Abstract

Artificial intelligence is a broad branch of computer science that enables the creation of analysis pathways that mimic human intelligence; it is a set of adaptive tools that can be used to predict outcomes from biological and clinical data. Artificial intelligence models have the potential to improve the efficiency of healthcare by integrating information including adverse drug events. Objective. To apply artificial intelligence techniques to adverse event and medication use information reported for the Colombian population. Methodology. Non-interventional research, with an analytical and retrospective component. The methods included data science mediated by artificial intelligence techniques. Study population: patients with national adverse event reporting between 2017 and 2019 available in the Colombia open data platform. Results. Female sex had a representation of 59.65% of the adverse events reported. The largest number of patients presented an outcome or exit type "recovered resolved" in 40.4%. The most common route of administration was the oral route (27.85%). The artificial intelligence algorithm allowed predictions close to 90% regarding recovery from an adverse reaction. Conclusion. The inclusion of artificial intelligence for the analysis of the variables: gender, adverse reaction recovery, route of administration, mechanism of the reaction, types of ADR and age of onset, allows the creation of a predictive tool that anticipates the presentation of possible outcomes related to the prescription of a drug.