Introdução: Os medicamentos são importantes elementos da maioria dos esquemas terapêuticos cobertos pelo Sistema Único de Saúde (SUS), representando significativa parcela do orçamento no país. O Complexo de Saúde vinculado ao Hospital das Clínicas atende toda a região noroeste do Estado de São Paulo e de outras partes do estado e do país, como centro de referência em tratamentos de alta complexidade, sendo frequente a prescrição de medicamentos de alto custo (MAC). Estima-se que 75,4% do orçamento geral para compra de medicamentos do complexo HCRP-FMRP-USP, são dedicados à aquisição de medicação não padronizada (medicamento especial) num total de aproximadamente R$ 46.313.170,08 (2015). Sendo assim, ferramentas para controle não só da prescrição, como também da aquisição e seu uso são fundamentais para otimizar a gestão do Hospital, evoluindo de um caráter reativo a um proativo, no qual a tomada de decisões tenha como base um histórico e indicadores de casos apresentados no complexo. Objetivo: Desenvolver uma plataforma eletrônica baseada na rede mundial de computadores, que possibilite a gestão entendida como documentação, rastreabilidade e inter-relacionamento entre os componentes da cadeia de decisão de medicamentos considerados especiais no Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo. Métodos: Compreendeu o desenvolvimento de um sistema que tem como características principais monitoramento, acompanhamento e controle da cadeia de decisão de medicamentos que são considerados especiais pela instituição. Este sistema também permite a tomada de decisões, o desenvolvimento de indicadores em tempo real para decisão administrativa e o controle que requer a cadeia de suprimento de medicamentos de alto custo em cada um dos seus componentes. Resultados: Maior e melhor comunicação entre as unidades de farmácia, o solicitante (médico), o Departamento de atenção à Saúde (DAS) e os locais do Complexo HC-FMRP-USP que compõem a cadeia de decisão do suprimento de medicamentos especiais (MAC); além disso, possibilitará organizar um histórico de dados que poderá ser transposto facilmente a indicadores para o plano assistencial à medida, que garanta a presença de um agente transformador. Conclusões: Uma plataforma eletrônica foi desenvolvida que permite armazenamento, gestão e o processamento de dados e informações respeito à cadeia de decisão do fornecimento de medicamentos não padronizados / Introduction: Medicines are important elements in health care, especially those covered by the Brazilian Unified Healthcare System - Sistema Único de Saúde (SUS), representing a significant portion of its budget. The health infrastructure linked to Hospital das Clínicas serves throughout the northwest region of State the São Paulo and other parts of the state and the country. It is, therefore, known as a reference center for highly complex treatments and, for this reason, frequently prescribes treatments with expensive drugs. Is estimated that 75.4% of the general budget of HCRP-FMRP-USP complex is dedicated to the acquisition of this type of medication, i.e., not standardized medication (special medication), that has a value of approximately USD $14.434.300 (2015). Therefore, tools for controlling not only the prescription, as well as the acquisition and use, becomes critical to optimize the management of the hospital, aiming to move from a reactive to proactive role, where decision-making is based on a history and on indicators of the cases presented in the complex. Objective: To develop an electronic platform based on the World Wide Web, which allows the management, documentation, traceability and interrelationship between the components of the considered decision chain of nonstandard medicines in the Clinics Hospital of Ribeirão Preto Medical School of the University of Sao Paulo. Methods: Include a software development that has, as main features, tracking, monitoring and control of decision chain of drugs, which are considered special by the institution. This software also allows making decisions, development of indicators in real-time and administrative decisions that require the regulatory control supply system of high cost of medicines in each of its components. Results: Further and improved communication between the pharmacy units, the applicant (physician), the Department of attention to health (DAS) and places from the HC-FMRP-USP complex that integrate the chain of decision of the special drug supply. Moreover, organize a data history, which easily can be implemented to indicators for the assistance plan as guaranteeing the presence of a transforming agent. Conclusions: Developed an electronic platform that enables storage, management and processing of data and information, considering the chain decision of nonstandard medicines supply.

Federated learning is a framework in machine learning that allows for training a model while maintaining data privacy. Moreover, it allows clients with their own data to collaborate in order to build a stronger, shared model. Federated learning is of particular interest to healthcare data, since it is of the utmost importance to respect patient privacy while still building useful diagnostic tools. However, healthcare data can be complicated — data format might differ across providers, leading to unexpected inputs and incompatibility between different providers. For example, electrocardiograms might differ in sampling rate or number of leads used, meaning that a classifier trained at one hospital might be useless to another. We propose using autoencoders to address this problem, transforming important information contained in electrocardiograms to a uniform input, where federated learning can then be used to train a strong classifier for multiple healthcare providers. Furthermore, we propose using statistically-guided hyperparameter tuning to ensure fast convergence of the model.

