HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Tipo de documento: | Dissertação |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10362/151158 |
Resumo: | This dissertation is part of the project FrailCare.AI, which aims to detect frailty in the elderly Portuguese population in order to optimize the SNS24 (telemonitoring) service, with the goal of suggesting health pathways to reduce the patients frailty. Frailty can be defined as the condition of being weak and delicate which normally increases with age and is the consequence of several health and non-health related factors. A patient health journey is recorded in Eletronic Health Record (EHR), which are rich but sparse, noisy and multi-modal sources of truth. These can be used to train predictive models to predict future health states, where frailty is just one of them. In this work, due to lack of data access we pivoted our focus to phenotype prediction, that is, predicting diagnosis. What is more, we tackle the problem of data-insufficiency and class imbalance (e.g. rare diseases and other infrequent occurrences in the training data) by integrating standardized healthcare ontologies within graph neural networks. We study the broad task of phenotype prediction, multi-task scenarios and as well few-shot scenarios - which is when a class rarely occurs in the training set. Furthermore, during the development of this work we detect some reproducibility issues in related literature which we detail, and also open-source all of our implementations introduding a framework to aid the development of similar systems. |
id |
RCAP_c5c22d7e6ce34f59c747e1912192e0aa |
---|---|
oai_identifier_str |
oai:run.unl.pt:10362/151158 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORKSNS24FrailtyHealth OutcomesDiagnosis PredictionFew-shot LearningGraph Neural NetworksDomínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e InformáticaThis dissertation is part of the project FrailCare.AI, which aims to detect frailty in the elderly Portuguese population in order to optimize the SNS24 (telemonitoring) service, with the goal of suggesting health pathways to reduce the patients frailty. Frailty can be defined as the condition of being weak and delicate which normally increases with age and is the consequence of several health and non-health related factors. A patient health journey is recorded in Eletronic Health Record (EHR), which are rich but sparse, noisy and multi-modal sources of truth. These can be used to train predictive models to predict future health states, where frailty is just one of them. In this work, due to lack of data access we pivoted our focus to phenotype prediction, that is, predicting diagnosis. What is more, we tackle the problem of data-insufficiency and class imbalance (e.g. rare diseases and other infrequent occurrences in the training data) by integrating standardized healthcare ontologies within graph neural networks. We study the broad task of phenotype prediction, multi-task scenarios and as well few-shot scenarios - which is when a class rarely occurs in the training set. Furthermore, during the development of this work we detect some reproducibility issues in related literature which we detail, and also open-source all of our implementations introduding a framework to aid the development of similar systems.A presente dissertação insere-se no projecto FrailCare.AI, que visa detectar a fragilidade da população idosa portuguesa com o objectivo de optimizar o serviço de telemonitoriza- ção do Sistema Nacional de Saúde Português (SNS24), e também sugerir acções a tomar para reduzir a fragilidade dos doentes. A fragilidade é uma condição de risco composta por multiplos fatores. Hoje em dia, grande parte da história clinica de cada utente é gravada digitalmente. Estes dados diversos e vastos podem ser usados treinar modelos preditivos cujo objectivo é prever futuros estados de saúde, sendo que fragilidade é só um deles. Devido à falta de accesso a dados, alteramos a tarefa principal deste trabalho para previsão de diágnosticos, onde exploramos o problema de insuficiência de dados e dese- quilíbrio de classes (por exemplo, doenças raras e outras ocorrências pouco frequentes nos dados de treino), integrando ontologias de conceitos médicos por meio de redes neu- ronais de gráfos. Exploramos também outras tarefas e o impacto que elas têm entre si. Para além disso, durante o desenvolvimento desta dissertação identificamos questões a nivel de reproducibilidade da literatura estudada, onde detalhamos e implementamos os conceitos em falta. Com o objectivo de reproducibilidade em mente, nós libertamos o nosso código, introduzindo um biblioteca que permite desenvlver sistemas semelhantes ao nosso.Martins, FlávioCorreia, NunoLondral, AnaRUNCardoso, Miguel Simão Dórdio2023-03-24T12:18:44Z2022-022022-02-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10362/151158enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-03-11T05:33:38Zoai:run.unl.pt:10362/151158Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T03:54:29.606005Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
title |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
spellingShingle |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK Cardoso, Miguel Simão Dórdio SNS24 Frailty Health Outcomes Diagnosis Prediction Few-shot Learning Graph Neural Networks Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
title_short |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
title_full |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
title_fullStr |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
title_full_unstemmed |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
title_sort |
HEALTH OUTCOME PATHWAY PREDICTION. A GRAPH-BASED FRAMEWORK |
author |
Cardoso, Miguel Simão Dórdio |
author_facet |
Cardoso, Miguel Simão Dórdio |
author_role |
author |
dc.contributor.none.fl_str_mv |
Martins, Flávio Correia, Nuno Londral, Ana RUN |
dc.contributor.author.fl_str_mv |
Cardoso, Miguel Simão Dórdio |
dc.subject.por.fl_str_mv |
SNS24 Frailty Health Outcomes Diagnosis Prediction Few-shot Learning Graph Neural Networks Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
topic |
SNS24 Frailty Health Outcomes Diagnosis Prediction Few-shot Learning Graph Neural Networks Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
description |
This dissertation is part of the project FrailCare.AI, which aims to detect frailty in the elderly Portuguese population in order to optimize the SNS24 (telemonitoring) service, with the goal of suggesting health pathways to reduce the patients frailty. Frailty can be defined as the condition of being weak and delicate which normally increases with age and is the consequence of several health and non-health related factors. A patient health journey is recorded in Eletronic Health Record (EHR), which are rich but sparse, noisy and multi-modal sources of truth. These can be used to train predictive models to predict future health states, where frailty is just one of them. In this work, due to lack of data access we pivoted our focus to phenotype prediction, that is, predicting diagnosis. What is more, we tackle the problem of data-insufficiency and class imbalance (e.g. rare diseases and other infrequent occurrences in the training data) by integrating standardized healthcare ontologies within graph neural networks. We study the broad task of phenotype prediction, multi-task scenarios and as well few-shot scenarios - which is when a class rarely occurs in the training set. Furthermore, during the development of this work we detect some reproducibility issues in related literature which we detail, and also open-source all of our implementations introduding a framework to aid the development of similar systems. |
publishDate |
2022 |
dc.date.none.fl_str_mv |
2022-02 2022-02-01T00:00:00Z 2023-03-24T12:18:44Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10362/151158 |
url |
http://hdl.handle.net/10362/151158 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799138133741142016 |