Comparing the <scp>GP</scp> ‐ <scp>CORE</scp> and <scp>WEMWBS</scp> Scores as Predictors of Future Mental Health Service Utilisation and Help‐Seeking Behaviour Among <scp>UK</scp> Students in Higher Education

ABSTRACT Background The General Population Clinical Outcomes in Routine Evaluation (GP‐CORE) and the Warwick‐Edinburgh Mental Well‐being Scale (WEMWBS) are two instruments often used to measure psychological distress or well‐being in student samples. This study aimed to directly compare the ability of the two instruments to yield information on psychological distress in this population. Specifically, the ability of the scores to predict mental health service utilisation was compared. Methods GP‐CORE and WEMWBS scores, sociodemographic and educational data were available for 1260 students at a UK University (Time 1) and for 265 students at 6‐month follow‐up (Time 2). Rasch calibration was performed on baseline data to evaluate test information, with item and person fit assessed using infit/outfit and item characteristic curves. Effect sizes (Cohen's d ) and discriminative ability (area under the curve—AUC) were estimated for predicting future mental health service utilisation. Results Both measures were acceptably reliable (α &gt; 0.8), though the WEMWBS was better at discriminating between respondents, as indicated by the person separation indices (GP‐CORE = 2.34; WEMWBS = 3.06). The GP‐CORE scores were more sensitive to service utilisation status (Cohen's d = 0.31, p = 0.02) compared to those from the WEMWBS (Cohen's d = 0.25, p = 0.06). The GP‐CORE scores were statistically significantly superior at discriminating between students reporting service utilisation vs. those that did not (AUC = 0.59 vs. 0.43, p = 0.02). Conclusion Although the WEMWBS scores may be superior at differentiating between different levels of ‘psychological well‐being’, the GP‐CORE scores are more predictive of service utilisation and may be more appropriate when screening students for potential mental health issues that warrant formal support.