Your AKT question-bank score may have stopped improving because repeated-item recognition is masking unresolved gaps, the question mix has become harder, review is too passive, or timing is exposing a different problem. The percentage alone cannot distinguish these explanations. Compare the conditions and inspect the errors before deciding that you need a new bank or more questions.
This article proposes a way to investigate your own revision process. It does not diagnose the cause of an individual plateau from a dashboard or claim that a particular score predicts an official AKT result.
Establish what the percentage actually measures
A score from repeated questions is not the same observation as a score from unfamiliar questions. A focused specialty block differs from a mixed block, and a timed attempt differs from one completed with unrestricted review. If those conditions have changed, the apparent plateau may not represent unchanged understanding.
Record the relevant context for a small number of recent sessions: whether the material was new, the topic selection, the timing conditions and whether you used explanations during the attempt. You do not need to build a complex analytics system. The purpose is to stop comparing unlike activities as though they were the same test.
An adaptive resource may also alter the difficulty or selection of material. That possibility makes raw percentages harder to interpret; it is not evidence that any particular platform deliberately lowered your score or that every difficult session represents progress.
Check for recognition without reasoning
Return to an item you answered correctly and hide the explanation. Can you state why the selected answer fits the case and why the closest alternative does not? Can you identify which detail would change the decision?
A familiar wording pattern can make the answer feel obvious even when the underlying distinction remains uncertain. This is particularly relevant when a bank has been repeated several times. Use an unfamiliar question or a changed fictional scenario to test whether the principle transfers.
Do not interpret this as an instruction to discard repeated material. Repeated questions can serve a review purpose. The error is treating the resulting percentage as equivalent to a fresh assessment of what you can do without recognition cues.
Look for the misconception that survives explanation reading
Consider a fictional trainee, Oliver, who keeps missing questions about the next step in a clinical pathway. After each attempt, he reads the explanation and feels that it makes sense. In the next unfamiliar case, he again chooses an option associated with the diagnosis rather than the action appropriate at that stage.
Oliver's review has supplied an answer but not changed the decision process. He begins asking what the question requires now, what information is already available and what the proposed action would add. He then explains why the plausible alternative is premature or addresses a different task, where that conclusion is supported by the case and current source.
The useful result to look for is a changed explanation on a new item. A better score on the original item would not establish that the recurring misconception has been resolved.
Audit breadth as well as weak topics
A plateau can conceal different movements across subjects. One area may improve while another remains neglected. A total percentage compresses those differences into a single figure.
The RCGP AKT preparation material, checked on 3 October 2026, links the blueprint and resources for clinical medicine, data interpretation and primary care organisation. Use that scope to check whether your recent activity has become too narrow.
Do not allow your favourite clinical subject to become a reliable source of reassuring scores. Equally, do not spend every session on one difficult area without checking the rest. A useful plan alternates targeted correction with fresh mixed practice so that improvement can be examined in context.
Investigate timing separately
A candidate may understand the material when untimed but lose accuracy when working at the required pace. Inspect where the time goes. Are you rereading because the task is unclear, spending too long between two options, or checking an answer without identifying any new evidence?
The AKT's standard format changed to 160 items in 160 minutes from October 2025, according to the RCGP change notice, checked on 3 October 2026. Use current instructions and any approved adjustments when designing practice conditions, rather than an outdated mock specification.
Do not respond to a timing problem by rushing every learning session. First clarify the process that is consuming time, then test a more workable approach under suitable conditions. Concept learning and pacing practice have different purposes.
Change one part of the review process
Choose a specific hypothesis about the plateau. For example, explanations are being read without an independent account of the reasoning. Change that step: commit to an answer, state why, inspect the correction and then explain the key distinction without looking.
Keep other aspects reasonably similar for a short review period so that you can understand what changed. If you simultaneously replace the bank, alter the topic mix and increase time pressure, the next percentage will be difficult to interpret.
This is a personal learning experiment, not a controlled trial. Treat the observations as useful feedback rather than claiming that the chosen change has a measured causal effect. Where the pattern remains unclear, discuss representative errors with a trainer who can challenge your interpretation.
Use guided questioning for the unresolved step
As described in iatroX's October 2026 product information, its Socratic Tutor opens from an attempted question and asks targeted follow-ups to explore the misconception. Its bank workflow also includes adaptive sequencing and spaced repetition. These features can support a more active review process, but they do not guarantee that a plateau will disappear.
A useful request is to examine why your chosen alternative was attractive before showing the full explanation. Ask which detail changes the decision and then attempt a related question independently. Check clinically important source claims rather than assuming that a coherent teaching conversation is necessarily correct.
Free question access is available without a trial expiry or verification gate under iatroX's product information checked on 3 October 2026. Ongoing tutoring and full bank access belong to the paid bundle. Test whether the workflow addresses your specific review problem before adding another resource.
Decide whether a new resource would add information
A second bank may be useful when you need genuinely unfamiliar material, a different explanation of a persistent concept or a function your existing setup lacks. It is less useful when it simply offers another percentage without changing how errors are reviewed.
Ask what result would change your plan. If a new mixed session reveals the same misconception, the priority remains concept repair. If the difficulty is concentrated in neglected topics, the next step is coverage. If performance differs mainly with timing, examine the process under timed conditions.
A plateau is a prompt to investigate, not a verdict on ability. The most useful change is the one that improves your account of the next unfamiliar question, not merely the appearance of the dashboard.
Frequently asked questions
Does a flat AKT score mean I am not learning?
Not necessarily: topic mix, familiarity, timing and difficulty may have changed. Inspect comparable sessions and the reasoning behind errors before drawing that conclusion.
Should I reset my bank and start again?
Only if repeating it serves a clear learning purpose, and keep repeated-item results separate from unfamiliar-question performance. A reset alone does not resolve the misconception behind an error.
When is a second bank useful?
It can help when it supplies unfamiliar questions, a needed explanation or a missing review function. It should not replace investigating why the existing process has stalled.