No description available.

tudies in teaching and non-teaching hospitals have shown that one- quarter to one-third of all patients receive an antimicrobial (AMC) drug during their hospital stay." 1-30 Many of these patients (ranging from 30 to 60%), particularly those on the surgical services, have no definite evidence of infection. Inquiry regarding the use of AMC drugs can be traced back more than two decades. In 1961, the Commission on Professional and Hospital Activities in its Professional Activity Study (CPHA-PAS) surveyed 24 hospitals." They found that approximately 27% of all patients were given an AMC drug; however, it was estimated by PAS that only about 12% of these patients should have received AMC therapy under the most conservative medical practice.

In the past three years, evidence based medicine emerged as a powerful force in an effort to improve quality and health outcomes, and to reduce cost of care. Computerized physician order entry (CPOE) applications brought safety and efficiency features to clinical settings, including ease of ordering medications via pre-defined sets. Order sets offer promise of standardized care beyond convenience features through evidence-based practices built upon a growing and powerful knowledge of clinical professionals to achieve potentially more consistent health outcomes with patients and to reduce frequency of medical errors, adverse drug effects, and unintended side effects during treatment. While order sets existed in paper form prior to the introduction of CPOE, their true potential was only unleashed with support of clinical informatics, at those healthcare facilities that installed CPOE systems and reap rewards of standardized care. Despite ongoing utilization of order sets at facilities that implemented CPOE, there is a lack of quantitative evidence behind their benefits. Comprehensive research into their impact requires a history of electronic medical records necessary to produce large population samples to achieve statistically significant results. The study, conducted at a large Midwest healthcare system consisting of several community and academic hospitals, was aimed at quantitatively analyzing benefits of the order sets applied to prevent venous thromboembolism (VTE) and treat pneumonia, congestive heart failure (CHF), and acute myocardial infarction (AMI) - testing hospital mortality, readmission, complications, and length of stay (LOS) as health outcomes. Results indicated reduction of acute VTE rates among non-surgical patients in the experimental group, while LOS and complications benefits were inconclusive. Pneumonia patients in the experimental group had lower mortality, readmissions, LOS, and complications rates. CHF patients benefited from order sets in terms of mortality and LOS, while there was no sufficient data to display results for readmissions and complications. Utilization of AMI order sets was insufficient to produce statistically significant results. Results will (1) empower health providers with evidence to justify implementation of order sets due to their effectiveness in driving improvements in health outcomes and efficiency of care and (2) provide researchers with new ideas to conduct health outcomes research.

Este trabalho faz parte de um projeto maior, o Annotation and Image Markup Project, que tem o objetivo de criar uma base de conhecimento médico sobre imagens radiológicas para identificação, acompanhamento e reasoning acerca de lesões tumorais em pesquisas sobre câncer e consultórios médicos. Esse projeto está sendo desenvolvido em conjunto com o Radiological Sciences Laboratory da Stanford University. O problema específico, que será abordado nesse trabalho, é que a maior parte das informações semânticas sobre imagens radiológicas não são capturados e relacionados às mesmas usando termos de ontologias biomédicas e padrões, como o RadLex e DICOM, o que impossibilita a sua avaliação automática por computadores, busca em arquivos médicos em hospitais, etc. Para tratar isso, os radiologistas precisam de uma solução computacional fácil, intuitiva e acessível para adicionar essas informações. Nesse trabalho foi desenvolvida uma solução Web para inclusão dessas anotações, o sistema ePAD. O aplicativo permite a recuperação de imagens médicas, como as imagens disponíveis em sistemas de informação hospitalares (PACS), o delineamento dos contornos de lesões tumorais, a associação de termos ontológicos a esses contornos e o armazenamento desses termos em uma base de conhecimento. Os principais desafios desse trabalho envolveram a aplicação de interfaces intuitivas baseadas em Rich Internet Applications e sua operação a partir de um navegador Web padrão. O primeiro protótipo funcional do ePAD atingiu seus objetivos ao demonstrar sua viabilidade técnica, sendo capaz de executar o mesmo trabalho básico de anotação de aplicações Desktop, como o OsiriX-iPad, sem o mesmo overhead. Também mostrou a sua utilidade a comunidade médica o que gerou o interesse de usuários potenciais / This work is a part of a larger project, the Annotation and Markup Project, which aims to create a medical knowledge base about radiological images to identify, monitor and reason about tumors in cancer research and medical practices. This project is being developed in conjunction with the Laboratory of Image Informatics at Stanford University. The specific problem that will be addressed in this work is that most of the semantic information about radiological images are not captured and related to them using terms of biomedical ontologies and standards, such as RadLex or DICOM, what makes it impossible to automatic evaluate them by computers, to search for them in hospital databases using semantic criteria, etc. To address this issue, radiologists need an easy, intuitive and affordable computational solution to add this semantic information. In this work, a web solution for adding the information was developed, the ePAD system. It allows the retrieval of medical images, such as images available in hospital information systems (PACS), the creation of contours around tumor lesions, the association of ontological terms to these contours, and the storage of this terms in a knowledge base. The main challenges of this work involved the creation of intuitive interfaces using Rich Internet Applications technology and the operation from a standard Web Browser. The first functional prototype of ePAD reached its goal of proving its technical feasibility. It was able to do the same basic annotation job of desktop applications, such as OsiriX-iPad, without the same overhead. It also showed to the medical community that it was a useful tool and that generated interest of potential early users

