• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 26
  • 8
  • 4
  • 2
  • 1
  • 1
  • Tagged with
  • 62
  • 62
  • 62
  • 12
  • 9
  • 9
  • 9
  • 9
  • 8
  • 7
  • 6
  • 6
  • 5
  • 5
  • 5
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
51

Fatphobia and Clinical Counseling Decision Making in Counselor Education Students

Forristal, Kaitlyn Michelle, Forristal January 2018 (has links)
No description available.
52

Non-Invasive Skin Cancer Classification from Surface Scanned Lesion Images

Dhinagar, Nikhil J. 12 June 2013 (has links)
No description available.
53

Information critical for social work practitioners in the decision making process: An empirical study of implicit knowledge using naturalistic decision making perspective

Hsu, Kai-Shyang 12 September 2006 (has links)
No description available.
54

[en] 3D PROPEDEUTICS VISUALIZATION: TECHNOLOGIES TO SUPPORT CLINICAL DECISION-MAKING / [pt] PROPEDÊUTICA COM VISUALIZAÇÃO 3D: TECNOLOGIAS PARA SUPORTE À TOMADA DE DECISÃO

LEONARDO FRAJHOF 22 October 2020 (has links)
[pt] O objetivo desta pesquisa é projetar serviços da propedêutica clínica considerando o uso da tecnologia de visualização tridimensional como linguagem e fonte de dados para apoiar o raciocínio das decisões clínicas. Visando projetar esse serviço, avaliamos o potencial de algumas tecnologias de visualização tridimensional que poderiam ser úteis para apoiar a propedêutica: visualização de modelos em ambientes virtuais (imersão em Realidade Virtual ); projeção de modelos tridimensionais na realidade visualizados por smartphones (Realidade Aumentada); modelos impressos em 3D. Optamos por projetar um produto que seria o objeto de investigação aprofundada nesta pesquisa: um ambiente virtual para a visualização de casos clínicos reais por meio de óculos de realidade aumentada. Para realizar esta pesquisa um artefato para a visualização de casos clínicos reais por meio de Realidade Virtual (RV) e Realidade Aumentada (RA) foi projetado. O artefacto foi nomeado como ProVis3D e foi desenvolvido para que os médicos assistentes visualizassem, de forma tridimensional, imagens médicas em posição anatômica, mantendo as relações espaciais entre órgãos e vasos de forma fidedigna, para o que é proposto como precisão pelo método, e correspondendo a sua posição no mundo real, ou seja, como estas estão dispostas no interior do corpo humano. Além de poderem ser visualizadas, há possibilidade de interação com as imagens, simulando a percepção táctil, e de produzir sua movimentação: tocar em órgãos segmentados, vasos e vísceras, separar estas estruturas, aumentar seu tamanho e fazer a rotação do objeto em torno de seu eixo. Uma avaliação empírica foi realizada nesta pesquisa, na perspectiva epistemológica-metodológica projetiva Design Science Research (DSR), utilizando como metodologia projetando diferentes casos, com cinco unidades de análise (cada médico convidado para usar o artefato da pesquisa consiste em uma unidade de análise nesta pesquisa). A abordagem foi observacional e interpretativa, com a finalidade de compreender e refletir sobre o comportamento do médico, especialmente a sua tomada de decisão, quando utiliza o artefato desenvolvido nesta pesquisa. Foram projetadas diferentes cenas para possibilitar a observação das experiências dos usuários do ProVis3D (médicos especialistas e não especialistas) ao visualizar imagens médicas em contextos anatômicos reais (reconstrução 3D a partir de exames de tomografia computadorizada e modelos 3D coloridos com todas as estruturas nas suas posições reais) e ao interagir com essas imagens por meio de Realidade Aumentada. A cena virtual consiste em uma imagem tridimensional dinâmica, que pode ser manipulada para apoiar um cirurgião no seu planejamento cirúrgico ou apoiar um radiologista a complementar seu laudo na estação de trabalho (workstation). Os médicos visualizam a cena inicialmente