Here is a realistic Tuesday-evening scenario. A GP working through a themed question set meets a vignette: a 58-year-old with newly diagnosed atrial fibrillation and a CHA2DS2-VASc score of 1 in a man. The GP selects "start anticoagulation", reasonably confident. The answer wanted "discuss, considering anticoagulation", and the explanation turns on exactly where current guidance places the threshold and how it frames shared decision-making around it. Mildly annoying, quickly read, easily forgotten.
That moment is the richest CPD prompt medicine routinely produces, and most doctors let it evaporate. The difference between letting it evaporate and turning it into appraisal-grade evidence is about four minutes of structured thinking. This article gives you the structure.
Description is not reflection
First, the distinction appraisal guidance keeps trying to teach. Description records that something happened: "Completed 30 cardiology questions, scored 73 percent, revised AF guidance." Reflection records what it meant: what you believed, why that was wrong, and what follows for your patients. Appraisers can tell the difference in one glance, because description could have been written by anyone, while reflection could only have been written by you. The template below is designed so that completing it honestly forces reflection, because three of its five questions are about your thinking rather than the event.
The five-question framework
For any missed question worth keeping, answer these in one or two sentences each:
What did I think? State the answer you chose and, more importantly, the reasoning behind it. "I chose immediate anticoagulation because I had internalised any elevated score as an automatic trigger."
What was incorrect? Name the actual error, which is usually narrower than the question topic. Not "AF management" but "I had collapsed a threshold-and-discussion recommendation into an automatic rule."
What did I learn? The corrected understanding, with its source. "Current NICE guidance frames this score in this population as a consider-and-discuss decision, weighing bleeding risk and patient preference."
Why does it matter? Connect it to real practice. "I review AF patients most weeks; the difference between an automatic rule and a shared decision changes those conversations."
What will I change? The behavioural commitment, however small. "I will check the threshold table rather than trusting my recalled rule, and I will frame borderline scores as decisions to share."
Five lines, written while the sting of the wrong answer is fresh, and you have a reflective entry stronger than most appraisal portfolios contain.
Letting the Tutor find the real misconception
The framework's hardest line is the second one, because the true error is often not where it appears. This is exactly what a Socratic tutor is for. When the iatroX Tutor engages on that AF question, it does not begin with the guideline; it asks what you chose and why. Articulating "any elevated score means anticoagulate" exposes the actual misconception, an over-generalised rule, which is a different and more fixable problem than not knowing the guideline. The dialogue does the diagnostic work, and the template's second and third lines write themselves from it. Doctors who use the Tutor first and the template second consistently produce sharper reflections, because they are reflecting on the diagnosed error rather than the surface one.
Closing the loop with spaced repetition
Reflection without retention is good paperwork about a gap that will reopen. The final step is scheduling: the missed topic goes back into the adaptive queue, returns as forgetting approaches, and your performance on the return visit tells you whether the correction held. When it does, your CPD entry gains its last line, the one almost no portfolio evidence has: "Re-tested on the topic three weeks later; correction retained." That single sentence converts a reflective note into documented, verified learning.
Copyable appraisal prompts
These are prompts to complete, deliberately not finished reflections; the value of the entry, and its legitimacy, lives in your answers being yours.
For a single missed question: "A question on [topic] revealed that I believed [X]. The current guidance in [source] states [Y]. This matters in my practice because [Z]. I will [specific change], and I re-tested on [date] with [result]."
For a themed session: "A [n]-question set on [area] on [date] identified gaps in [two or three topics]. The most significant was [gap], where I had [misconception]. Having reviewed [sources], I will [change]. Remaining topics are scheduled for spaced review."
For confirmed currency: "A [n]-question assessment across [area] scored [x] percent, confirming my practice is aligned with current guidance. The one area flagged for review was [topic], revisited on [date]."
Frequently asked questions
Should every missed question become a CPD entry?
No; you would drown. Reserve the full template for misses that genuinely surprised you or touch your regular practice, and let spaced repetition quietly handle the rest. A dozen deep entries a year beat three hundred shallow ones.
Can AI write the reflection from my session data?
It can scaffold, prompt and organise, and it should not write the substance; current college guidance is explicit that reflection must remain yours, and an appraiser reading a generated entry learns nothing about you. Use the Tutor dialogue to find the misconception, then write the five lines yourself.