Coding Electronic Medical Records (EMRs) with diagnosis and procedure codes is an essential task for billing, secondary data analyses, and monitoring health trends. Both speed and accuracy of coding are critical. While coding errors could lead to more patient-side financial burden and misinterpretation of a patient’s well-being, timely coding is also needed to avoid backlogs and additional costs for the healthcare facility. Therefore, it is necessary to develop automated diagnosis and procedure code recommendation methods that can be used by professional medical coders. The main difficulty with developing automated EMR coding methods is the nature of the label space. The standardized vocabularies used for medical coding contain over 10 thousand codes. The label space is large, and the label distribution is extremely unbalanced - most codes occur very infrequently, with a few codes occurring several orders of magnitude more than others. A few codes never occur in training dataset at all. In this work, we present three methods to handle the large unbalanced label space. First, we study how to augment EMR training data with biomedical data (research articles indexed on PubMed) to improve the performance of standard neural networks for text classification. PubMed indexes more than 23 million citations. Many of the indexed articles contain relevant information about diagnosis and procedure codes. Therefore, we present a novel method of incorporating this unstructured data in PubMed using transfer learning. Second, we combine ideas from metric learning with recent advances in neural networks to form a novel neural architecture that better handles infrequent codes. And third, we present new methods to predict codes that have never appeared in the training dataset. Overall, our contributions constitute advances in neural multi-label text classification with potential consequences for improving EMR coding.

abstract: In the current millennium, extensive use of computers and the internet caused an exponential increase in information. Few research areas are as important as information extraction, which primarily involves extracting concepts and the relations between them from free text. Limitations in the size of training data, lack of lexicons and lack of relationship patterns are major factors for poor performance in information extraction. This is because the training data cannot possibly contain all concepts and their synonyms; and it contains only limited examples of relationship patterns between concepts. Creating training data, lexicons and relationship patterns is expensive, especially in the biomedical domain (including clinical notes) because of the depth of domain knowledge required of the curators. Dictionary-based approaches for concept extraction in this domain are not sufficient to effectively overcome the complexities that arise because of the descriptive nature of human languages. For example, there is a relatively higher amount of abbreviations (not all of them present in lexicons) compared to everyday English text. Sometimes abbreviations are modifiers of an adjective (e.g. CD4-negative) rather than nouns (and hence, not usually considered named entities). There are many chemical names with numbers, commas, hyphens and parentheses (e.g. t(3;3)(q21;q26)), which will be separated by most tokenizers. In addition, partial words are used in place of full words (e.g. up- and downregulate); and some of the words used are highly specialized for the domain. Clinical notes contain peculiar drug names, anatomical nomenclature, other specialized names and phrases that are not standard in everyday English or in published articles (e.g. "l shoulder inj"). State of the art concept extraction systems use machine learning algorithms to overcome some of these challenges. However, they need a large annotated corpus for every concept class that needs to be extracted. A novel natural language processing approach to minimize this limitation in concept extraction is proposed here using distributional semantics. Distributional semantics is an emerging field arising from the notion that the meaning or semantics of a piece of text (discourse) depends on the distribution of the elements of that discourse in relation to its surroundings. Distributional information from large unlabeled data is used to automatically create lexicons for the concepts to be tagged, clusters of contextually similar words, and thesauri of distributionally similar words. These automatically generated lexical resources are shown here to be more useful than manually created lexicons for extracting concepts from both literature and narratives. Further, machine learning features based on distributional semantics are shown to improve the accuracy of BANNER, and could be used in other machine learning systems such as cTakes to improve their performance. In addition, in order to simplify the sentence patterns and facilitate association extraction, a new algorithm using a "shotgun" approach is proposed. The goal of sentence simplification has traditionally been to reduce the grammatical complexity of sentences while retaining the relevant information content and meaning to enable better readability for humans and enhanced processing by parsers. Sentence simplification is shown here to improve the performance of association extraction systems for both biomedical literature and clinical notes. It helps improve the accuracy of protein-protein interaction extraction from the literature and also improves relationship extraction from clinical notes (such as between medical problems, tests and treatments). Overall, the two main contributions of this work include the application of sentence simplification to association extraction as described above, and the use of distributional semantics for concept extraction. The proposed work on concept extraction amalgamates for the first time two diverse research areas -distributional semantics and information extraction. This approach renders all the advantages offered in other semi-supervised machine learning systems, and, unlike other proposed semi-supervised approaches, it can be used on top of different basic frameworks and algorithms. / Dissertation/Thesis / Ph.D. Biomedical Informatics 2011

