Dolor de espalda en estudiantes de entre 12 y 17 añosaproximación multifactorial basada en árboles de decisión

Abstract

Objective The aim of this study was to use a decision tree as a data analysis technique in order to establish combinations of risk factors that may predispose schoolchildren to suffer back pain. Material and methods The study sample consisted of 337 young people between 12 and 17 years of age from a high school in Castellon, Spain. The subjects completed the Adolescent Sedentary Activity Questionnaire and the Kuorinka Nordic questionnaire. The study also included an assessment of the body mass index, body fat percentage, the characteristics of the school backpack, the time of transportation of the backpack, and habitual physical activity (latter by means of accelerometers). These factors were used as input variables in a decision tree, with the objective of classifying adolescents according to whether they had or had not a back pain episode. Results The tree showed an accuracy of 78.39% in student classification. The sensitivity of the classification was 90.77%, while the specificity was 63.21%. Conclusions The decision tree is a useful analysis technique that has allowed finding combinations of risk factors associated with low back pain in adolescents. The findings suggest that practicing physical activity, the body mass index, body fat percentage, time of transportation of the backpack, minutes of sedentary activities, and family history of back pain are factors related to back pain in adolescents.