A carregar...
Publicação
Estimação da duração de cirurgias para suporte ao planeamento cirúrgico em hospitais públicos
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.03 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O sector da saúde é um dos pilares do desenvolvimento das sociedades, pelo que assume extrema importância na vida da população, proporcionando um aumento constante da esperança média de vida e do nível de bem-estar. Contudo, o ambiente económico que se viveu nos últimos anos tem levado a sucessivos cortes orçamentais no sector da saúde, pelo que se tornou cada vez mais necessária uma melhor racionalização de recursos e uma maior eficiência no uso dos meios disponíveis. Neste sentido, a procura por melhores planos cirúrgicos tem vindo a ser muito desenvolvida, contribuindo para isto os estudos desenvolvidos na área da Investigação Operacional. No entanto, existe um factor que não tem sido tão abordado ao longo destes estudos, que diz respeito à consideração da estocasticidade da duração das cirurgias aquando da realização dos planos cirúrgicos. Deste modo, este trabalho surge no âmbito de um projecto em parceria com um hospital público e foca-se no estudo da incerteza presente nestas durações para a obtenção de planos cirúrgicos mais robustos, desenvolvendo, para isso, um estudo realizado anteriormente por Catarina Mateus sobre planeamento cirúrgico, no âmbito da sua dissertação do mestrado em Estatística e Investigação Operacional, que é a base de trabalho desta dissertação. É apresentado um método para a obtenção de estimativas de durações cirúrgicas e de tempos totais dos pacientes em salas de operação, através da regressão de um modelo linear, bem como a estimação de uma função distribuição (Log-Normal), para cada uma das durações mencionadas. Este estudo é realizado para a dimensão do problema considerada no trabalho referido anteriormente, no que respeita ao número de variáveis a considerar. Inserindo estimativas das durações cirúrgicas e dos tempos totais dos pacientes em sala de operação nos modelos estudados no trabalho anterior para a marcação semanal de cirurgias electivas, os resultados mostram que estes modelos (VA – Versão Administração, VC – Versão Cirurgiões e VM – Versão Mista) obtêm, no geral, bons resultados ao nível de cirurgias marcadas, tendo em conta os seus objectivos. Uma comparação entre as estimativas obtidas pelo método aqui adoptado e pelo método utilizado no estudo base, que considera apenas as médias das durações das cirurgias e dos tempos totais dos pacientes em sala de operação de cada código de procedimento cirúrgico, é realizada no final do trabalho. Esta comparação contempla, para cada tipo de duração mencionada, a média e o desvio padrão das diferenças entre as estimativas obtidas por cada um dos métodos e o tempo efectivamente incorrido pelo hospital aquando da realização das cirurgias correspondentes. É também realizada uma análise de robustez ao plano cirúrgico obtido com recurso ao método de estimação de durações desenvolvido nesta dissertação, que inclui também a percentagem de alteração média das estimativas obtidas por cada um dos estudos realizados face aos tempos reais. Através destas análises e comparações com o estudo base, verificam-se alguns casos em que os resultados obtidos são melhores através do método da regressão linear e outros através do método utilizado no trabalho original. Por fim, são dadas algumas orientações de caminhos futuros a percorrer para a obtenção de estimativas melhores e mais precisas para as durações, quer cirúrgicas, quer do tempo total do paciente em sala de operação.
