that acknowledges the patient's concerns, and that explains the consequences of non-adherence clearly without frightening them into avoidance. You probably say similar things across many patients — but arriving at exactly the right words in a ten-minute slot is not always easy.

How AI Helps

You can ask an AI to draft talking points for a specific adherence conversation. Give it the condition, the common reasons patients stop that particular medication, and the tone you want. It will produce a structured set of points you can review and adapt.

You can also ask it to produce a short written explanation for the patient — something they can read at home when the anxiety about side effects kicks in at 11pm. Giving patients written material to counter internet misinformation is a recognised approach; AI can help you produce that material more quickly.

A Real Example

Dr Patel is seeing "Mrs G", a 68-year-old with osteoporosis who stopped her oral bisphosphonate six months ago citing jaw concerns. He wants to have a careful conversation that acknowledges her concerns, explains the actual risk in perspective, and explores whether an alternative approach might work.

He asks the AI:

"Draft talking points for a consultation where I need to discuss bisphosphonate adherence with a patient who has stopped because of concerns about jaw complications. I want the tone to be non-judgmental, empathetic, and honest about risk. Cover: acknowledging her concern, explaining the actual frequency of the complication in patients not having dental surgery, the risks of untreated osteoporosis, and the possibility of discussing alternatives. Bullet points, not a script. No specific drug names or doses."

The AI produces five clear talking points. Dr Patel reads them, removes one that he feels underplays the legitimacy of her concern, and adds a note about asking what she has read and where. He now has a framework for the conversation rather than arriving unprepared.

Try It Yourself

Draft consultation talking points for an adherence conversation with a {{patient_age}}-year-old patient with {{condition}} who has {{adherence_issue}}. 

I want to:
1. Acknowledge their concern without dismissing it
2. Explain what the evidence says about {{specific_concern}} in plain language
3. Explain the clinical risk of not treating the condition
4. Open a discussion about whether alternatives exist

Tone: warm, non-judgmental, honest. Bullet points. No specific drug names or doses.

Do not include real patient names, dates of birth, or any identifying details.

Things to Watch For

AI talking points are starting points, not scripts. Every patient is different. Use the AI output to prepare, then respond to the actual person in front of you.

Check any statements about risk. If the AI produces a statement like "the complication affects fewer than 1 in 100 patients", verify this against your own clinical knowledge before saying it to a patient. Do not quote statistics you have not verified.

Empathy cannot be outsourced. The AI can give you words. The relationship, the tone of voice, the moment of genuine acknowledgment — that is yours.

Do not share AI-generated talking points directly with patients as if they are official clinical documents. They are your preparation tool, not a patient resource.

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