Natural Language Processing to Extract Information from Portuguese-Language Medical Records

Studies that use medical records are often impeded due to the information presented in narrative fields. However, recent studies have used artificial intelligence extract and process secondary health data from electronic records. The aim of this study was develop a neural network uses unstructured capture regarding symptoms, diagnoses, medications, conditions, exams, treatment. Data 30,000 patients hospitalized Clinical Hospital Botucatu Medical School (HCFMB), São Paulo, Brazil, were obtained, creating corpus with 1200 clinical texts. A natural language algorithm for text extraction convolutional networks pattern recognition evaluate model goodness-of-fit indices. results showed good accuracy, considering complexity model, an F-score 63.9% precision 72.7%. patient condition class reached 90.3% medication 87.5%. proposed will facilitate detection relationships between diseases symptoms prevalence incidence, addition detecting identification disease evolution, effects prescribed medications.