Workflow: Summarising a Complex Case History

Complex gynaecological cases accumulate years of notes, letters, and results. Reading through them before a new consultation takes time you rarely have. This workflow shows you how to use an AI tool to produce a structured summary — without ever sharing identifiable patient data with the tool.

1. Read through the notes yourself first

   This step is not optional. Before you summarise anything, read the key documents yourself. You need to understand the case well enough to know what is clinically relevant. The tool will only work with what you give it — and you are the one who decides what that is. Allocate five to eight minutes for a complex case with extensive history.

2. Create an anonymised handwritten or typed list of the key facts

   On a blank document or piece of paper, write down the essential clinical facts: age range, presenting complaint, key diagnoses, investigations performed, interventions, current medications (categories, not names, if preferred), and outstanding questions. Use no names, dates of birth, or any other identifying detail. This becomes your working summary.

3. Open the AI tool and describe the task

   Open your chosen tool and begin by telling it what you want. For example: "I am going to give you a list of anonymised clinical facts about a patient. I want you to organise them into a structured summary under the following headings: Background, Investigations, Treatments to Date, Current Status, Outstanding Questions. Do not add any clinical interpretation. Use only the information I provide."

4. Paste your anonymised fact list

   Copy your anonymised fact list into the tool. Keep it factual and bullet-pointed. Avoid narrative sentences, which are harder for the tool to reorganise correctly. The tool will produce a structured summary.

5. Check the output for accuracy

   Read the summary and compare it against your own understanding of the case. Look for: facts that have been misrepresented or combined incorrectly, important clinical details that were lost in the reorganisation, and any interpretive language the tool added that was not in your list. This step is essential. The tool reorganises — it does not understand.

6. Ask for a briefer version if needed

   If the summary is longer than you need for a quick consultation prep, ask the tool: "Summarise this into five bullet points for a pre-consultation briefing." Use the shorter version to refresh your understanding before you enter the room, and keep the full version in your notes.

7. Add your clinical interpretation separately

   The tool's output is a structural summary. It does not contain your assessment of what this case means, what you are concerned about, or what you plan to do. Add that yourself, in your own notes. Keep the AI-generated summary clearly separate from your clinical commentary.

8. Do not save the AI-generated summary in the patient record

   This summary was created from anonymised facts and does not constitute a verified clinical record. Do not paste it into the patient's notes as if it were. Use it as a personal reference tool for consultation preparation, then discard it.

What this means for you

This workflow does not save you from reading the notes. It reorganises information you have already reviewed, into a format that is faster to reference during a consultation. The value is in the structure, not in offloading the clinical reading.

When not to use this workflow

Do not use this approach when the case is straightforward enough that reading the last letter and clinic note gives you everything you need. This workflow is for genuinely complex cases — multiple specialities, years of treatment history, conflicting prior investigations — where the volume of information is the problem.

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