em RV e posteriormente em RA. Duas questões foram elaboradas para o estudo: 1) Será que o artefato realmente possibilita obter informação de valor para a tomada de decisão clínica?; 2) Será que essa tecnologia de visualização realmente cria uma linguagem que possibilita aos médicos produzirem novos conhecimentos? A partir desse estudo, foi demonstrado que artefato ProVis3D tem potencial para apoiar a decisão clínica e que a tecnologia interativa de visualização tridimensional, em Realidade Aumentada, cria uma linguagem, faz os médicos conversarem de um modo diferente, sobre coisas que não estão habituados, possibilitando aos médicos produzirem novos conhecimentos. / [en] The objective of this research is to design a clinical propaedeutic services considering the use of three-dimensional visualization technology as a language and data source to support the reasoning of clinical decisions. In order to design this service, we evaluated the potential of some three-dimensional visualization technologies that could be useful to support the propaedeutics: visualization of models in virtual environments (immersion in Virtual Reality); projection of three-dimensional models in reality viewed by smartphones (Augmented Reality); 3D printed models. We opt to design a product that should be the object of further investigation in this research: a virtual environment for viewing real clinical cases through augmented reality glasses. In order to accomplish this research, an artifact for visualization of real clinical cases through Virtual Reality (VR) and Augmented Reality (AR) was designed The artifact was named as ProVis3D and was designed for assisting physicians to view medical images in an anatomical position and three-dimensional manner, maintaining spatial relationships between organs and vessels reliably as corresponding to their position in the real world and arranged within the human body. Besides being able to be visualized, there is a possibility for interact with the images, simulating the tactile perception, and producing its movement: touching segmented organs, vessels and viscera, separating these structures, increasing their size and rotating the object around them. its axis. An empirical evaluation was performed in this research, in the projective epistemological-methodological perspective Design Science Research (DSR), designing different cases with five units of analysis (each physician invited to use the research artifact consists of one unit of analysis in this research). The approach was observational and interpretative, with the purpose of understanding and reflecting on the physician s behavior, especially his decision making, when using the artifact developed in this research. Different scenes have been designed in order to observe experiences on ProVis3D s users (medical specialists and non-specialists) when viewing medical images in real anatomical contexts (3D reconstruction from CT scans and full-color 3D models with all structures in place) and interacting with these images through augmented reality. The virtual scene consists of a dynamic three-dimensional image that can be manipulated to support a surgeon on his surgical planning or support a radiologist to complement his workstation report. Doctors view the scene initially in RV and later in RA. Two questions were elaborated for the study: 1) Does the artifact really make it possible to obtain valuable information for clinical decision making? 2) Does this visualization technology really create a language that enables doctors to produce new knowledge? It has been demonstrated that ProVis3D artifact has the potential to support clinical decision making and that the augmented reality interactive three-dimensional visualization technology creates a language; it makes doctors talk differently about things they are not used to, enabling doctors to produce new knowledge.
55

