RESUMO
RESUMO Objetivo: Criar e validar um modelo de predição de choque séptico ou hipovolêmico a partir de variáveis de fácil obtenção coletadas na admissão de pacientes internados em uma unidade de terapia intensiva. Métodos: Estudo de modelagem preditiva com dados de coorte concorrente realizada em um hospital do interior do nordeste brasileiro. Foram incluídos pacientes com 18 anos ou mais sem uso de droga vasoativa no dia da admissão e que foram internados entre novembro de 2020 e julho de 2021. Foram testados os algoritmos de classificação do tipo Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost para a construção do modelo. O método de validação utilizado foi o k-fold cross validation. As métricas de avaliação utilizadas foram recall, precisão e área sob a curva Receiver Operating Characteristic. Resultados: Foram utilizados 720 pacientes para criação e validação do modelo. Os modelos apresentaram alta capacidade preditiva com área sob a curva Receiver Operating Characteristic de 0,979; 0,999; 0,980; 0,998 e 1,00 para os algoritmos de Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost, respectivamente. Conclusão: O modelo preditivo criado e validado apresentou elevada capacidade de predição do choque séptico e hipovolêmico desde o momento da admissão de pacientes na unidade de terapia intensiva.
ABSTRACT Objective: To create and validate a model for predicting septic or hypovolemic shock from easily obtainable variables collected from patients at admission to an intensive care unit. Methods: A predictive modeling study with concurrent cohort data was conducted in a hospital in the interior of northeastern Brazil. Patients aged 18 years or older who were not using vasoactive drugs on the day of admission and were hospitalized from November 2020 to July 2021 were included. The Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost classification algorithms were tested for use in building the model. The validation method used was k-fold cross validation. The evaluation metrics used were recall, precision and area under the Receiver Operating Characteristic curve. Results: A total of 720 patients were used to create and validate the model. The models showed high predictive capacity with areas under the Receiver Operating Characteristic curve of 0.979; 0.999; 0.980; 0.998 and 1.00 for the Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost algorithms, respectively. Conclusion: The predictive model created and validated showed a high ability to predict septic and hypovolemic shock from the time of admission of patients to the intensive care unit.
RESUMO
Objectives. This study sought to compare the estimated glomerular filtration rate and the indication of dose adjustment of antimicrobials when using Cockcroft-Gault or Modification of Diet in Renal Disease. Methods. A cross-sectional study was performed with patients admitted to the intensive care unit of a Brazilian general hospital. The glomerular filtration rate was calculated for patients on all days using the Cockcroft-Gault and Modification of Diet in Renal Disease equations. The difference in estimated glomerular filtration and the dose adjustment indication of antimicrobials were assessed. Results. A total of 631 patients were included in this study. The median estimated glomerular filtration was significantly higher when estimated using Modification of Diet in Renal Disease (100.3 mL/ min/1.73 m2) than the estimation by Cockcroft-Gault (83.2 mL/min) [p<0.001]. Greater differences in estimations produced by the two formulae were observed in patients at extremes of weight and age, and a different dose adjustment was indicated for all antimicrobials assessed. Conclusions. These results demonstrate a significant difference in estimated glomerular filtration rate values when calculated using either Cockcroft-Gault or Modification of Diet in Renal Disease as well as in the indication of dose adjustment in an intensive care unit