Medición de la mortalidad de pacientes internados por enfermedad cardiovascular mediante el desarrollo de un algoritmo de inteligencia artificial / Measurement of mortality during hospitalization due cardiovascular disease through the development of an artificial intelligence algorithm

Introducción: determinar la causa de muerte de los pacientes internados con enfermedad cardiovascular es de suma importancia para poder tomar medidas y así mejorar la calidad su atención y prevenir muertes evitables. Objetivos: determinar las principales causas de muerte durante la internación por enfermedades cardiovasculares. Desarrollar y validar un algoritmo para clasificar automáticamente a los pacientes fallecidos durante la internación con enfermedades cardiovasculares Diseño del estudio: estudio exploratorio retrospectivo. Desarrollo de un algoritmo de clasificación. Resultados: del total de 6161 pacientes, el 21,3% (1316) se internaron por causas cardiovasculares; las enfermedades cerebrovasculares representan el 30,7%, la insuficiencia cardíaca el 24,9% y las enfermedades cardíacas isquémicas el 14%. El algoritmo de clasificación según motivo de internación cardiovascular vs. no cardiovascular alcanzó una precisión de 0,9546 (IC 95%: 0,9351-0,9696). El algoritmo de clasificación de causa específica de internación cardiovascular alcanzó una precisión global de 0,9407 (IC 95%: 0,8866-0,9741). Conclusiones: la enfermedad cardiovascular representa el 21,3% de los motivos de internación de pacientes que fallecen durante su desarrollo. Los algoritmos presentaron en general buena performance, particularmente el de clasificación del motivo de internación cardiovascular y no cardiovascular y el clasificador según causa específica de internación cardiovascular. (AU)

Introduction: determining the cause of death of hospitalized patients with cardiovascular disease is of the utmost importance in order to take measures and thus improve the quality of care of these patients and prevent preventable deaths. Objectives: to determine the main causes of death during hospitalization due to cardiovascular diseases.To development and validate a natural language processing algorithm to automatically classify deceased patients according to their cause for hospitalization. Design: retrospective exploratory study. Development of a natural language processing classification algorithm. Results: of the total 6161 patients in our sample who died during hospitalization, 21.3% (1316) were hospitalized due to cardiovascular causes. The stroke represent 30.7%, heart failure 24.9%, and ischemic cardiac disease 14%. The classification algorithm for detecting cardiovascular vs. Non-cardiovascular admission diagnoses yielded an accuracy of 0.9546 (95% CI 0.9351, 0.9696), the algorithm for detecting specific cardiovascular cause of admission resulted in an overall accuracy of 0.9407 (95% CI 0.8866, 0.9741). Conclusions: cardiovascular disease represents 21.3% of the reasons for hospitalization of patients who die during hospital stays. The classification algorithms generally showed good performance, particularly the classification of cardiovascular vs non-cardiovascular cause for admission and the specific cardiovascular admission cause classifier. (AU)

Development and validation of various phenotyping algorithms for Diabetes Mellitus using data from electronic health records.

BACKGROUND AND OBJECTIVE: Recent progression towards precision medicine has encouraged the use of electronic health records (EHRs) as a source for large amounts of data, which is required for studying the effect of treatments or risk factors in more specific subpopulations. Phenotyping algorithms allow to automatically classify patients according to their particular electronic phenotype thus facilitating the setup of retrospective cohorts. Our objective is to compare the performance of different classification strategies (only using standardized problems, rule-based algorithms, statistical learning algorithms (six learners) and stacked generalization (five versions)), for the categorization of patients according to their diabetic status (diabetics, not diabetics and inconclusive; Diabetes of any type) using information extracted from EHRs. METHODS: Patient information was extracted from the EHR at Hospital Italiano de Buenos Aires, Buenos Aires, Argentina. For the derivation and validation datasets, two probabilistic samples of patients from different years (2005: n = 1663; 2015: n = 800) were extracted. The only inclusion criterion was age (≥40 & <80 years). Four researchers manually reviewed all records and classified patients according to their diabetic status (diabetic: diabetes registered as a health problem or fulfilling the ADA criteria; non-diabetic: not fulfilling the ADA criteria and having at least one fasting glycemia below 126 mg/dL; inconclusive: no data regarding their diabetic status or only one abnormal value). The best performing algorithms within each strategy were tested on the validation set. RESULTS: The standardized codes algorithm achieved a Kappa coefficient value of 0.59 (95% CI 0.49, 0.59) in the validation set. The Boolean logic algorithm reached 0.82 (95% CI 0.76, 0.88). A slightly higher value was achieved by the Feedforward Neural Network (0.9, 95% CI 0.85, 0.94). The best performing learner was the stacked generalization meta-learner that reached a Kappa coefficient value of 0.95 (95% CI 0.91, 0.98). CONCLUSIONS: The stacked generalization strategy and the feedforward neural network showed the best classification metrics in the validation set. The implementation of these algorithms enables the exploitation of the data of thousands of patients accurately.