Drafting Chemotherapy Side-Effect Information Guides

The problem

Patients starting chemotherapy are anxious. Many are frightened by what they have read online — which is often inaccurate, alarming, or focused on worst-case experiences. They arrive for their first cycle with a list of fears and leave with a verbal explanation that they will have largely forgotten by the time they feel their first side effects.

Written information helps. But many printed side-effect guides are outdated, written at too high a reading level, or generic to a drug class when the patient needs something that feels relevant to their specific situation. When patients do not understand what is normal, they ring the oncology helpline for every headache. When the information is clear, they know what to expect and when to genuinely seek help.

How AI helps

An AI tool can draft a patient-facing side-effect guide for a specific treatment category. You describe the treatment type, the patient profile, and the key side effects to cover. The tool drafts. You review against current clinical guidance, adjust the language for your patient, and use it.

The tool is particularly good at producing graduated information — explaining the difference between expected side effects that the patient manages at home, and symptoms that need prompt clinical assessment. That distinction is exactly what reduces unnecessary calls while catching the genuinely important ones.

A real example

Dr Preethi is a clinical oncologist preparing a patient in her fifties for her first cycle of platinum-based chemotherapy for ovarian cancer. The unit's existing information sheet is three years old and does not reflect the current management of nausea, which her team has significantly improved since then.

She opens an AI tool and types:

Write a patient information guide about the common side effects of platinum-based 
chemotherapy. The patient is a woman in her mid-fifties. This is her first course 
of chemotherapy.

Cover:
- Nausea and vomiting (and that anti-nausea medication will be prescribed)
- Fatigue
- Hair loss (note that this varies by specific regimen — acknowledge uncertainty 
  without being alarming)
- Increased infection risk and what to watch for
- The difference between expected symptoms to manage at home 
  and symptoms that need same-day clinical contact

Tone: calm, clear, and reassuring without minimising. Do not include 
specific drug names or doses. Maximum 450 words.

The tool produces a draft. Dr Preethi reads it. She adjusts the nausea section to reflect her unit's current anti-emetic protocol (without specific doses, as the tool correctly omitted these). She adds the oncology nurse helpline number. She removes one sentence that she feels understates the fatigue patients on her regimen typically experience. The guide is ready in twelve minutes.

Try it yourself

Things to watch for

The tool does not know your unit's protocols. Management of nausea, infection risk thresholds, and contact pathways vary between centres. Any clinical thresholds in the draft — "contact us if your temperature is above X" — must be replaced with your own unit's specific guidance. Do not leave the tool's suggestions in place.

It may oversimplify or misframe some side effects. Hair loss, fatigue, and the impact on daily life are experienced very differently by different patients. The tool's language tends toward the middle — not alarming, but not always specific enough. Adjust the draft for the individual in front of you.

Immunotherapy and targeted therapy have very different side-effect profiles. Be specific in your prompt about the treatment category. A generic "chemotherapy side effects" guide will not prepare a patient on checkpoint inhibitor therapy for immune-related reactions. The more specific your prompt, the more relevant the output.

The emotional content requires your voice. The best side-effect guides acknowledge that treatment is hard and that it is normal to feel frightened. A tool can produce that sentiment. Whether it reads as human and kind — or as formulaic — depends on how much you adjust it.

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