Predictive validity of an in-house pediatric examination towards the NBME Pediatrics Subject examination

Abstract

Hassib Narchi

Analyze the predictive validity of the scores obtained by sixth year medical students in our institution in the in-house summative pediatric examination (IHE) and in each of its individual components at the end the pediatric clerkship in relation to scores obtained in the National Board for Medical Examiners Pediatrics Subject Examination (NBME-P). In a cohort of 152 students over five academic years (August 2004 - July 2009) we analyzed the correlation and the predictive validity of the former (and of each of its individual assessment tools) for the latter. In the univariate model, the overall IHE scores (p<0.001), the oral examination (p<0.001), the non-clinical skills (p<0.001), the academic year (p=0.01) and the clinical examination (p=0.04) were significantly correlated with and had a positive linear relationship with the NBME scores. In the multivariable analysis, only the oral examination remained significantly correlated with the NBME-P scores (p<0.001). The predicted NBME scores can be calculated from the students’ oral assessment score as follows: Predicted NBME-P scores = 25 marks + (1.4 x oral score). Having the best predictive validity, the oral examination in the IHE can identify which students are at risk of obtaining low NBME-P marks.