Predictive model for acute myocardial infarction in working-age population: a machine learning approach

Abstract

Cardiovascular diseases are the leading cause of mortality in Latin America, particularly acute myocardial infarction (AMI), which is the primary cause of atherosclerotic cardiovascular morbidity. This study aims to develop a predictive model for the probability of AMI occurrence in the working-age population, based on atherogenic indices, paraclinical variables, and anthropometric measures. The research conducted a cross-sectional study involving 427 workers aged 40 years or older in Popayán, Colombia. Out of this population, 202 individuals were screened with a 95% confidence interval and a 5% error margin. Epidemiological, anthropometric, and paraclinical data were collected. A binary logistic regression model was employed to identify variables directly associated with the probability of AMI. Predictive classification models were generated using statistical software JASP and the programming language Python. During the training stage, JASP produced a model with an accuracy of 87.5%, while Python generated a model with an accuracy of 90.2%. In the validation stage, JASP achieved an accuracy of 93%, and Python reached 95%. These results establish an effective model for predicting the probability of AMI in the working population.