Performance of doctors with specific learning difficulties in the UK GP licensing Applied Knowledge Test data interpretation questions

Problem

Specific Learning Difficulties (SpLDs) affect the way information is learned and processed and can have a significant impact on learning including reading and writing as well as data interpretation. Substantial numbers of students who have a SpLD enrol in medical and GP training, but they can be disadvantaged in certain assessments, particularly if reasonable adjustments for their needs are not provided. The aim of this study was to evaluate and compare performance of SpLD candidates and candidates with no declared SpLD on all four different categories of data interpretation questions: text only, table, chart, and calculation.

Approach

We used a retrospective study design examining data comprising candidates’ results from the UK GP licensing Applied Knowledge Test (AKT). Ten percent of questions in the AKT are on evidence-based practice including data interpretation and the critical appraisal skills needed to interpret research data. Data interpretation questions were selected and divided into the four categories (text only, table, chart, calculation). Proportions of correct responses were calculated for each category and each candidate. Descriptive statistics of distributions and comparison tests were run between candidates declaring a SpLD and those without. Multivariable predictive models were run for each data interpretation category to establish if having a declared SpLD was a predictor of performance after accounting for ethnicity and country of primary medical qualification (PMQ).

Findings

Out of a total of 18,925 candidates who undertook the AKT23 to AKT44, 920 candidates had a declared learning difficulty representing 4.9% of the sample. On average, candidates with a SpLD performed significantly better on chart-type questions when compared to the other three data interpretation question categories (p<0.001). They also performed significantly better on chart-type questions compared to candidates without a SpLD (p=0.001). Having a SpLD was no longer a significant predictor of better performance on chart-type questions (B= 0.00, 95%CI -0.01, 0.01, p=0.310) after accounting for ethnicity and country of PMQ. On average, candidates with a SpLD had a similar performance on data interpretation type questions when compared to candidates without a SpLD.

Consequences

Candidates with a declared SpLD tended to perform better on chart-type data interpretation questions but this difference was no longer valid when considering other demographic factors. The tendency to perform better on this question category may be linked to the use of graphic organisers and visual displays as strategies to improve reading skills in those with a SpLD and reliance on graphic schemes to improve readability. Overall, these results indicate that different data interpretation question types do not affect performance of candidates with SpLD on the AKT.