• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Doctoral Thesis
DOI
10.11606/T.5.2012.tde-11072012-113541
Document
Author
Full name
Omar Asdrubal Vilca Mejia
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Paulo, 2012
Supervisor
Committee
Stolf, Noedir Antonio Groppo (President)
Antunes, Manuel de Jesus
Braile, Domingo Marcolino
Dallan, Luis Alberto Oliveira
Néto, Vicente Avila
Title in Portuguese
Predição de mortalidade em cirurgia de coronária e/ou valva no InCor: validação de dois modelos externos e comparação com o modelo desenvolvido localmente (InsCor)
Keywords in Portuguese
Estudos de validação
Fatores de risco
Medição de risco
Modelos probabilísticos
Mortalidade hospitalar
Ponte de artéria coronária
Procedimentos cirúrgicos cardiovasculares
Abstract in Portuguese
Objetivo: Novas tendências na avaliação de risco trazem evidências de que modelos externos recalibrados ou remodelados funcionam melhor localmente. O objetivo deste estudo foi validar dois modelos externos e formular um modelo local, comparando-os na predição de mortalidade nos pacientes operados de coronária e/ou valva no InCor-HCFMUSP. Método: Entre 2007 e 2009, 3.000 pacientes foram sequencialmente operados de coronária e/ou valva no InCor-HCFMUSP. No banco de dados, foi realizada a validação dos modelos 2000 Bernstein-Parsonnet (2000BP) e EuroSCORE (ES), mediante testes de calibração e discriminação. O InsCor de 2.000 pacientes foi elaborado mediante a utilização de técnicas de bootstrap. Nos próximos 1.000 pacientes foi realizada a validação interna do modelo, e seu desempenho medido frente ao 2000BP e ES. Resultados: Houve uma diferença significativa na prevalência dos fatores de risco entre as populações do estudo, ES e 2000BP. Na validação externa dos modelos, o ES apresentou uma boa calibração (P=0,596); no entanto, o 2000BP revelou-se inadequado (P=0,047). Na discriminação, a área abaixo da curva ROC revelou-se boa para ambos os modelos, ES (0,79) e 2000BP (0,80). Utilizando a técnica de bootstrap, 10 variáveis: idade >70 anos, sexo feminino, cirurgia associada, infarto do miocárdio <90 dias, reoperação, cirurgia da valva aórtica, cirurgia da valva tricúspide, creatinina <2mg/dl, fração de ejeção <30% e estado pré-operátorio crítico (eventos), foram selecionadas para formulacão do InsCor. Na validação interna do InsCor, a calibração foi adequada, com P=0,184. Na discriminação, a área abaixo da curva ROC foi boa (0,79). Neste grupo, a área abaixo da curva ROC foi de 0,81 e 0,82 para o ES e 2000BP, respectivamente, mostrando-se apropriada para ambos os modelos. Conclusões: O InsCor e o ES tiveram melhor desempenho que o 2000BP em todas as fases da validação; pórem o novo modelo, além de se identificar com os fatores de risco locais, é mais simples e objetivo para a predição de mortalidade nos pacientes operados de coronária e/ou valva no InCor-HCFMUSP
Title in English
Mortality prediction in coronary bypass surgery and/or heart valve surgery at InCor: Validation of two external risk models and comparison to the locally developed model (InsCor)
Keywords in English
Cardiovascular surgical procedures
Coronary artery bypass
Hospital mortality
Models statistical
Risk assessment
Risk factors
Validation studies
Abstract in English
Background: New trends in risk assessment bring evidence that recalibrated or remodeled external models work best locally. The aim of this study was to validate two external models and formulate a local model, comparing them to predict mortality in patients who underwent coronary bypass surgery and/or heart valve surgery at InCor-HCFMUSP. Method Between 2007 and 2009, 3.000 patients were sequentially operated to coronary bypass surgery and/or heart valve surgery at InCor-HCFMUSP. The database was assessment to validate the models 2000 Bernstein-Parsonnet (2000BP) and EuroSCORE (ES) through calibration and discrimination tests. The InsCor of 2,000 patients (2/3 of database) was elaborated using bootstrap techniques. Over the next 1000 patients (1/3 of database) the internal validation of the InsCor was performed and its performance compared against the 2000BP and ES. Results: Significant difference in the prevalence of risk factors was found among the external and study populations (P<0,001). In the external validation of these models, the ES showed good calibration (P = 0.596); however, 2000BP was inadequate (P = 0.047). In discrimination, the area under the ROC curve was good for both models, ES (0.79) and 2000BP (0.80). With the bootstrap technique, 10 variables: age> 70 years, female, CABG + valve surgery, myocardial infarction <90 days, reoperation, aortic valve surgery, tricuspid valve surgery, creatinine <2mg/dl, ejection fraction <30% and critical preoperative state (events) were chosen to formulate the InsCor. In the validation of InsCor, the calibration was appropriate with P = 0.184. In discrimination, the area under the ROC curve was good (0.79). In this group, the area under the ROC curve was 0.81 and 0.82 for ES and 2000BP, respectively, being suitable for both models. Conclusions: The InsCor and ES outperformed the 2000BP at all stages of validation, but the new model, besides identifying itself with the local risk factors, is more simple and objective for the prediction of mortality in patients who underwent coronary bypass surgery and/or heart valve surgery at InCor-HCFMUSP.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2012-07-12
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
Centro de Informática de São Carlos
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2021. All rights reserved.