Este trabalho faz parte de um projeto maior, o Annotation and Image Markup Project, que tem o objetivo de criar uma base de conhecimento médico sobre imagens radiológicas para identificação, acompanhamento e reasoning acerca de lesões tumorais em pesquisas sobre câncer e consultórios médicos. Esse projeto está sendo desenvolvido em conjunto com o Radiological Sciences Laboratory da Stanford University. O problema específico, que será abordado nesse trabalho, é que a maior parte das informações semânticas sobre imagens radiológicas não são capturados e relacionados às mesmas usando termos de ontologias biomédicas e padrões, como o RadLex e DICOM, o que impossibilita a sua avaliação automática por computadores, busca em arquivos médicos em hospitais, etc. Para tratar isso, os radiologistas precisam de uma solução computacional fácil, intuitiva e acessível para adicionar essas informações. Nesse trabalho foi desenvolvida uma solução Web para inclusão dessas anotações, o sistema ePAD. O aplicativo permite a recuperação de imagens médicas, como as imagens disponíveis em sistemas de informação hospitalares (PACS), o delineamento dos contornos de lesões tumorais, a associação de termos ontológicos a esses contornos e o armazenamento desses termos em uma base de conhecimento. Os principais desafios desse trabalho envolveram a aplicação de interfaces intuitivas baseadas em Rich Internet Applications e sua operação a partir de um navegador Web padrão. O primeiro protótipo funcional do ePAD atingiu seus objetivos ao demonstrar sua viabilidade técnica, sendo capaz de executar o mesmo trabalho básico de anotação de aplicações Desktop, como o OsiriX-iPad, sem o mesmo overhead. Também mostrou a sua utilidade a comunidade médica o que gerou o interesse de usuários potenciais / This work is a part of a larger project, the Annotation and Markup Project, which aims to create a medical knowledge base about radiological images to identify, monitor and reason about tumors in cancer research and medical practices. This project is being developed in conjunction with the Laboratory of Image Informatics at Stanford University. The specific problem that will be addressed in this work is that most of the semantic information about radiological images are not captured and related to them using terms of biomedical ontologies and standards, such as RadLex or DICOM, what makes it impossible to automatic evaluate them by computers, to search for them in hospital databases using semantic criteria, etc. To address this issue, radiologists need an easy, intuitive and affordable computational solution to add this semantic information. In this work, a web solution for adding the information was developed, the ePAD system. It allows the retrieval of medical images, such as images available in hospital information systems (PACS), the creation of contours around tumor lesions, the association of ontological terms to these contours, and the storage of this terms in a knowledge base. The main challenges of this work involved the creation of intuitive interfaces using Rich Internet Applications technology and the operation from a standard Web Browser. The first functional prototype of ePAD reached its goal of proving its technical feasibility. It was able to do the same basic annotation job of desktop applications, such as OsiriX-iPad, without the same overhead. It also showed to the medical community that it was a useful tool and that generated interest of potential early users