The health sector is one of the cornerstones of the development of societies and is therefore of paramount importance in the lives of the population, providing a steady increase in the average life expectancy and level of well-being. However, the economic environment lived in the last years resulted in successive budget cuts in the health sector, making it increasingly necessary a better rationalization of resources and an improving efficiency in the use of available resources. In this sense, the search for better surgical plans has been very developed, contributing to this, the studies developed in the Operational Research area. However, there is a factor that has been not so much approached during these studies, which refers to the stochasticity of the duration of the surgeries during the realization of the surgical plans. Thus, this work arises from a project in partnership with a public hospital and focuses on the study of the uncertainty present in these durations to obtain more robust surgical plans, developing, for this, a study previously performed by Catarina Mateus on surgical planning, within the scope of her dissertation on the Master's Degree in Statistics and Operational Research, which is the basis of this dissertation. The study presents a method for obtaining estimates of surgical durations and total times of patients in operating rooms, through the regression of a linear model, as well as the estimation of a distribution function (Log-Normal), for each one of mentioned durations. This study is carried out for the dimension of the problem considered in the previously mentioned work, in what concerns to the number of variables to be considered. Inserting estimates of the surgical durations and total time of patients in operating room in the models studied in the previous work for the weekly marking of elective surgeries, the results show that these models (AV – Administration Version, SV – Surgeons Version and MV – Mixed Version) obtain, in general, good results in the level of marked surgeries, taking into account its objectives. One comparasion between the estimates obtained by the method adopted here and by the method used in the basis work, which only consideres the averages of the surgeries’s durations and of the total duration of the pacients in the operating room of each code of cirurgical procedure, is carried out at the end of the work. This compararion contemplates, for each type of mentioned duration, the average and the standard deviation of the differences between the estimates obtained by each one of the methods and the time actually incurred by the hospital when performing the corresponding surgeries. An analysis of robustness to the surgical plane obtained using the duration estimation method developed in this dissertation is also performed, which also includes the percentage of average change in the estimates obtained by each of the studies performed in relation to the actual times. Through these analyzes and comparisons with the base study, there are some cases in which the obtained results are better through the method of linear regression and others through the method used in the original work. Finally, some guidelines are given for future paths to be taken in order to obtain better and more accurate estimates to the durations, whether surgical or total patient time in the operating room.
The health sector is one of the cornerstones of the development of societies and is therefore of paramount importance in the lives of the population, providing a steady increase in the average life expectancy and level of well-being. However, the economic environment lived in the last years resulted in successive budget cuts in the health sector, making it increasingly necessary a better rationalization of resources and an improving efficiency in the use of available resources. In this sense, the search for better surgical plans has been very developed, contributing to this, the studies developed in the Operational Research area. However, there is a factor that has been not so much approached during these studies, which refers to the stochasticity of the duration of the surgeries during the realization of the surgical plans. Thus, this work arises from a project in partnership with a public hospital and focuses on the study of the uncertainty present in these durations to obtain more robust surgical plans, developing, for this, a study previously performed by Catarina Mateus on surgical planning, within the scope of her dissertation on the Master's Degree in Statistics and Operational Research, which is the basis of this dissertation. The study presents a method for obtaining estimates of surgical durations and total times of patients in operating rooms, through the regression of a linear model, as well as the estimation of a distribution function (Log-Normal), for each one of mentioned durations. This study is carried out for the dimension of the problem considered in the previously mentioned work, in what concerns to the number of variables to be considered. Inserting estimates of the surgical durations and total time of patients in operating room in the models studied in the previous work for the weekly marking of elective surgeries, the results show that these models (AV – Administration Version, SV – Surgeons Version and MV – Mixed Version) obtain, in general, good results in the level of marked surgeries, taking into account its objectives. One comparasion between the estimates obtained by the method adopted here and by the method used in the basis work, which only consideres the averages of the surgeries’s durations and of the total duration of the pacients in the operating room of each code of cirurgical procedure, is carried out at the end of the work. This compararion contemplates, for each type of mentioned duration, the average and the standard deviation of the differences between the estimates obtained by each one of the methods and the time actually incurred by the hospital when performing the corresponding surgeries. An analysis of robustness to the surgical plane obtained using the duration estimation method developed in this dissertation is also performed, which also includes the percentage of average change in the estimates obtained by each of the studies performed in relation to the actual times. Through these analyzes and comparisons with the base study, there are some cases in which the obtained results are better through the method of linear regression and others through the method used in the original work. Finally, some guidelines are given for future paths to be taken in order to obtain better and more accurate estimates to the durations, whether surgical or total patient time in the operating room.
Descrição
Tese de mestrado, Estatística e Investigação Operacional (Estatística e Investigação Operacional) Universidade de Lisboa, Faculdade de Ciências, 2018
Palavras-chave
Tempos cirúrgicos Tempos totais dos pacientes em sala de operação Agrupamento de códigos de procedimento Modelo de regressão linear Distribuição Log-Normal Teses de mestrado - 2018