Application of artificial neural networks in assigned leadership and academic success in medical graduates

Abstract

Objective: To apply an artificial neural networks analysis (ANN) model to identify variables that predict assigned leadership and academic success in graduates of six generations of medical school. Method: Analytical, retrospective, comparative study. A total of 1434 graduates participated. A questionnaire was sent to them by e-mail including a voluntary participation consent. A multivariate statistical analysis using multi-layer perceptron ANN, decision trees and driver analysis was performed. Results: The ANN identified seven independent variables that predicted professional success and eight for leadership in medical graduates. The decision trees identified significant differences in the variables professional performance (p = 0.000), age (p = 0.005) and continuing education activities (p = 0.034) related to professional success, and for leadership the variables gender (p = 0.000), high school grades (p = 0.042), performing clinical practice during the social service year (p = 0.002) and continuing education activities (p = 0.011). Conclusions: The ANN identified the main independent predictor variables of professional success and leadership of the graduates. This study opens up two new lines of research little studied with the techniques of in the area of medicine.