Modelo predictivo de riesgo académico en estudiantes de medicina

Background: Academic performance is an important issue for universities. Andrés Bello University's academic risk factors identification system focuses on performance variables, such as scores on university entrance examinations and scores in high school. Since medical students have high entrance scores, their academic risk is not detected. Aim: To create an academic risk predictive model for medical students using neuro-didactic variables. Material and Methods: Prospective cohort study with 189 first-year medical students. After signing the informed consent, questionnaires were applied to measure variables related to executive functions and reward system. These were analyzed with logistic regression models and classification trees. Results: The variables that modified the probability of academic risk were identified by the logistic regression model with a global accuracy of 0.74 and by the algorithm of the classification tree, with a sensitivity and specificity of 68 and 79% respectively. The identified variables for executive functions, were the scores obtained in the science, mathematics, and emotional intelligence tests of the university entrance examination. Family pressure and who they live with, were variables associated with the reward system. Conclusions: A low score in the science test of the university entrance examination is the variable that initially determines the probability of academic risk. It is modulated by variables related to reward system. In the most complex branch, the terminal node is represented by the emotional intelligence variable of executive functions.

[A predictive model for academic risk in medicine students]. / Modelo predictivo de riesgo académico en estudiantes de medicina.

BACKGROUND: Academic performance is an important issue for universities. Andrés Bello University's academic risk factors identification system focuses on performance variables, such as scores on university entrance examinations and scores in high school. Since medical students have high entrance scores, their academic risk is not detected. AIM: To create an academic risk predictive model for medical students using neuro-didactic variables. MATERIAL AND METHODS: Prospective cohort study with 189 first-year medical students. After signing the informed consent, questionnaires were applied to measure variables related to executive functions and reward system. These were analyzed with logistic regression models and classification trees. RESULTS: The variables that modified the probability of academic risk were identified by the logistic regression model with a global accuracy of 0.74 and by the algorithm of the classification tree, with a sensitivity and specificity of 68 and 79% respectively. The identified variables for executive functions, were the scores obtained in the science, mathematics, and emotional intelligence tests of the university entrance examination. Family pressure and who they live with, were variables associated with the reward system. CONCLUSIONS: A low score in the science test of the university entrance examination is the variable that initially determines the probability of academic risk. It is modulated by variables related to reward system. In the most complex branch, the terminal node is represented by the emotional intelligence variable of executive functions.