Introduction:
Borderline regression and contrasting groups are methods used to set passing scores in the group of Examinee-centered methods according to absolute standards. At the Center for Elaboration Competency and Innovation in Clinical Simulation – Pham Ngoc Thach University of Medicine (CECICS-PTNU), we implemented these methods for the formative assessment of the Observed Structured Clinical Examination (OSCE) station for metered dose inhaler (MDI) use instruction for third-year medical students of the Vietnam-Germany program (MS3VG).

Background:
Borderline regression and contrasting group methods were based on observational assessments of students’ performance during OSCE. The results were classified into different levels, such as 3 levels (Borderline, Pass and Fail) or 5 levels (Fail, Borderline Fail, Borderline Pass, Pass and High Pass). Borderline regression method used both the checklist score and the global rating score of all students to determine the passing score. For the contrasting groups method, the assessors classified students as either “Pass” or “Fail” after observing their overall performances. This method was based on the student’s actual performance rather than the checklist score. The intersection of two graphs by the checklist score distributions represented the cut score.

Methods:
Before conducting the assessment, faculty members at CECICS-PNTU were selected as either observers or assessors and received training on these two assessment methods. The Learning Space system included two circuits for 22 MS3VG, with each circuit having six OSCEs. OSCE MDI was one station among those six. For each running circuit, one faculty member provided checklist scoring as well as global rating scoring. The checklist score was previously performed on Content Validity Index and Content Validity Ratio (CVI-CVR) with 15 items. The maximum score of the checklist was 30 points. The results of 22 MS3VG were analyzed using the borderline regression and contrasting group methods.

Results:
Based on the borderline regression method, the results of 22 MS3VG included checklist scores (ordinate) and global rating scores (abscissa), with a running regression equation of Y = 4.1667 X + 12.636, R2 = 0.618. When X = 2 (equivalent to borderline pass), the pass score was 20.97. Therefore, 2 students (9.1%) failed and 20 students passed (90.9%).
For the contrasting groups method, the smoothed graph also showed that the pass score was 21. In contrast, the traditional relative standard setting method showed that 100% of the students passed.

Conclusion:
Through extensive training for faculty members as assessors, our results showed that the pass score setting was consistent between borderline regression method (20.97/30 points) and contrasting method (21/30 points). The Fail rate of two methods was 9.1% and the Pass rate was 90.9%. Borderline regression and Contrasting methods are more reliable and legally valued compared to the traditional relative standard setting due to their focus on the actual performance and ability of students. Therefore, these two methods are commonly used in high stakes examinations as well as important exams in medical education and healthcare to enhance the quality of teaching, assessment and patient safety.

Keywords: Standard setting, pass score, Borderline regression, Contrasting group, MDI, OSCE, assessment, medical education, clinical simulation.