Why active recall is the only revision strategy that survives the GPhC Common Registration Assessment

The trainee pharmacists who pass the Common Registration Assessment first time and the trainees who do not rarely differ in how many hours they worked. They differ in how those hours were spent.

This is not a motivational point. It is a statement about cognition. The CRA is not a test of how much pharmacy you have read. It is a test of what you can retrieve, apply, and adjudicate against the clock when a clinical scenario you have never seen before is in front of you. Those are different abilities, and they are produced by different study methods.

This article sets out, in concrete terms, why active recall outperforms every other revision technique on this exam, what the science says about why, and how to translate that into a revision week that actually moves the needle.

What the CRA is actually measuring

The Common Registration Assessment is a single-day, computer-based examination delivered through Surpass test centres. Part 1 contains 40 pharmaceutical calculations with numerical free-entry responses, allotted 120 minutes. Part 2 contains single-best-answer questions delivered at roughly 75 seconds per question, drawn from clinical therapeutics, law, governance, regulation, and pharmacy practice. Calculators are permitted in both papers. Both parts must be passed in the same sitting; there is no compensation between them, and the cap is three attempts.

From 2026, the assessment has been harmonised across all training pathways and the independent prescribing learning outcome (LO37) has been removed from the question pool. The pass standard is set using a modified Angoff methodology with one Standard Error of Measurement added, then maintained between sittings via Item Response Theory. In practice this means the pass mark moves with question difficulty, but the standard expected of you does not.

The recent failure rates are instructive. The November 2024 sitting saw 42 per cent of candidates fail. June 2025 was better, but still 23 per cent. These are not edge cases. They are the predictable outcome when capable graduates revise the wrong way.

The reason this happens is almost always the same. The CRA does not reward recognition. It rewards retrieval and judgement under time pressure. Most candidates train recognition.

The fluency illusion

Cognitive psychology has a name for the trap most trainees fall into during their final months of preparation. It is the fluency illusion: the feeling, when re-reading a chapter or a set of notes, that you understand the material. Fluency is the experience of processing something easily. It is not the same as the ability to produce that information from memory when prompted by a question.

Re-reading produces high fluency very quickly. After three passes through a chapter on antibiotic stewardship, most of the text feels familiar. That sense of familiarity is taken, wrongly, as evidence of mastery. The text is recognised. The clinical decision it supports has not been retrieved a single time.

In the now-classic 2006 study by Roediger and Karpicke at Washington University, students who re-read a passage repeatedly performed worse on a delayed test than students who read once and then practised retrieving the information from memory, even when the retrieval group felt less confident in the moment. The follow-up work by Karpicke and Blunt in 2011 found that retrieval practice produced better learning than concept mapping, despite concept mapping being widely taught as an active study method.

The 2013 review in Psychological Science in the Public Interest by Dunlosky and colleagues evaluated ten common study techniques across hundreds of trials. Two emerged as high utility across age groups, content domains, and assessment formats: distributed practice (spacing study sessions over time) and practice testing (retrieving information from memory under test-like conditions). Highlighting, summarising, and re-reading were rated low utility.

This is the empirical foundation on which any serious approach to the CRA needs to rest. The exam is built around applied retrieval. The revision strategy has to be too.

What active recall actually is

Active recall is the practice of forcing yourself to produce information from memory rather than recognising it on a page. In the context of CRA revision, this looks like:

• Closing the BNF chapter and writing out, from memory, the indications, cautions, and monitoring requirements for a drug class before checking what you missed.

• Reading a clinical vignette, pausing before scrolling to the answer options, and committing to a diagnosis or action.

• Working through a calculation on paper without consulting a worked example, then reconstructing the method from your own steps once finished.

The discomfort of not being able to recall something is not a sign that the technique is failing. It is the technique. Each effortful retrieval, even one that ends in error, strengthens the memory trace and the retrieval pathway. Robert Bjork at UCLA has spent four decades describing this as a desirable difficulty. The conditions that make learning feel slow and frustrating in the moment are often the same conditions that make it durable.

The corollary is harder. Methods that feel productive in the moment, including colour-coding notes, watching summary videos at 1.5 speed, and re-reading with a highlighter, are usually producing fluency without learning.

Spaced repetition: the second engine

Active recall on its own is powerful. Active recall combined with spaced repetition is the dominant evidence-based study system for high-stakes professional examinations.

Hermann Ebbinghaus, working in Berlin in the 1880s, demonstrated that newly learned material decays exponentially in the absence of retrieval. A single review at the right interval flattens the forgetting curve dramatically. Three or four spaced reviews, timed to land just before forgetting would otherwise occur, produce something close to durable memory.

