Made available in DSpace on 2017-07-10T17:11:51Z (GMT). No. of bitstreams: 1 WILSON JUNG.pdf: 3777203 bytes, checksum: 02dd354bc8c0d25187fd3960d5d56152 (MD5) Previous issue date: 2012-04-25 / Lately, many human knowledge fields use computer systems to support data management which are the foundation to the decision making process. Data Quality (DQ) is a key feature whose absence can undermine the usefulness of the information and the processes that use it. There can be found in the literature several cases of DQ problems with impact in many areas, resulting in economic and social losses. Therefore, DQ research aims to study data problems causes and proposes assessment methods and processes to assist in quality assurance. In healthcare, data constitutes an important element used as the basis for applying medical treatments and procedures to patients, thus requiring a high quality level. The data is also used in the research and application of computational knowledge discovery methods, such as Data Mining. Therefore, the goal of this work is to study the implementation of principles to assist DQ guarantee during the medical software development. This goal motivated the development of a case study related to Coloproctology, in which a surgery data management system prototype was de- veloped in partnership with the Coloproctology Service of FCM - UNICAMP. The interaction with domain experts was a key factor during the development process, providing the adequate data structure modeling that composes the system. A module to monitor specific data problems has also been incorporated into the prototype to assist the appropriate information insertion as much as the control of patients records which have DQ problems. The prototype has been evaluated by computer and healthcare s colaborators, who, after using the system, answered to a qualitative DQ assessment form. The assessment s results pointed out the prototype suitability to the activities it is aimed for, guided specific functionalities review and may support the proposed software evolution and future related work. / Atualmente, diversas áreas do conhecimento humano fazem uso de sistemas computacionais para auxiliar no gerenciamento de dados, que são a base para o processo de tomada de decisão. A Qualidade de Dados (QD) constitui uma característica fundamental cuja ausência pode comprometer a utilidade da informação e os processos que a utilizam. Na literatura são apresentados diversos casos que relatam o impacto de problemas de QD nas mais diversas áreas, represen- tando perdas econômicas e sociais. Assim, a área de QD visa o estudo das causas de problemas nos dados e a proposição de métodos de avaliação e processos que auxiliem na garantia da qualidade. Na área da saúde os dados constituem elementos importantes que são utilizados como base para a aplicação de tratamentos e procedimentos médicos aos pacientes, fatores que exigem um nível elevado de qualidade. Esses dados também são utilizados em pesquisas e aplicações de métodos computacionais de extração de conhecimento, como a Mineração de Dados. Assim, o objetivo deste trabalho consiste em estudar a aplicação de princípios que auxiliem na garantia da QD durante o desenvolvimento de um sistema computacional médico. Tal objetivo motivou a realização de um estudo de caso relacionado à especialidade da Coloproctologia, no qual foi desenvolvido o protótipo de um sistema para gerenciamento de dados de cirurgia coloproctológica em parceria com o Serviço de Coloproctologia da FCM - UNICAMP. A interação com os especialistas de domínio constituiu um fator fundamental durante o processo de desenvolvimento, possibilitando a modelagem adequada da estrutura dos dados que forma o sistema. Também foi incorporado ao protótipo um módulo para monitoramento de problemas específicos nos dados, auxiliando tanto no preenchimento adequado da informação quanto no controle dos registros de pacientes que apresentam problemas de QD. Ao final, o protótipo foi subme- tido à avaliação por colaboradores da área da computação e da saúde, que após a utilização do sistema responderam a um formulário para avaliação qualitativa de QD. Os resultados da avaliação indicaram a adequação do protótipo para as atividades a que é destinado, orientaram para a revisão de funcionalidades específicas e poderão auxiliar na evolução do sistema proposto e em trabalhos futuros.

