Closing the Confidence Gap: Piloting an In-House Revision Programme to Improve UKMLA Preparedness, Confidence, and Question Comprehension

Technical Summary: Closing the Confidence Gap in UKMLA Preparation

Problem Statement
The introduction of the United Kingdom Medical Licensing Assessment (UKMLA) Applied Knowledge Test (AKT) in 2024 established a new high-stakes national threshold for provisional General Medical Council (GMC) registration. Unlike previous systems where individual medical schools managed final examinations, the UKMLA requires all graduates to demonstrate a standardized level of core knowledge and clinical reasoning. This shift presents significant challenges for final-year students, who must synthesize years of learning into a specific question format (single best answer) mapped to the MLA Content Map.

A critical gap exists in institutional support: formal university-led revision for the AKT is limited, forcing students to rely heavily on commercial question banks. This reliance raises concerns regarding equity, as the cost of commercial resources may create disparities among students from different socioeconomic backgrounds. Furthermore, while evidence supports active learning strategies like retrieval practice and spaced repetition, students often lack awareness of these principles or the guidance to apply them effectively to the new exam format.

Methodology
This study was a single-centre service evaluation conducted at a major teaching hospital in the North West of England involving 156 eligible final-year medical students. The intervention was a two-session, institution-led revision programme designed to be low-resource and accessible.

• Session 1 (Online): A 40-minute structured session delivered via Microsoft Teams by a recent graduate who had passed the pilot UKMLA AKT. Content covered exam structure, question types, content distribution, and revision strategies.
• Session 2 (In-Person): A 75-minute interactive session held one month later, utilizing a fast-paced, question-based format with curated AKT-style multiple-choice questions (MCQs) and official practice paper samples.

Data Collection and Analysis:

• Instrumentation: Pre- and post-session questionnaires were administered for Session 1, and a post-session questionnaire for Session 2. Instruments were developed specifically for this evaluation using a five-point Likert scale (1 = 'strongly disagree' to 5 = 'strongly agree').
• Metrics: For Session 1, four domains were assessed: (1) self-reported preparedness, (2) understanding of exam structure, (3) confidence in approaching AKT-style MCQs, and (4) confidence in the appropriateness of revision resources. Session 2 assessed enjoyment, relevance, and interactivity.
• Statistical Analysis: Pre- and post-session distributions for Session 1 were compared using the Mann-Whitney U test (for ordinal data) and Fisher's exact test (for dichotomized "agree/strongly agree" responses). Free-text responses were analyzed descriptively.

Key Results

• Session 1 Outcomes:

  • Response Rates: 27 pre-session and 18 post-session questionnaires were completed (17.3% and 11.5% response rates, respectively).
  • Preparedness: The proportion of students agreeing they felt prepared increased from 7.4% to 33.3% (+25.9 percentage points; p = 0.045).
  • Exam Structure Understanding: Agreement rose from 29.6% to 83.3% (+53.7 percentage points; p < 0.001).
  • MCQ Confidence: Agreement increased from 33.3% to 61.1% (+27.8 percentage points; p = 0.022).
  • Resource Confidence: The most significant shift occurred here, with agreement rising from 18.5% to 94.4% (+75.9 percentage points; p < 0.001). Uncertainty regarding resources dropped from 66.7% to 5.6%.
  • Follow-up Intent: 88.9% of respondents intended to attend the follow-up session.

• Session 2 Outcomes:

  • Attendance: 50 students attended (32.0% of the eligible cohort).
  • Feedback: Of 42 respondents, 97.6% agreed or strongly agreed that the session was enjoyable, relevant to the AKT, and fun/interactive.
  • Qualitative Feedback: Students valued the interactive format and immediate explanations. Constructive suggestions included replacing hand-raising with digital polling tools to ensure inclusive participation and slowing the pace to accommodate students with learning differences (e.g., dyslexia).

Key Contributions

1. Institutional Intervention Model: The paper demonstrates that a brief, low-resource, institution-led revision series can produce measurable improvements in self-reported student outcomes.
2. Resource Confidence: The study highlights that uncertainty regarding revision resources is a major source of student anxiety. A structured programme signposting evidence-based resources can address this disproportionately large stressor.
3. Test-Format Knowledge: The results suggest that understanding the exam structure is a distinct, malleable domain that can be significantly improved through a single educational session, separate from clinical knowledge mastery.
4. Equity Implications: By providing free, structured guidance, institutions can potentially reduce reliance on costly commercial question banks, thereby addressing potential barriers to equitable performance.

Significance and Claims
The authors position this work as a pilot evaluation rather than a definitive proof of efficacy. The paper claims that:

• A structured, recent-graduate-led intervention can improve students' self-reported preparedness, understanding of exam structure, and confidence in resources.
• The findings warrant multi-centre evaluation to determine if these self-reported shifts translate into actual performance improvements (Level 4 of Kirkpatrick's hierarchy).
• The model offers a sustainable, near-peer support mechanism that could be embedded into the final-year curriculum to address the informational gap left by the transition to the UKMLA.

The authors explicitly note limitations, including small sample sizes, low response rates (risk of self-selection bias), the inability to track individual-level changes due to anonymous data collection, and the reliance on self-reported confidence rather than actual exam scores. Consequently, the paper concludes modestly, suggesting the model is promising but requires further validation in larger studies linking intervention exposure to actual UKMLA AKT performance.