Proposição e validação de ferramenta de avaliação da desnutrição hospitalar, com base no Global Leadership Initiative on Malnutrition: protocolo do Estudo GLIM-BR
Carregando...
Autores
Lopes, Giovanna Guimarães
Piovacari, Silvia Maria Fraga
Pereira, Adriano José
Orientador
Co-orientadores
Citações na Scopus
Tipo de documento
Artigo Científico
Data
2022
Título da Revista
ISSN da Revista
Título do Volume
Resumo
Introdução: O consenso do Global Leadership Initiative on Malnutrition (GLIM) publicou uma proposta de diretrizes para estruturação de uma ferramenta, composta por critérios diagnósticos fenotípicos e etiológicos. Posteriormente, foi publicado o Guia de Validação do GLIM, que visa estimular e direcionar iniciativas de validação desta nova ferramenta. O objetivo deste artigo é compartilhar o protocolo do Estudo GLIM-BR, em andamento, o qual irá propor e validar um novo instrumento nacional de classificação diagnóstica de desnutrição hospitalar baseado no GLIM, utilizando desfechos clínicos. Método: Estudo de validação prospectivo observacional multicêntrico, dividido em 3 fases, realizado em 4 hospitais brasileiros terciários, com pacientes internados e com expectativa de permanência maior que 48h. Modelos preditivos baseados em aprendizado de máquina/inteligência artificial serão utilizados na definição do conjunto ótimo de variáveis e cutoffs capazes de predizer desfechos clínicos, como óbito e tempo de permanência. Resultados: Na fase 1, das 12 variáveis apresentadas e discutidas no painel de opinião de especialistas, 8 tiveram aprovação sem alteração. As demais variáveis foram ajustadas por meio do consenso Grupo GLIM-BR. Todas as variáveis foram escolhidas com base na literatura atual (racional teórico) e utilizando, sempre que possível, outras ferramentas já validadas. A Fase 2 já possui resultados preliminares (subestudo) apresentados em Congresso internacional (ESPEN) e que serão submetidos para publicação em periódico científico internacional nos próximos meses. A Fase 3 está em curso e as variáveis de interesse selecionadas para serem avaliadas pelo modelo preditivo do estudo GLIM-BR, em cada uma das categorias propostas pelo GLIM, estão divulgadas neste artigo, juntamente com detalhes do protocolo de pesquisa em curso. Conclusão: Almeja-se desenvolver uma ferramenta validada para diagnóstico da desnutrição hospitalar, que contorne limitações identificadas em ferramentas de avaliação nutricional vigentes, prática e pronta para uso pela comunidade de nutricionistas nos serviços hospitalares.
Introduction: Global Leadership Initiative on Malnutrition (GLIM) consensus has proposed a developing guideline, based on phenotypic and etiological diagnostic criteria for the identification and classification of malnutrition, which aims universal standardization. Subsequently, a GLIM Validation Guide was published, to foster and drive international validation initiatives based on the published standards. This manuscript brings details about the GLIM-BR study protocol, which is currently in progress, and was planned to propose and validate a new diagnostic classification instrument for hospital malnutrition based on GLIM, using clinical outcomes. Methods: Multicenter prospective observational validation trial, divided into 3 phases, carried out in 4 Brazilian tertiary hospitals, with patients hospitalized for longer than 48 hours. Machine learning/artificial intelligence predictive models will be employed in order to identify an optimal subset of variables and thresholds associated with clinically relevant outcomes (as length of stay and death). Results: In phase 1, 8 out 12 variables presented and discussed in the expert opinion panel were approved without changes. The other variables were adjusted by consensus in the GLIM-BR Group of researchers. Initial variables were chosen based on the current literature and using, whenever possible, other already validated tools. Phase 2 already has partial results published in an international congress (oral presentation at ESPEN) and the detailed results are planned to be submitted as an original manuscript in an international journal in the next few months. Phase 3 is ongoing, and the variables of interest selected to be analyzed by the GLIM- BR predictive model, in each of the categories proposed by GLIM are hereby described. Conclusion: The goal is to develop a validated tool for the diagnosis of hospital malnutrition, which circumvents limitations identified in current nutritional asses sment tools, practical enough and ready for use by the community of nutritionists in hospital services
Introduction: Global Leadership Initiative on Malnutrition (GLIM) consensus has proposed a developing guideline, based on phenotypic and etiological diagnostic criteria for the identification and classification of malnutrition, which aims universal standardization. Subsequently, a GLIM Validation Guide was published, to foster and drive international validation initiatives based on the published standards. This manuscript brings details about the GLIM-BR study protocol, which is currently in progress, and was planned to propose and validate a new diagnostic classification instrument for hospital malnutrition based on GLIM, using clinical outcomes. Methods: Multicenter prospective observational validation trial, divided into 3 phases, carried out in 4 Brazilian tertiary hospitals, with patients hospitalized for longer than 48 hours. Machine learning/artificial intelligence predictive models will be employed in order to identify an optimal subset of variables and thresholds associated with clinically relevant outcomes (as length of stay and death). Results: In phase 1, 8 out 12 variables presented and discussed in the expert opinion panel were approved without changes. The other variables were adjusted by consensus in the GLIM-BR Group of researchers. Initial variables were chosen based on the current literature and using, whenever possible, other already validated tools. Phase 2 already has partial results published in an international congress (oral presentation at ESPEN) and the detailed results are planned to be submitted as an original manuscript in an international journal in the next few months. Phase 3 is ongoing, and the variables of interest selected to be analyzed by the GLIM- BR predictive model, in each of the categories proposed by GLIM are hereby described. Conclusion: The goal is to develop a validated tool for the diagnosis of hospital malnutrition, which circumvents limitations identified in current nutritional asses sment tools, practical enough and ready for use by the community of nutritionists in hospital services
Palavras-chave
Titulo de periódico
BRASPEN Journal
Título de Livro
URL na Scopus
Idioma
Inglês
URL permanente
Notas
Membros da banca
Área do Conhecimento CNPQ
CIENCIAS DA SAUDE::NUTRICAO
CIENCIAS DA SAUDE::SAUDE COLETIVA::SAUDE PUBLICA
CIENCIAS DA SAUDE::MEDICINA::CLINICA MEDICA
CIENCIAS DA SAUDE::SAUDE COLETIVA::EPIDEMIOLOGIA
ENGENHARIAS::ENGENHARIA BIOMEDICA
ENGENHARIAS::ENGENHARIA BIOMEDICA::BIOENGENHARIA
CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::PROBABILIDADE E ESTATISTICA APLICADAS
CIENCIAS DA SAUDE::SAUDE COLETIVA::SAUDE PUBLICA
CIENCIAS DA SAUDE::MEDICINA::CLINICA MEDICA
CIENCIAS DA SAUDE::SAUDE COLETIVA::EPIDEMIOLOGIA
ENGENHARIAS::ENGENHARIA BIOMEDICA
ENGENHARIAS::ENGENHARIA BIOMEDICA::BIOENGENHARIA
CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::PROBABILIDADE E ESTATISTICA APLICADAS