Predictive validity of the UKCAT for medical school undergraduate performance: a national prospective cohort study

Tiffin, Paul, Mwandigha, Lazaro M., Paton, Lewis W., Hesselgreaves, H., Mclachlan, John Charles orcid iconORCID: 0000-0001-5493-2645, Finn, Gabrielle M. and Kasim, Adetayo S. (2016) Predictive validity of the UKCAT for medical school undergraduate performance: a national prospective cohort study. BMC Medicine, 14 (140).

[thumbnail of Version of Record]
Preview
PDF (Version of Record) - Published Version
Available under License Creative Commons Attribution.

1MB

Official URL: https://doi.org/10.1186/s12916-016-0682-7

Abstract

Background

The UK Clinical Aptitude Test (UKCAT) has been shown to have a modest but statistically significant ability to predict aspects of academic performance throughout medical school. Previously, this ability has been shown to be incremental to conventional measures of educational performance for the first year of medical school. This study evaluates whether this predictive ability extends throughout the whole of undergraduate medical study and explores the potential impact of using the test as a selection screening tool.

Methods

This was an observational prospective study, linking UKCAT scores, prior educational attainment and sociodemographic variables with subsequent academic outcomes during the 5 years of UK medical undergraduate training. The participants were 6812 entrants to UK medical schools in 2007–8 using the UKCAT. The main outcome was academic performance at each year of medical school. A receiver operating characteristic (ROC) curve analysis was also conducted, treating the UKCAT as a screening test for a negative academic outcome (failing at least 1 year at first attempt).

Results

All four of the UKCAT scale scores significantly predicted performance in theory- and skills-based exams. After adjustment for prior educational achievement, the UKCAT scale scores remained significantly predictive for most years. Findings from the ROC analysis suggested that, if used as a sole screening test, with the mean applicant UKCAT score as the cut-off, the test could be used to reject candidates at high risk of failing at least 1 year at first attempt. However, the ‘number needed to reject’ value would be high (at 1.18), with roughly one candidate who would have been likely to pass all years at first sitting being rejected for every higher risk candidate potentially declined entry on this basis.

Conclusions

The UKCAT scores demonstrate a statistically significant but modest degree of incremental predictive validity throughout undergraduate training. Whilst the UKCAT could be considered a fairly crude screening tool for future academic performance, it may offer added value when used in conjunction with other selection measures. Future work should focus on the optimum role of such tests within the selection process and the prediction of post-graduate performance.


Repository Staff Only: item control page