MULTIPLE-CHOICE QUESTION EXAMINATIONS IN PRE-CLINICAL COURSES AS POTENTIAL ENDPOINTS FOR ANALYSIS OF QUALITY OF MEDICAL SCHOOL ADMISSION CRITERIA

Introduction:
Various forms of admission tests are used in most western countries to select the best candidates for medical schools. Predictive ability of selected examination criteria with reference to student's results obtained in pre-clinical and clinical courses constitutes an important factor that determines good selection of candidates.

Aim of study:
Assessment of the quality of final tests in physiology and pathophysiology courses as well as retrospective analysis of the medical school admission criteria with reference to the results of pre-clinical education.

Materials and Methods
862 full-time and part-time Warsaw Medical University students, including 416 students admitted in the academic year 2008/09 and 446 students admitted in the academic year 2010/11 to the two medical faculties. The results of the biology, chemistry, and physics exams were analysed to assess the admission criteria. The results of two physiology exams and four pathophysiology exams were used to analyse the quality of multiple-choice question (MCQ) examinations in pre-clinical courses.

The quality of the exams was assessed by establishing the level of simplicity and differentiating power of particular questions and the reliability of the exam was assessed by estimating Cronbach's α coefficient. Moreover, the analysis involved the correlation between the results of physiology and pathophysiology exams as well as between particular results of these exams and scores obtained by a candidate during the admission procedure. The Pearson product-moment correlation coefficient was used to assess the correlation.
The significance level for all analyses was established at p<0.05.

Results:
A detailed analysis of the quality of physiology and pathophysiology exams demonstrated high reliability of all MCQ tests (α > 0.08). Furthermore, neither the number of distractors, nor the number of test versions significantly influenced the quality of these tests. Optimisation of examination tools consisting in elimination of questions with a negative correlation improved the quality of tests (on average by +0.02). Difficulty of particular tests was diverse and fell within 0.55 and 0.80. A correlation analysis showed a statistically significant relation between the physiology exam results and pathophysiology exam results (r value between 0.521 and 0.624, p < 0.001). The total reliability of pre-clinical exams was also high and fell within α = 0.872 – 0.922.

Assessment of the selective ability of the adopted admission criteria demonstrated that, while in the case of students admitted to the full-time programme the total score obtained in the three key courses (biology, chemistry and physics) had a positive correlation with the results of pre-clinical exams (p<0.01), there was no such correlation in the case of part-time students. Moreover, an analysis of individual criteria showed that chemistry was the best independent selection criterion, whereas biology and physics indicated low or no correlation.

Conclusions:
The accuracy of admission of appropriate candidates for medical school should be evaluated on several levels during the course of studies as well as when monitoring a medical career of a graduate. There is a need to strive for a proper admission procedure that would allow for selection of the best candidates for prospective doctors who would efficiently learn complex pre-clinical and clinical skills.