Avaliação de um instrumento de auxílio à tomada de decisão para a priorização de vagas em unidades de terapia intensiva / Evaluation of a decision-aid tool for prioritization of admissions to the intensive care unit

Ramos, João Gabriel Rosa 02 May 2018 (has links)
Introdução: Triagem para admissão em unidades de terapia intensiva (UTIs) é realizada rotineiramente e é comumente baseada somente no julgamento clínico, o que pode mascarar vieses e preconceitos. Neste estudo, foram avaliadas a reprodutibilidade e validade de um algoritmo de apoio a decisões de triagem em UTI. Também foi avaliado o efeito da implementação de um instrumento de auxílio à tomada de decisão para a priorização de vagas de UTI nas decisões de admissão em UTI. Foi avaliada, ainda, a acurácia da predição prognóstica dos médicos na população de pacientes em deterioração clínica aguda. Métodos: Para o primeiro objetivo do estudo, um algoritmo computadorizado para auxiliar as decisões de priorização de vagas em UTI foi desenvolvido para classificar pacientes nas categorias do sistema de priorização da \"Society of Critical Care Medicine (SCCM)\". Nove médicos experientes (experts) avaliaram quarenta vinhetas clínicas baseadas em pacientes reais. A referência foi definida como as prioridades classificadas por dois investigadores com acesso ao prontuário completo dos pacientes. As concordâncias entre as prioridades do algoritmo com as prioridades da referência e com as prioridades dos experts foram avaliadas. As correlações entre a prioridade do algoritmo e o julgamento clínico de adequação da admissão na UTI em contexto com e sem escassez de vagas também foram avaliadas. A validade foi ainda avaliada através da aplicação do algoritmo, retrospectivamente em uma coorte de 603 pacientes com solicitação de vagas de UTI, para correlação com desfechos clínicos. Para o segundo objetivo do estudo, um estudo prospectivo, quaseexperimental foi conduzido, antes (maio/2014 a novembro/2014, fase 1) e após (novembro/2014 a maio/2015, fase 2) a implementação de um instrumento de auxílio à tomada de decisão, que foi baseado no algoritmo descrito acima. Foi avaliado o impacto da implementação do instrumento de auxílio à tomada de decisão na ocorrência de admissões potencialmente inapropriadas na UTI em uma coorte de pacientes com solicitações urgentes de vaga de UTI. O desfecho primário foi a proporção de solicitações de vaga potencialmente inapropriadas que foram admitidas na UTI em até 48 horas após a solicitação. Solicitações de vaga potencialmente inapropriadas foram definidas como pacientes prioridade 4B, conforme diretrizes da SCCM de 1999, ou prioridade 5, conforme diretrizes da SCCM de 2016. Foram realizadas análises multivariadas com teste de interação entre fase e prioridades para avaliação dos efeitos diferenciados em cada estrato de prioridade. Para o terceiro objetivo do estudo, a predição prognóstica realizada pelo médico solicitante foi registrada no momento da solicitação de vaga de UTI. Resultados: No primeiro objetivo do estudo, a concordância entre as prioridades do algoritmo e as prioridades da referência foi substancial, com uma mediana de kappa de 0,72 (IQR 0,52-0,77). As prioridades do algoritmo evidenciaram uma maior reprodutibilidade entre os pares [kappa = 0,61 (IC95% 0,57-0,65) e mediana de percentagem de concordância = 0,64 (IQR 0,59-0,70)], quando comparada à reprodutibilidade entre os pares das prioridades dos experts [kappa = 0,51 (IC95% 0,47-0,55) e mediana de percentagem de concordância = 0,49 (IQR 0,44-0,56)], p=0,001. As prioridades do algoritmo também foram associadas ao julgamento clínico de adequação da admissão na UTI (vinhetas com prioridades 1, 2, 3 e 4 seriam admitidas no último leito de UTI em 83,7%, 61,2%, 45,2% e 16,8% dos cenários, respectivamente, p < 0,001) e com desfechos clínicos reais na coorte retrospectiva, como admissão na UTI, consultas com equipe de cuidados paliativos e mortalidade hospitalar. No segundo objetivo do estudo, 2374 solicitações urgentes de vaga de UTI foram avaliadas, das quais 1184 (53,8%) pacientes foram admitidos na UTI. A implementação do instrumento de auxílio à tomada de decisão foi associada com uma redução de admissões potencialmente inapropriadas na UTI, tanto utilizando a classificação de 1999 [adjOR (IC95%) = 0,36 (0,13-0,97), p = 0,043], quanto utilizando a classificação de 2016 [adjOR (IC95%) = 0,35 (0,13-0,96, p = 0,041)]. Não houve diferença em mortalidade entre as fases 1 e 2 do estudo. No terceiro objetivo do estudo, a predição prognóstica do médico solicitante foi associada com mortalidade. Ocorreram 593 (34,4%), 215 (66,4%) e 51 (94,4%) óbitos nos grupos com prognóstico de sobrevivência sem sequelas graves, sobrevivência com sequelas graves e nãosobrevivência, respectivamente (p < 0,001). Sensibilidade foi 31%, especificidade foi 91% e a área sob a curva ROC foi de 0,61 para predição de mortalidade hospitalar. Após análise multivariada, a gravidade da doença aguda, funcionalidade prévia e admissão na UTI foram associadas com uma maior chance de erro prognóstico, enquanto que uma predição de pior prognóstico foi associada a uma menor chance de erro prognóstico. O grau de expertise do médico solicitante não teve efeito na predição prognóstica. Discussão/Conclusão: Neste estudo, um algoritmo de apoio a decisões de triagem em UTI demonstrou boa