Reprodutibilidade de provas sobre anatomia do aparelho locomotor humano em estudantes de educação física

Abstract

Various theoretical and practical assessment methods are used in the teaching of human anatomy; however, there is a scarcity of studies investigating the reproducibility of these evaluations, that is, the extent to which repeated measures in stable subjects, under different conditions, yield similar results. In this context, the present study aimed to test the reproducibility of practical and theoretical examinations on the anatomy of the human locomotor system in Physical Education students. We conducted a cross-sectional observational study with the application of a digital practical exam (slides) and a traditional theoretical exam (multiple choice) on the anatomy of the human locomotor system. The study comprised two evaluation moments (test and retest), separated by seven to eight days. Each session lasted up to 100 minutes: 25 minutes for the practical exam and 75 minutes for the theoretical exam. A total of 102 Physical Education students participated, aged ≥18 years (20.9 ± 3.4 years), who had already completed the locomotor anatomy course. Most students were male (62.7%), dedicated exclusively to their studies (69.6%), came from public high schools (68.6%), attended university five days per week (73.5%), did not participate in extracurricular academic activities (52%), did not receive financial aid (79.4%), and all had attended classes using synthetic anatomical models (100%). In addition, the mean daily study time outside the university was 2.1 ± 1.2 hours. The assessments included 100 questions (50 practical and 50 theoretical), each scored from 0 to 10. The Wilcoxon test compared scores between test and retest for both practical and theoretical exams, as well as for the total score. Relative reproducibility was assessed using ICC3,1; absolute reproducibility with the coefficient of variation (CV%) and the minimal detectable difference (MDD and %MDC); absolute agreement with the Bland–Altman method; and internal consistency with Cronbach’s alpha. All analyses were performed in JASP software, with a significance level of 5%. Participants obtained median scores of 3.0 u.a. and 3.4 u.a. (practical) and 6.5 u.a. and 7.2 u.a. (theoretical) in the test and retest, respectively. The total score was 9.6 u.a. and 10.5 u.a., with a significant increase (Δ = +0.4 u.a.; p < 0.001). ICC3,1 indicated excellent reproducibility for the practical exam (0.963) and the total score (0.926), and good reproducibility for the theoretical exam (0.722). CV% was 11.2% and 10.0% for the practical and theoretical exams, respectively; and MDD was 1.0 u.a. (27.8%) and 2.1 u.a. (31.0%), in that order. A small systematic bias was found for practical, theoretical, and total scores (0.3 u.a., 0.5 u.a., and 0.8 u.a., respectively). Cronbach’s alpha was very high for the practical exam (0.955 in the test; 0.953 in the retest) and high for the theoretical exam (0.812 in the test; 0.868 in the retest). The anatomy examinations of the human locomotor system demonstrated good to excellent reproducibility and high to very high internal consistency, indicating score stability and reliability.