À ce jour, les mesures de risque des cancers ou d’efficacité de leur suivi, se font à partir de recueils de données médicales spécifiques initiés par les médecins épidémiologistes. Ces recueils disposent néanmoins de certaines limites : perte d’information, biais de déclaration, absence de données pour un risque non connu, biais de mesure (par exemple pour les données de nature médico-économiques). Le partage sécurisé de données médicales entre différentes structures médicales publiques et/ou privées est à ce jour en pleine mutation technologique. Les technologies proposées doivent rendre possible un partage électronique et sécurisé de ces données de manière à les rendre disponible à tout instant dans le cadre de l’observation sanitaire à l’évaluation de prises en charge ou de politiques de santé. Pour répondre à ces besoins, l’infrastructure GINSENG se base sur des informations produites dans le cadre des soins, sans nouvelles modalités de recueil, permettant à la fois une vitesse d’accès à l’information et une exhaustivité accrue. Ce recueil se fait par ailleurs avec de meilleures garanties d’anonymat et un chaînage de l’information médicale pour chaque patient. Une autorisation de la CNIL a été octroyée à l’infrastructure informatique du projet ainsi qu’à son utilisation pour le suivi des cancers en octobre 2013. Depuis le portail web e-ginseng.com, les médecins habilités s’authentifient grâce à leur Carte de Professionnel de Santé (CPS). Chaque patient, dont les données médicales sont réparties dans les établissements de santé, est identifié avec son accord, par les attributs suivants : nom, prénom, année et mois de naissance ainsi que son code postal de résidence avant d’être assigné à un numéro d’identification unique et anonyme. La mise à jour des données médicales de chaque patient est réalisée une fois par semaine ; chaque médecin peut alors consulter toutes les informations médicales relatives à chaque patient par une simple connexion au réseau. Ces informations lui apparaissent sous forme d’une arborescence d’évènements médicaux. Par exemple, un médecin chargé du suivi des patients dans le cadre du dépistage organisé pourra accéder directement depuis le portail web aux informations médicales dont il aura besoin pour établir une fiche médicale exhaustive du parcours du patient pour lequel un cancer aurait été détecté ou bien une suspicion de cancer qui se serait avérée négative suite à plusieurs examens médicaux. Un médecin épidémiologiste peut également réaliser des requêtes statistiques d’envergure sur les données médicales afin de répondre à des questions d’intérêt en santé publique. Pour aller plus loin, les requêtes épidémiologiques lancées sur les données médicales peuvent être couplées à des informations d’utilité publique recueillies sur d’autres bases de données en accès libre sur internet. L’infrastructure informatique GINSENG est actuellement déployée pour le suivi des cancers en région Auvergne entre les structures de gestion du dépistage organisé du cancer (SGDO) et le cabinet d’anatomie et cytologie pathologiques (ACP) Sipath-Unilabs. Le recours à un hébergeur de données de santé (HADS), nommé Informatique de sécurité (IDS), est également proposé pour le stockage des informations confidentielles des patients. Cette infrastructure permet actuellement de collecter toutes les informations médicales d’intérêt pour le suivi des cancers et l’évaluation des pratiques médicales. Les équipes de bio-statistiques et de santé publique du CHU de Clermont-Ferrand établissent actuellement les analyses épidémiologiques d’intérêt à partir des données collectées par le réseau. / The implementation of a grid network to support large-scale epidemiology analysis (based on distributed medical data sources) and medical data sharing require medical data integration and semantic alignment. In this thesis, we present the GINSENG (Global Initiative for Sentinel eHealth Network on Grid) network that federates existing Electronic Health Records through a rich metamodel (FedEHR), a semantic data model (SemEHR) and distributed query toolkits. A query interface based on the VIP platform, and available through the e-ginseng.com web portal helps medical end-users in the design of epidemiological studies and the retrieval of relevant medical data sets.

Abstract Developments in information technology have had a big impact in healthcare, producing vast amounts of data and increasing demands associated with their secure transfer, storage and analysis. To serve them, biomedical data need to carry patient information and records or even extra biomedical images or signals required for multimodal applications. The proposed solution is to host this information in data using data hiding algorithms through the introduction of imperceptible modifications achieving two main purposes: increasing data management efficiency and enhancing the security aspects of confidentiality, reliability and availability. Data hiding achieve this by embedding the payload in objects, including components such as authentication tags, without requirements in extra space or modifications in repositories. The proposed methods satisfy two research problems. The first is the hospital-centric problem of providing efficient and secure management of data in hospital networks. This includes combinations of multimodal data in single objects. The host data were biomedical images and sequences intended for diagnoses meaning that even non-visible modifications can cause errors. Thus, a determining restriction was reversibility. Reversible data hiding methods remove the introduced modifications upon extraction of the payload. Embedding capacity was another priority that determined the proposed algorithms. To meet those demands, the algorithms were based on the Least Significant Bit Substitution and Histogram Shifting approaches. The second was the patient-centric problem, including user authentication and issues of secure and efficient data transfer in eHealth systems. Two novel solutions were proposed. The first method uses data hiding to increase the robustness of face biometrics in photos, where due to the high robustness requirements, a periodic pattern embedding approach was used. The second method protects sensitive user data collected by smartphones. In this case, to meet the low computational cost requirements, the method was based on Least Significant Bit Substitution. Concluding, the proposed algorithms introduced novel data hiding applications and demonstrated competitive embedding properties in existing applications. / Tiivistelmä Modernit terveydenhuoltojärjestelmät tuottavat suuria määriä tietoa, mikä korostaa tiedon turvalliseen siirtämiseen, tallentamiseen ja analysointiin liittyviä vaatimuksia. Täyttääkseen nämä vaatimukset, biolääketieteellisen tiedon täytyy sisältää potilastietoja ja -kertomusta, jopa biolääketieteellisiä lisäkuvia ja -signaaleja, joita tarvitaan multimodaalisissa sovelluksissa. Esitetty ratkaisu on upottaa tämä informaatio tietoon käyttäen tiedonpiilotusmenetelmiä, joissa näkymättömiä muutoksia tehden saavutetaan kaksi päämäärää: tiedonhallinnan tehokkuuden nostaminen ja luottamuksellisuuteen, luotettavuuteen ja saatavuuteen liittyvien turvallisuusnäkökulmien parantaminen. Tiedonpiilotus saavuttaa tämän upottamalla hyötykuorman, sisältäen komponentteja, kuten todentamismerkinnät, ilman lisätilavaatimuksia tai muutoksia tietokantoihin. Esitetyt menetelmät ratkaisevat kaksi tutkimusongelmaa. Ensimmäinen on sairaalakeskeinen ongelma tehokkaan ja turvallisen tiedonhallinnan tarjoamiseen sairaaloiden verkoissa. Tämä sisältää multimodaalisen tiedon yhdistämisen yhdeksi kokonaisuudeksi. Tiedon kantajana olivat biolääketieteelliset kuvat ja sekvenssit, jotka on tarkoitettu diagnosointiin, missä jopa näkymättömät muutokset voivat aiheuttaa virheitä. Siispä määrittävin rajoite oli palautettavuus. Palauttavat tiedonpiilotus-menetelmät poistavat lisätyt muutokset, kun hyötykuorma irrotetaan. Upotuskapasiteetti oli toinen tavoite, joka määritteli esitettyjä algoritmeja. Saavuttaakseen nämä vaatimukset, algoritmit perustuivat vähiten merkitsevän bitin korvaamiseen ja histogrammin siirtämiseen. Toisena oli potilaskeskeinen ongelma, joka sisältää käyttäjän henkilöllisyyden todentamisen sekä turvalliseen ja tehokkaaseen tiedonsiirtoon liittyvät haasteet eHealth-järjestelmissä. Työssä ehdotettiin kahta uutta ratkaisua. Ensimmäinen niistä käyttää tiedonpiilotusta parantamaan kasvojen biometriikan kestävyyttä valokuvissa. Korkeasta kestävyysvaatimuksesta johtuen käytettiin periodisen kuvion upottamismenetelmää. Toinen menetelmä suojelee älypuhelimien keräämää arkaluontoista käyttäjätietoa. Tässä tapauksessa, jotta matala laskennallinen kustannus saavutetaan, menetelmä perustui vähiten merkitsevän bitin korvaamiseen. Yhteenvetona ehdotetut algoritmit esittelivät uusia tiedonpiilotussovelluksia ja osoittivat kilpailukykyisiä upotusominaisuuksia olemassa olevissa sovelluksissa.