reprodutibilidade e validade. Além disso, a implementação de um instrumento de auxílio à tomada de decisões para priorização de vagas de UTI foi associada a uma redução de admissões potencialmente inapropriadas na UTI. Também foi encontrado que a predição prognóstica dos médicos solicitantes foi associada a mortalidade hospitalar, porém a acurácia foi pobre, principalmente devido a uma baixa sensibilidade para detectar risco de morte / Introduction: Intensive care unit (ICU) admission triage is performed routinely and is often based solely on clinical judgment, which could mask biases. In this study, we sought to evaluate the reliability and validity of an algorithm to aid ICU triage decisions. We also aimed to evaluate the effect of implementing a decision-aid tool for ICU triage on ICU admission decisions. We also evaluated the accuracy of physician\'s prediction of hospital mortality in in acutely deteriorating patients. Methods: For the first objective of the study, a computerized algorithm to aid ICU triage decisions was developed to classify patients into the Society of Critical Care Medicine\'s prioritization system. Nine senior physicians evaluated forty clinical vignettes based on real patients. Reference standard was defined as the priorities ascribed by two investigators with full access to patient\'s records. Agreement of algorithm-based priorities with the reference standard and with intuitive priorities provided by the physicians were evaluated. Correlations between algorithm prioritization and physician\'s judgment of appropriateness of ICU admission in scarcity and non-scarcity settings were also evaluated. Validity was further assessed by retrospectively applying this algorithm to 603 patients with requests for ICU admission for association with clinical outcomes. For the second objective of the study, a prospective, quasi-experimental study was conducted, before (May 2014 to November 2014, phase 1) and after (November 2014 to May 2015, phase 2) the implementation of a decision-aid tool for ICU admission triage, which was based on the aforementioned algorithm. We assessed the impact of the implementation of the decision-aid tool in potentially inappropriate ICU admissions in a cohort of patients referred for urgent ICU admission. Primary outcome was the proportion of potentially inappropriate ICU referrals that were admitted to the ICU in 48 hours following referral. Potentially inappropriate ICU referrals were defined as priority 4B patients, as described by the 1999 Society of Critical Care Medicine (SCCM) guidelines and as priority 5 patients, as described by the 2016 SCCM guidelines. We conducted multivariate analyses and evaluated the interaction between phase and triage priorities to assess for differential effects in each priority strata. For the third objective of the study, physicians\' prognosis and other variables were recorded at the moment of ICU referral. Results: On the first objective of the study, agreement between algorithm-based priorities and the reference standard was substantial, with a median kappa of 0.72 (IQR 0.52-0.77). Algorithm-based priorities demonstrated higher interrater reliability [overall kappa of 0.61 (95%CI 0.57-0.65) and median percent agreement of 0.64 (IQR 0.59-0.70)] than physician\'s intuitive prioritization [overall kappa of 0.51 (95%CI 0.47-0.55) and median percent agreement of 0.49 (IQR 0.44-0.56)], p=0.001. Algorithm-based priorities were also associated with physicians\' judgment of appropriateness of ICU admission (priorities 1, 2, 3 and 4 vignettes would be admitted to the last ICU bed in 83.7%, 61.2%, 45.2% and 16.8% of the scenarios, respectively, p < 0.001) and with actual ICU admission, palliative care consultation and hospital mortality in the retrospective cohort. On the second objective of the study, of 2374 urgent ICU referrals, 1184 (53.8%) patients were admitted to the ICU. Implementation of the decision-aid tool was associated with a reduction of potentially inappropriate ICU admissions using the 1999 [adjOR (95% CI) = 0.36 (0.13-0.97), p = 0.043] or 2016 [adjOR (95%CI) = 0.35 (0.13-0.96, p = 0.041)] definitions. There was no difference on mortality between phases 1 and 2. On the third objective of the study, physician\'s prognosis was associated to hospital mortality. There were 593 (34.4%), 215 (66.4%) and 51 (94.4%) deaths in the groups ascribed a prognosis of survival without disabilities, survival with severe disabilities or no survival, respectively (p < 0.001). Sensitivity was 31%, specificity was 91% and the area under the ROC curve was 0.61 for prediction of mortality. After multivariable analysis, severity of illness, performance status and ICU admission were associated to an increased likelihood of incorrect classification, while worse predicted prognosis was associated to a lower chance of incorrect classification. Physician\'s level of expertise had no effect on predictive ability. Discussion/Conclusion: In this study, a ICU admission triage algorithm demonstrated good reliability and validity. Moreover, the implementation of a decision-aid tool for ICU triage was associated with a reduction of potentially inappropriate ICU admissions. It was also found that physician\'s prediction was associated to hospital mortality, but overall accuracy was poor, mainly due to low sensitivity to detect mortality risk
56

