The method used to set the pass mark in an objective structured clinical examination defines the performance of candidates for certification as rheumatologists.

Background: The Mexican Accreditation Council for Rheumatology certifies trainees (TR) on an annual basis using both a multiple-choice question (MCQ) test and an objective structured clinical examination (OSCE). For 2013 and 2014, the OSCE pass mark (PM) was set by criterion referencing as >= 6 (CPM), whereas overall rating of borderline performance method (BPM) was added for 2015 and 2016 accreditations. We compared OSCE TR performance according to CPM and BPM, and examined whether correlations between MCQ and OSCE were affected by PM.& para;& para;Methods: Forty-three (2015) and 37 (2016) candidates underwent both tests. Altogether, OSCE were integrated by 15 validated stations; one evaluator per station scored TR performance according to a station-tailored check-list and a Likert scale (fail, borderline, above range) of overall performance. A composite OSCE score was derived for each candidate. Appropriate statistics were used.& para;& para;Results: Mean (=standard derivation [SD]) MCQ test scores were 6.6 +/- 0.6 (2015) and 6.4 +/- 0.6 (2016) with 5 candidates receiving a failing score each year. Mean (+/- SD) OSCE scores were 7.4 +/- 0.6 (2015) and 7.3 +/- 0.6 (2016); no candidate received a failing CPM score in either 2015 or 2016 OSCE, although 21 (49%) and 19 (51%) TR, respectively, received a failing BPM score (calculated as 7.3 and 7.4, respectively). Stations for BPM ranged from 4.5 to 9.5; overall, candidates showed better performance in CPM.& para;& para;In all, MCQ correlated with composite OSCE, r = 0.67 (2015) and r = 0.53 (2016); P <=.001. Trainees with a passing BPM score in OSCE had higher MCQ scores than those with a failing score.& para;& para;Conclusions: Overall, OSCE-PM selection impacted candidates' performance but had a limited affect on correlation between clinical and practical examinations. (C) 2016 Elsevier Espafia, S.L.U. and Sociedad Espanola de Reumatologia y Colegio Mexicano de Reumatologia. All rights reserved.