Predictive validity of a selection centre testing non-technical skills for recruitment to training in anaesthesia

Methods. Expert consensus identified non-technical skills suitable for assessment at the point of selection. Four stations—structured interview, portfolio review, presentation, and simulation—were developed, the latter two being realistic scenarios of work-related tasks. Evaluation of the selection centre focused on applicant and assessor feedback ratings, inter-rater agreement, and internal consistency reliability coefficients. Predictive validity was sought via correlations of selection centre scores with subsequent workplace-based ratings of appointed trainees. Results. Two hundred and twenty-four candidates were assessed over two consecutive annual recruitment rounds; 68 were appointed and followed up during training. Candidates and assessors demonstrated strong approval of the selection centre with more than 70% of ratings ‘good’ or ‘excellent’. Mean inter-rater agreement coefficients ranged from 0.62 to 0.77 and internal consistency reliability of the selection centre score was high (Cronbach’s a¼0.88 –0.91). The overall selection centre score was a good predictor of workplace performance during the first year of appointment. Conclusions. An assessment centre model based on the rating of non-technical skills can produce a reliable and valid selection tool for recruitment to speciality training in anaesthesia. Early results on predictive validity are encouraging and justify further development and evaluation.