Avaliação de um instrumento de auxílio à tomada de decisão para a priorização de vagas em unidades de terapia intensiva / Evaluation of a decision-aid tool for prioritization of admissions to the intensive care unit

João Gabriel Rosa Ramos 02 May 2018 (has links)
Introdução: Triagem para admissão em unidades de terapia intensiva (UTIs) é realizada rotineiramente e é comumente baseada somente no julgamento clínico, o que pode mascarar vieses e preconceitos. Neste estudo, foram avaliadas a reprodutibilidade e validade de um algoritmo de apoio a decisões de triagem em UTI. Também foi avaliado o efeito da implementação de um instrumento de auxílio à tomada de decisão para a priorização de vagas de UTI nas decisões de admissão em UTI. Foi avaliada, ainda, a acurácia da predição prognóstica dos médicos na população de pacientes em deterioração clínica aguda. Métodos: Para o primeiro objetivo do estudo, um algoritmo computadorizado para auxiliar as decisões de priorização de vagas em UTI foi desenvolvido para classificar pacientes nas categorias do sistema de priorização da \"Society of Critical Care Medicine (SCCM)\". Nove médicos experientes (experts) avaliaram quarenta vinhetas clínicas baseadas em pacientes reais. A referência foi definida como as prioridades classificadas por dois investigadores com acesso ao prontuário completo dos pacientes. As concordâncias entre as prioridades do algoritmo com as prioridades da referência e com as prioridades dos experts foram avaliadas. As correlações entre a prioridade do algoritmo e o julgamento clínico de adequação da admissão na UTI em contexto com e sem escassez de vagas também foram avaliadas. A validade foi ainda avaliada através da aplicação do algoritmo, retrospectivamente em uma coorte de 603 pacientes com solicitação de vagas de UTI, para correlação com desfechos clínicos. Para o segundo objetivo do estudo, um estudo prospectivo, quaseexperimental foi conduzido, antes (maio/2014 a novembro/2014, fase 1) e após (novembro/2014 a maio/2015, fase 2) a implementação de um instrumento de auxílio à tomada de decisão, que foi baseado no algoritmo descrito acima. Foi avaliado o impacto da implementação do instrumento de auxílio à tomada de decisão na ocorrência de admissões potencialmente inapropriadas na UTI em uma coorte de pacientes com solicitações urgentes de vaga de UTI. O desfecho primário foi a proporção de solicitações de vaga potencialmente inapropriadas que foram admitidas na UTI em até 48 horas após a solicitação. Solicitações de vaga potencialmente inapropriadas foram definidas como pacientes prioridade 4B, conforme diretrizes da SCCM de 1999, ou prioridade 5, conforme diretrizes da SCCM de 2016. Foram realizadas análises multivariadas com teste de interação entre fase e prioridades para avaliação dos efeitos diferenciados em cada estrato de prioridade. Para o terceiro objetivo do estudo, a predição prognóstica realizada pelo médico solicitante foi registrada no momento da solicitação de vaga de UTI. Resultados: No primeiro objetivo do estudo, a concordância entre as prioridades do algoritmo e as prioridades da referência foi substancial, com uma mediana de kappa de 0,72 (IQR 0,52-0,77). As prioridades do algoritmo evidenciaram uma maior reprodutibilidade entre os pares [kappa = 0,61 (IC95% 0,57-0,65) e mediana de percentagem de concordância = 0,64 (IQR 0,59-0,70)], quando comparada à reprodutibilidade entre os pares das prioridades dos experts [kappa = 0,51 (IC95% 0,47-0,55) e mediana de percentagem de concordância = 0,49 (IQR 0,44-0,56)], p=0,001. As prioridades do algoritmo também foram associadas ao julgamento clínico de adequação da admissão na UTI (vinhetas com prioridades 1, 2, 3 e 4 seriam admitidas no último leito de UTI em 83,7%, 61,2%, 45,2% e 16,8% dos cenários, respectivamente, p < 0,001) e com desfechos clínicos reais na coorte retrospectiva, como admissão na UTI, consultas com equipe de cuidados paliativos e mortalidade hospitalar. No segundo objetivo do estudo, 2374 solicitações urgentes de vaga de UTI foram avaliadas, das quais 1184 (53,8%) pacientes foram admitidos na UTI. A implementação do instrumento de auxílio à tomada de decisão foi associada com uma redução de admissões potencialmente inapropriadas na UTI, tanto utilizando a classificação de 1999 [adjOR (IC95%) = 0,36 (0,13-0,97), p = 0,043], quanto utilizando a classificação de 2016 [adjOR (IC95%) = 0,35 (0,13-0,96, p = 0,041)]. Não houve diferença em mortalidade entre as fases 1 e 2 do estudo. No terceiro objetivo do estudo, a predição prognóstica do médico solicitante foi associada com mortalidade. Ocorreram 593 (34,4%), 215 (66,4%) e 51 (94,4%) óbitos nos grupos com prognóstico de sobrevivência sem sequelas graves, sobrevivência com sequelas graves e nãosobrevivência, respectivamente (p < 0,001). Sensibilidade foi 31%, especificidade foi 91% e a área sob a curva ROC foi de 0,61 para predição de mortalidade hospitalar. Após análise multivariada, a gravidade da doença aguda, funcionalidade prévia e admissão na UTI foram associadas com uma maior chance de erro prognóstico, enquanto que uma predição de pior prognóstico foi associada a uma menor chance de erro prognóstico. O grau de expertise do médico solicitante não teve efeito na predição prognóstica. Discussão/Conclusão: Neste estudo, um algoritmo de apoio a decisões de triagem em UTI demonstrou boa