Concepção de um sistema de apoio à decisão para acompanhamento nutricional de pacientes submetidos a cirurgia bariátrica

AUTOR(ES)

Magda Rosa Ramos Quadros

DATA DE PUBLICAÇÃO

2006

RESUMO

Nowadays, the morbid obesity is considered a public health concern and the bariatric surgery is the chosen method to treat this illness. It is known that nutritional deficiencies can occur in the post-operative, such as: anemia, lack of B12 and B1 vitamins, folic acid and malnutrition. This dissertation aims to propose a support system for the for the decision-making in the health area, that helps the professional to diagnose the main nutritional deficiencies in the bariatric surgery post-operative. Bayesians Networks were used to implement this system and the database training was made at shell Netica. A golden-pattern was used as reference to evaluate the results, wich was elaborated by a team of four specialists in the nutrition area. Using the golden-pattern, the accurate rate was calculated, which was found among the specialists varying from 80 to 93,33%, according to the nutritional diagnostic and the accurate rate of the system related to the golden-pattern, that was 100% for all the nutritional diagnoses, showing a similar performance to the golden-pattern. The sensitivity and the system specificity were also calculated in order to detect the risk or presence of each nutritional diagnoses and the ROC curve was projected by them for each of them. The sensitivity and specificity found were 95% and the ROC curves projected showed that the system has good performance, i. e., it is able to represent and help the specialists in the daily tasks. The use of this system being viable to support the decision during the nutritional follow-up of patients submitted to the bariatric surgery was concluded. Bayesians Networks may also be considered as efficient tools to reproduce the specialists knowledge when there are several data, wich represent different values, involving the diagnostic definition.

ASSUNTO(S)

engenharia medica inteligencia artificial - aplicações médicas teoria bayesiana de decisão estatística obesidade - cirurgia

Documentos Relacionados

• Acompanhamento psicológico tardio em pacientes submetidos à cirurgia bariátrica
• Estado nutricional e qualidade de vida em pacientes submetidos à cirurgia bariátrica
• Características psicológicas de pacientes submetidos a cirurgia bariátrica
• Concepção, desenvolvimento e validação de um sistema de apoio à decisão comercial - SADEC
• Repercussões nutricionais em pacientes submetidos à cirurgia bariátrica