A useful CPSA summary compresses the consultation without changing the evidence. It tells the listener what the patient reported, what was actually observed, what those findings suggest and what should happen next. The common mistake is not excessive length alone: it is allowing an inference to become an apparently established fact.
The GMC's CPSA guidance, checked on 13 September 2026, explains that UK medical schools deliver the assessment against GMC requirements. Schools do not therefore have to use identical station instructions. Practise the summary requested by your own assessment, rather than assuming every encounter ends with the same handover.
A summary is an argument with its evidence attached
Consider the difference between "the patient has recurrent falls" and "the patient describes feeling unsteady but has not reported a fall". Both are brief. Only one may represent the consultation accurately. The second also leaves the next clinician a clearer question to investigate.
A good summary preserves consequential uncertainty. It need not repeat every question or include an exhaustive negative history. It should, however, make clear when an important issue was denied, was not asked about, or could not be established. Those categories cannot safely be merged into "no concerns".
Choose the audience before choosing the detail. A supervising clinician may need the reasoning and outstanding decisions. A patient may need a plain-language explanation and agreement about the next step. An examiner's instruction might require only a focused presentation of findings. The same conversation can support all three outputs, but they are not interchangeable.
Try this original encounter
In this fictional practice case, an adult describes intermittent unsteadiness after starting a new medicine. They think the timing may be connected, but cannot remember the exact start date. They have not fallen. They are worried about driving to work. No physical examination or investigation results have been supplied in the exercise.
A weak summary says: "This is medication-related dizziness with a normal examination and no safety concerns. I would reassure the patient."
That sentence invents a cause, invents an examination, removes the patient's concern and closes the decision prematurely. It sounds clinically fluent because it resembles a familiar presentation. Fluency is precisely why it needs scrutiny.
A more faithful summary is: "The patient reports intermittent unsteadiness and suspects an association with a recently started medicine, although the chronology needs clarification. They have not reported a fall, and examination findings are not available in this exercise. Their immediate concern is driving to work; I would clarify the symptom and relevant risks before agreeing an assessment and advice plan."
This is an educational illustration of evidence handling, not a clinical management recommendation. The improved version does not diagnose the fictional patient. It identifies the work still required before doing so.
Use four separate lines before combining them
Write one line under each heading below, then speak the summary without reading the template. The worksheet is an original study aid, not an official CPSA mark sheet.
| Layer | What belongs here | Check before speaking |
|---|---|---|
| Reported information | Symptoms, timeline and concerns attributed to the patient or another source | Did the person actually say this? |
| Observed information | Findings or results genuinely supplied or obtained | Have I turned an unperformed examination into a normal one? |
| Interpretation | A working explanation and relevant uncertainty | Can I point to the supporting information? |
| Next step | The decision or information required, with responsibility where appropriate | Does the plan address the unresolved problem? |
A useful compression test is to remove one sentence and ask whether the listener loses something that changes their understanding or action. A sentence that merely repeats the presenting symptom may be expendable. A sentence explaining why the usual plan cannot yet be agreed may be essential.
Do not shorten a summary by deleting the distinction between evidence and interpretation. Shorten repetition instead.
Review spoken and written versions differently
Record a fictional encounter with a consenting practice partner, then produce a spoken summary and a short written one. The partner should compare both with their role information, not with an imagined ideal consultation.
In the spoken version, listen for unclear sequencing, unexplained abbreviations and the point at which the main decision becomes apparent. In the written version, inspect wording that could survive as a misleading record: "normal", "denies", "agreed" and "discussed" all imply particular evidence or actions.
Now change one detail in the role information. Perhaps the patient did fall, or the symptom preceded the medicine. Rewrite only the portions affected. This tests whether the summary is built from the encounter rather than retrieved as a memorised paragraph.
Keep an error log containing the unsupported phrase, the information actually available and the corrected wording. A small collection of your own recurring assumptions is more actionable than copying an entire model presentation.
Where Geeky Medics and iatroX help
This article is published by iatroX and includes iatroX among the resources considered. As published on 13 September 2026, Geeky Medics offers clinical-skills resources and virtual-patient practice. Demonstrations can help learners recognise an organised encounter, while repeated practice creates material to summarise. Neither removes the need to check local CPSA instructions.
The September 2026 iatroX Simulations release includes UKMLA CPSA practice, transcript-linked feedback and Tutor-led follow-up. A practical use is to compare your summary with the actual transcript and investigate where you introduced an unsupported assumption. This is a proposed review method, not a claim that every summary is automatically assessed or that feedback predicts an examination result.
For a knowledge gap exposed by the encounter, question practice and Socratic Tutor provide a different activity from further presentation rehearsal. For difficulty observing or eliciting physical findings, supervised bedside work remains necessary.
Choose the missing activity, not another collection of scripts
A student who does not know how to organise an encounter may benefit from a clear demonstration first. Someone whose encounters are adequate but summaries are unreliable needs a transcript-and-summary audit. Someone who misses physical findings needs observation and supervision rather than another conversational subscription.
Per iatroX's September 2026 product information, the UK learning subscription is £99 paid upfront for a year, equivalent to £8.25 a month billed annually, or £29 monthly. It combines paid question banks, Tutor, the study planner, simulations and CPD tools; simulations are not a separate add-on. One complete simulation is free, so the relevant decision is whether the practice and review process helps with your identified problem before paying for continued use.
Frequently asked questions
Should every CPSA summary use the same structure?
A consistent private organising method can help, but the final output should follow the task and audience specified by your medical school. A patient explanation and a clinician handover need different wording and emphasis.
Should I include a normal finding if the scenario does not mention it?
No. State that the finding is unavailable or that the examination has not been performed rather than converting missing information into reassurance.
Can I use an AI transcript as the sole check of my performance?
No. It can support review of the recorded interaction, but physical findings, local assessment requirements and the correctness of feedback still require appropriate checking.