reprodutibilidade e validade. Além disso, a implementação de um instrumento de auxílio à tomada de decisões para priorização de vagas de UTI foi associada a uma redução de admissões potencialmente inapropriadas na UTI. Também foi encontrado que a predição prognóstica dos médicos solicitantes foi associada a mortalidade hospitalar, porém a acurácia foi pobre, principalmente devido a uma baixa sensibilidade para detectar risco de morte / Introduction: Intensive care unit (ICU) admission triage is performed routinely and is often based solely on clinical judgment, which could mask biases. In this study, we sought to evaluate the reliability and validity of an algorithm to aid ICU triage decisions. We also aimed to evaluate the effect of implementing a decision-aid tool for ICU triage on ICU admission decisions. We also evaluated the accuracy of physician\'s prediction of hospital mortality in in acutely deteriorating patients. Methods: For the first objective of the study, a computerized algorithm to aid ICU triage decisions was developed to classify patients into the Society of Critical Care Medicine\'s prioritization system. Nine senior physicians evaluated forty clinical vignettes based on real patients. Reference standard was defined as the priorities ascribed by two investigators with full access to patient\'s records. Agreement of algorithm-based priorities with the reference standard and with intuitive priorities provided by the physicians were evaluated. Correlations between algorithm prioritization and physician\'s judgment of appropriateness of ICU admission in scarcity and non-scarcity settings were also evaluated. Validity was further assessed by retrospectively applying this algorithm to 603 patients with requests for ICU admission for association with clinical outcomes. For the second objective of the study, a prospective, quasi-experimental study was conducted, before (May 2014 to November 2014, phase 1) and after (November 2014 to May 2015, phase 2) the implementation of a decision-aid tool for ICU admission triage, which was based on the aforementioned algorithm. We assessed the impact of the implementation of the decision-aid tool in potentially inappropriate ICU admissions in a cohort of patients referred for urgent ICU admission. Primary outcome was the proportion of potentially inappropriate ICU referrals that were admitted to the ICU in 48 hours following referral. Potentially inappropriate ICU referrals were defined as priority 4B patients, as described by the 1999 Society of Critical Care Medicine (SCCM) guidelines and as priority 5 patients, as described by the 2016 SCCM guidelines. We conducted multivariate analyses and evaluated the interaction between phase and triage priorities to assess for differential effects in each priority strata. For the third objective of the study, physicians\' prognosis and other variables were recorded at the moment of ICU referral. Results: On the first objective of the study, agreement between algorithm-based priorities and the reference standard was substantial, with a median kappa of 0.72 (IQR 0.52-0.77). Algorithm-based priorities demonstrated higher interrater reliability [overall kappa of 0.61 (95%CI 0.57-0.65) and median percent agreement of 0.64 (IQR 0.59-0.70)] than physician\'s intuitive prioritization [overall kappa of 0.51 (95%CI 0.47-0.55) and median percent agreement of 0.49 (IQR 0.44-0.56)], p=0.001. Algorithm-based priorities were also associated with physicians\' judgment of appropriateness of ICU admission (priorities 1, 2, 3 and 4 vignettes would be admitted to the last ICU bed in 83.7%, 61.2%, 45.2% and 16.8% of the scenarios, respectively, p < 0.001) and with actual ICU admission, palliative care consultation and hospital mortality in the retrospective cohort. On the second objective of the study, of 2374 urgent ICU referrals, 1184 (53.8%) patients were admitted to the ICU. Implementation of the decision-aid tool was associated with a reduction of potentially inappropriate ICU admissions using the 1999 [adjOR (95% CI) = 0.36 (0.13-0.97), p = 0.043] or 2016 [adjOR (95%CI) = 0.35 (0.13-0.96, p = 0.041)] definitions. There was no difference on mortality between phases 1 and 2. On the third objective of the study, physician\'s prognosis was associated to hospital mortality. There were 593 (34.4%), 215 (66.4%) and 51 (94.4%) deaths in the groups ascribed a prognosis of survival without disabilities, survival with severe disabilities or no survival, respectively (p < 0.001). Sensitivity was 31%, specificity was 91% and the area under the ROC curve was 0.61 for prediction of mortality. After multivariable analysis, severity of illness, performance status and ICU admission were associated to an increased likelihood of incorrect classification, while worse predicted prognosis was associated to a lower chance of incorrect classification. Physician\'s level of expertise had no effect on predictive ability. Discussion/Conclusion: In this study, a ICU admission triage algorithm demonstrated good reliability and validity. Moreover, the implementation of a decision-aid tool for ICU triage was associated with a reduction of potentially inappropriate ICU admissions. It was also found that physician\'s prediction was associated to hospital mortality, but overall accuracy was poor, mainly due to low sensitivity to detect mortality risk
57

