StackMed journal
How Many Practice Questions Should You Do for MRCP Part 1?
Set a useful MRCP Part 1 question target using blueprint coverage, review time and performance—not somebody else's total.
There is no official practice-question target for MRCP Part 1, and no number guarantees a pass. The useful target is the number you can answer, review and revisit while covering the full blueprint.
Start with coverage
The exam contains 200 best-of-five questions across two papers. The official blueprint spreads them unevenly across 17 headings: clinical sciences receives 25 questions, clinical pharmacology and therapeutics 15, and nine large clinical specialties about 14 each.
A large question total can hide narrow revision. Before chasing a number, check every domain:
- Have you completed an initial set?
- Have you reviewed the main errors?
- Have you met the topic again in a later mixed set?
- Can you separate its commonly confused diagnoses and treatments?
Weight your practice towards large and weak domains without leaving small specialties untouched.
Count questions you have reviewed
Pressing submit is only the start. A question has done its job when you can explain why the preferred answer fits, why your closest alternative loses and what you will notice next time. Correct guesses belong in the review pile too.
If 25 questions reveal several important gaps, stopping to resolve them is better than rushing into the next 25. Completion rate is a poor target when review quality is falling.
Set a weekly number from your available time
Measure one normal study session. How many questions can you attempt and properly review in an hour? Multiply that by the sessions you can realistically protect around your rota. That is a better starting target than a total copied from social media.
Review time will vary. A familiar correct answer needs little attention; a misconception spanning physiology and pharmacology may need reading and another focused set. Leave room for both.
Change the target as revision progresses
Use three phases:
- Sample: short sets across the blueprint. Progress is slower because you are finding gaps.
- Repair: more questions from weak domains, with delayed retesting of previous errors.
- Integrate: more mixed and timed sets to test recognition, pacing and consistency.
Research in health-professions education supports distributed retrieval, but it does not prescribe a universal MRCP question count. Revisiting difficult material is part of the workload, not an interruption to it.
When to do more new questions
Increase breadth when major domains remain unsampled, mixed sets keep exposing new gaps, or timing deteriorates in longer blocks.
Prioritise review when your marked list is growing faster than you revisit it, the same distinctions keep causing errors, or you recognise stems without being able to explain the answer.
A bank should be large enough to support varied practice; it does not follow that you must exhaust it. StackMed contains 13,000+ MRCP Part 1 questions so you can build broad, focused and repeat sessions without relying on a small fixed pool. Explore the MRCP Part 1 question bank, or read the question-bank workflow before setting your weekly target.
Sources
Put this into practice
Use the StackMed MRCP Part 1 question bank for mixed or specialty-focused best-of-five sessions with explanations for every option.