In practical terms, this means a question on creatinine clearance encountered today should be encountered again in two days, then five, then ten, then three weeks. Tools that automate this scheduling, including modern adaptive question banks, do the algebra for you. The effect is not magic. It is simply the deliberate exploitation of how human memory consolidates.

The reason this matters for the CRA specifically is that the syllabus is enormous. Trainees attempt to keep the entire BNF, the Medicines Ethics and Practice guide, the GPhC standards, NICE clinical guidelines, and the Misuse of Drugs Regulations in active memory across a six- to twelve-month preparation window. Without spacing, retention collapses by exam day. With it, retention holds.

Calculations are a separate skill, not a topic

Part 1 deserves its own treatment. The most common mistake trainees make is treating calculations as a chapter of revision rather than a skill to drill. They are not the same thing.

A calculation is a procedure executed under time pressure. Three minutes per question is not generous, and the Surpass interface requires numerical free entry rather than multiple choice, which removes the safety net of pattern-matching to plausible answers. The errors that fail Part 1 are almost always procedural: a unit conversion missed, a decimal place dropped, the Cockcroft-Gault equation applied with the wrong weight, a displacement value forgotten in a reconstitution.

Active recall applied to calculations means timed drills, not worked examples. It means doing twenty questions in sixty minutes, then reviewing every error to identify whether the failure was conceptual (you do not know the method) or procedural (you know the method but slipped). Conceptual gaps need teaching. Procedural slips need repetition until they stop happening, which is a different intervention. Conflating the two is why trainees plateau.

The pass mark for Part 1 in the most recent published sittings has hovered around 60 to 65 per cent of marks available, depending on question difficulty. That is not a mark you reach by understanding the methods. It is a mark you reach by being able to execute them quickly and accurately on a Tuesday afternoon when you are tired.

The pitfalls trainees fall into anyway

Five errors recur, almost invariably, in the trainees who fail.

Doing easy questions for the dopamine. The questions that feel good to answer are the ones you already know. A revision session that finishes with a 90 per cent score on questions you have done before has taught you nothing. The session should be uncomfortable. The score should be lower than you would like.

Neglecting Part 1 until the final month. Calculations require sustained drilling. Six weeks is not enough to build accuracy and speed simultaneously. Twelve weeks is realistic. Begin Part 1 from week one of revision.

Reading feedback without rewriting the method. When a question is wrong, the temptation is to read the explanation, nod, and move on. Without writing out the corrected approach in your own words, the correction does not consolidate. It is recognised, not learned.

Avoiding the Surpass interface until late. The Surpass platform is not Anki. The on-screen calculator, the highlighting tool, the PDF resource viewer, and the time pressure of the actual interface need to feel familiar by exam day. Practising on paper is not preparation. It is preparation-adjacent.

Studying alone, indefinitely. A single-page summary written for someone else, or a five-minute teach-back to a peer over a coffee, is one of the most efficient retrieval exercises available. The Feynman technique, properly executed, exposes every gap in a topic in roughly the time it takes to drink the coffee.

A revision week that respects the evidence

There is no single correct schedule. There is only the question of whether your week does the things the science says matter.

A defensible structure looks like this. Two to three hours per day across most days, divided between two activities. The first is timed question practice from a curated bank, with deliberate exposure to weak topics rather than comfortable ones. The second is targeted reading, but only in response to errors identified during the question session, not as a free-floating activity.

Two days a week, do a longer session of 90 minutes of mixed-mode questions under timed conditions, simulating exam pressure. One day a week, do a Part 1 drill of 40 calculations in 120 minutes, sat down, no interruptions, no looking things up.

Once a fortnight, sit a full mock paper. Once a month, review the Board of Assessors feedback documents from previous sittings to identify the topics current cohorts are getting wrong. Those topics are likely to be in your sitting too.

What competence looks like

Confidence and competence are different things, and the CRA is designed to expose the gap between them. A trainee who has read the BNF cover to cover three times is fluent in pharmacology. A trainee who can produce, under time pressure, the correct dose adjustment for a renally impaired patient on a drug she has not seen since November is competent.

Active recall builds the second of these. Re-reading builds the first, and the first is not what you are tested on. The most useful thing a trainee pharmacist can do in the months before sitting the CRA is internalise that distinction and design every revision session around producing answers from memory rather than recognising them on a page.

The exam, in the end, is not asking whether you have studied. It is asking whether you can be trusted on a Tuesday afternoon with a real patient and a difficult prescription. Active recall is the only revision method that trains for that question directly.