Macrocognition in the Health Care Built Environment (m-HCBE): A Focused Ethnographic Study of 'Neighborhoods' in a Pediatric Intensive Care Unit: A Dissertation

O'Hara Sullivan, Susan 12 December 2016 (has links)
Objectives: The objectives of this research were to describe the interactions (formal and informal) in which macrocognitive functions occur and their location on a pediatric intensive care unit (PICU); describe challenges and facilitators of macrocognition using three constructs of space syntax (openness, connectivity, and visibility); and analyze the health care built environment (HCBE) using those constructs to explicate influences on macrocognition. Background: In high reliability, complex industries, macrocognition is an approach to develop new knowledge among interprofessional team members. Although macrocognitive functions have been analyzed in multiple health care settings, the effect of the HCBE on those functions has not been directly studied. The theoretical framework, “Macrocognition in the Health Care Built Environment” (m-HCBE) addresses this relationship. Methods: A focused ethnographic study was conducted, including observation and focus groups. Architectural drawing files used to create distance matrices and isovist field view analyses were compared to panoramic photographs and ethnographic data. Results: Neighborhoods comprised of corner configurations with maximized visibility enhanced team interactions as well as observation of patients, offering the greatest opportunity for informal situated macrocognitive interactions (SMIs). Conclusions: Results from this study support the intricate link between macrocognitive interactions and space syntax constructs within the HCBE. These findings help to advance the m-HCBE theory for improving physical space by designing new spaces or refining existing spaces, or for adapting IPT practices to maximize formal and informal SMI opportunities; this lays the groundwork for future research to improve safety and quality for patient and family care.
58

Predictive Models for Coronary Heart Disease Prognosis using Ensemble Learning : master's thesis / Прогностические модели для прогнозирования ишемической болезни сердца с использованием коллективного обучения

Шах, Брахим, Shah, Brahim January 2024 (has links)
This thesis investigates the application of ensemble learning techniques in developing predictive models for coronary heart disease prognosis, aiming to enhance diagnostic capabilities and improve patient outcomes in cardiovascular medicine. By leveraging advanced computational methods and machine learning algorithms, the study focuses on automating the detection of myocardial infarction and heart conduction disorders using a deep learning model trained on ECG signals from a diverse dataset. The research methodology involves a systematic review of highly relevant papers, exclusion criteria to ensure the specificity of the study, and a search process in reputable academic libraries. Through a comparative analysis of selected papers and an in-depth exploration of machine learning approaches, the thesis aims to contribute to the advancement of predictive modeling techniques in cardiology. The findings of this research have the potential to significantly impact the field of cardiovascular care by providing more accurate prognostic tools for coronary heart disease management. / В данной работе исследуется применение методов коллективного обучения при разработке прогностических моделей для прогнозирования ишемической болезни сердца с целью расширения диагностических возможностей и улучшения результатов лечения пациентов в сердечно-сосудистой медицине. Используя передовые вычислительные методы и алгоритмы машинного обучения, исследование направлено на автоматизацию выявления инфаркта миокарда и нарушений сердечной проводимости с использованием модели глубокого обучения, обученной на основе сигналов ЭКГ из различных наборов данных. Методология исследования включает систематический обзор наиболее значимых статей, критерии исключения для обеспечения специфичности исследования и процесс поиска в авторитетных академических библиотеках. Благодаря сравнительному анализу избранных работ и углубленному изучению подходов к машинному обучению, диссертация призвана внести вклад в развитие методов прогностического моделирования в кардиологии. Результаты этого исследования могут оказать существенное влияние на сферу сердечно-сосудистой помощи, предоставив более точные инструменты прогнозирования для лечения ишемической болезни сердца.
59

Enhancing association rules algorithms for mining distributed databases : integration of fast BitTable and multi-agent association rules mining in distributed medical databases for decision support