Orientador: Saide Jorge Calil / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-08T01:44:24Z (GMT). No. of bitstreams: 1 Botelho_MariaLuciadeAzevedo_D.pdf: 2737756 bytes, checksum: 9bd75c71b8b603eca3b3c8d5482e2050 (MD5) Previous issue date: 2006 / Resumo: Um detalhado levantamento de programas de computador aplicados ao ensino foi realizado tendo como objetivo conhecer os recursos disponíveis no mercado. Paralelamente, foi feita uma pesquisa bibliográfica buscando as necessidades da comunidade acadêmica em termos de recursos computacionais. Foi constatado, então, que apesar de serem consideradas importantes, havia poucas alternativas para a realização de aulas virtuais que demandassem pequeno esforço operacional e recursos simples de infra-estrutura para utilização. O objetivo do trabalho foi definido então, como sendo desenvolver e avaliar um sistema que viabiliza a realização de aulas virtuais de dois tipos, as on-line e as off-line, exigindo pouca experiência de informática dos usuários, e adequado aos recursos mais comumente disponíveis nas universidades. Alguns requisitos básicos de funcionalidade foram definidos, visando dotar o sistema com máxima facilidade de operação, como por exemplo, interface padrão, ou seja, aparência idêntica em todas as operações, ajuda disponível em todos os níveis e permitir aproveitamento de shows de slides e textos já existentes. Foram elaboradas duas plataformas de trabalho, uma para o professor, que permite criar, alterar e realizar uma aula, e outra plataforma para o aluno, que permite assistir à aula. A avaliação do sistema foi realizada através da execução de dois Planos de Testes, que utilizaram instrumentos padronizados como os Critérios de Avaliação de Qualidade de Software, aos quais foram atribuídas notas, e os Questionários de Avaliação do sistema, que foram preenchidos pelos professores e pelos alunos envolvidos. As aulas off-line obtiveram notas máximas em todos os quesitos, e as aulas on-line obtiveram médias acima de 1,78 (numa escala de 0 a 2). Todos os professores responderam que gostaram de realizar as aulas utilizando o sistema; 75% disseram que gostariam de empregá-lo em seu trabalho e 25% disseram que talvez pudessem utilizá-lo. Dentre os alunos, somente 2,33% responderam que não gostaram da aula virtual, e 4,65% informaram que não gostariam de ter mais aulas realizadas com o sistema no seu curso. Foi desenvolvida uma Discussão sobre os motivos que resultaram no pior desempenho das aulas on-line, e a principal causa detectada foi a dificuldade de realização deste tipo de evento utilizando a Internet comercial, que apresenta problemas de grande volume de tráfego de dados. Dentre as conclusões apresentadas, destaca-se que o VirtuAula é uma interessante alternativa para instituições de ensino público brasileiras, pois sua aplicação é original, não se encontrando similares nacionais com todas as funcionalidades reunidas, e por ter baixo custo operacional, não apresentando ônus nem risco de contravenção, por haver a possibilidade de cessão gratuita de uso / Abstract: A detailed search for digital programs applied to teaching processes was performed to identify the available resources on the world market. It was also carried out a survey on public and private libraries looking for the requirements of the academic community regarding computational resources to such purpose. It was found that although considered important, there were few alternatives for the development of virtual classes that demands little operational effort as well as a simplified infrastructure for its use. The goal of this work is to develop and assess a system - VirtuAula - to assemble and present on-line and off-line virtual classes, requiring low experience on informatics for its users as well as being adequate to common resources available in universities. The basic functional requirements defined for developing the VirtuAula were a standard interface, which means identical browse for all operations, a user-friendly help desk and the possibility to use already prepared slide shows and texts. Two work platforms were elaborated, one for teachers, which allow them to create, change and carry out the virtual class, and a second platform for students to attend this virtual class. For the system assessment two tests plans were used; standard tools as the Evaluation Criteria of the Software Quality (grading method), and questionnaires that were filled by the involved teachers and students. The off-line classes reached the maximum grading score in all the evaluation topics, while on-line classes reached an average score over 1,78 (in a 0-2 scale). All the involved teachers answered that they would like to carry out virtual classes using this system; 75% declared that they would like to use it for their work, and 25% declared that they could use it. Among the students, only 2,33% dislike the virtual classes using the VirtuAula while 4,65% informed that they would not like to have such kind of classes in their courses. Looking for the reas ons for the lower performance of on-line classes in this survey, the major cause was the difficulty to carry out such event on the present commercial Internet system due to its low performance during very heavy data transfer. Among the conclusions presented here, it can be depicted that the system is an interesting alternative tool for public schools in Brazil due to its originality (no similar software), low cost and user free possibility / Doutorado / Engenharia Biomedica / Doutor em Engenharia Elétrica