Abdo, Walid Adly Atteya January 2012 (has links)
Over the past few years, mining data located in heterogeneous and geographically distributed sites have been designated as one of the key important issues. Loading distributed data into centralized location for mining interesting rules is not a good approach. This is because it violates common issues such as data privacy and it imposes network overheads. The situation becomes worse when the network has limited bandwidth which is the case in most of the real time systems. This has prompted the need for intelligent data analysis to discover the hidden information in these huge amounts of distributed databases. In this research, we present an incremental approach for building an efficient Multi-Agent based algorithm for mining real world databases in geographically distributed sites. First, we propose the Distributed Multi-Agent Association Rules algorithm (DMAAR) to minimize the all-to-all broadcasting between distributed sites. Analytical calculations show that DMAAR reduces the algorithm complexity and minimizes the message communication cost. The proposed Multi-Agent based algorithm complies with the Foundation for Intelligent Physical Agents (FIPA), which is considered as the global standards in communication between agents, thus, enabling the proposed algorithm agents to cooperate with other standard agents. Second, the BitTable Multi-Agent Association Rules algorithm (BMAAR) is proposed. BMAAR includes an efficient BitTable data structure which helps in compressing the database thus can easily fit into the memory of the local sites. It also includes two BitWise AND/OR operations for quick candidate itemsets generation and support counting. Moreover, the algorithm includes three transaction trimming techniques to reduce the size of the mined data. Third, we propose the Pruning Multi-Agent Association Rules algorithm (PMAAR) which includes three candidate itemsets pruning techniques for reducing the large number of generated candidate itemsets, consequently, reducing the total time for the mining process. The proposed PMAAR algorithm has been compared with existing Association Rules algorithms against different benchmark datasets and has proved to have better performance and execution time. Moreover, PMAAR has been implemented on real world distributed medical databases obtained from more than one hospital in Egypt to discover the hidden Association Rules in patients' records to demonstrate the merits and capabilities of the proposed model further. Medical data was anonymously obtained without the patients' personal details. The analysis helped to identify the existence or the absence of the disease based on minimum number of effective examinations and tests. Thus, the proposed algorithm can help in providing accurate medical decisions based on cost effective treatments, improving the medical service for the patients, reducing the real time response for the health system and improving the quality of clinical decision making.
60

Enhancing association rules algorithms for mining distributed databases. Integration of fast BitTable and multi-agent association rules mining in distributed medical databases for decision support.

Abdo, Walid A.A. January 2012 (has links)
Over the past few years, mining data located in heterogeneous and geographically distributed sites have been designated as one of the key important issues. Loading distributed data into centralized location for mining interesting rules is not a good approach. This is because it violates common issues such as data privacy and it imposes network overheads. The situation becomes worse when the network has limited bandwidth which is the case in most of the real time systems. This has prompted the need for intelligent data analysis to discover the hidden information in these huge amounts of distributed databases. In this research, we present an incremental approach for building an efficient Multi-Agent based algorithm for mining real world databases in geographically distributed sites. First, we propose the Distributed Multi-Agent Association Rules algorithm (DMAAR) to minimize the all-to-all broadcasting between distributed sites. Analytical calculations show that DMAAR reduces the algorithm complexity and minimizes the message communication cost. The proposed Multi-Agent based algorithm complies with the Foundation for Intelligent Physical Agents (FIPA), which is considered as the global standards in communication between agents, thus, enabling the proposed algorithm agents to cooperate with other standard agents. Second, the BitTable Multi-Agent Association Rules algorithm (BMAAR) is proposed. BMAAR includes an efficient BitTable data structure which helps in compressing the database thus can easily fit into the memory of the local sites. It also includes two BitWise AND/OR operations for quick candidate itemsets generation and support counting. Moreover, the algorithm includes three transaction trimming techniques to reduce the size of the mined data. Third, we propose the Pruning Multi-Agent Association Rules algorithm (PMAAR) which includes three candidate itemsets pruning techniques for reducing the large number of generated candidate itemsets, consequently, reducing the total time for the mining process. The proposed PMAAR algorithm has been compared with existing Association Rules algorithms against different benchmark datasets and has proved to have better performance and execution time. Moreover, PMAAR has been implemented on real world distributed medical databases obtained from more than one hospital in Egypt to discover the hidden Association Rules in patients¿ records to demonstrate the merits and capabilities of the proposed model further. Medical data was anonymously obtained without the patients¿ personal details. The analysis helped to identify the existence or the absence of the disease based on minimum number of effective examinations and tests. Thus, the proposed algorithm can help in providing accurate medical decisions based on cost effective treatments, improving the medical service for the patients, reducing the real time response for the health system and improving the quality of clinical decision making.

Page generated in 0.0857 seconds