This thesis discusses rendering capabilities of web browsers of accelerated 3D scene rendering. It specifically deals with direct volumetric medical data visualization. It focuses on the usage of ray casting algorithm, its quality and its realistic rendering options. One of the goals was to create an application that demonstrates the ability to render three-dimensional volume data in a web browser using WebGL. The application is written in JavaSript and its 3D rendering core uses the Three.js library.

This Thesis deals with the problems of medical image data management. It analyzes a metadata extraction problem from medical format DICOM and their efficient saving. System is designed as a client-server application. System provides an interface for DICOM data storing, interface for searching, data editing and exporting. System cooperatives with the collaborative system VCE. Implementation was carried out in C++ and PHP language with support of SQLite and MySQL database.

In the context of health improving, the measurement of vital parameters such as heart rate (HR) can provide solutions for health monitoring, prevention and screening for certain chronic diseases. Among the different technologies for HR measuring, photoplethysmography (PPG) technique embedded in smart watches is the most commonly used in the field of consumer electronics since it is comfortable and does not require any user intervention. To be able to provide an all day and night long HR monitoring method, difficulties associated with PPG signals vulnerability to Motion Artifact (MA) must be overcome. Conventional signal processing solutions (power spectral density analysis) have limited generalization capability as they are specific to certain types of movements, highlighting the interest of machine learning tools, particularly deep learning (DL). Since DL models in the literature are trained on data from a different sensor than the internal sensor, transfer learning may prove unsuccessful. This work proposes a DL approach for estimating HR from wrist PPG signals. The model is trained on internal data with a greater demographic diversity of participants. This project also illustrates the contribution of multi-path and multi-wavelength PPG instead of the conventional single green PPG solution. This work presents several models, called DeepTime, with selected input channels and wavelengths: Mono_Green, Multi_Green, Multi_Wavelength, and Multi_Channel_Multi_Wavelength. They take temporal PPG signals as inputs along with 3D acceleration and provide HR estimation every 2 seconds with an 8-second initialization. This convolutional neural network comprised of several input branches improves the existing Withings internal method’s overall Mean Absolute Error (MAE) from 9.9 BPM to 6.9 BPM on the holdout test set. This work could be completed and improved by adding signal temporal history using recurrent layers, such as Long-Short-Term-Memory (LSTM), training the model with a bigger dataset, improving preprocessing steps or using a more elaborate loss function that includes a trust score. / I sammanhanget av förbättring av hälsouppföljning kan mätning av vitala parametrar som hjärtfrekvens (HR) erbjuda lösningar för förebyggande och screening av vissa kroniska sjukdomar. Bland olika tekniker för mätning av HR är fotoplethysmografi (PPG) integrerad i smartklockor den vanligast använda inom elektronikområdet eftersom den är bekväm och inte kräver något användaringripande. För att erbjuda en kontinuerlig HRövervakningsmetod utgör sårbarheten hos PPG-signaler för rörelseartefakter (MA) en stor utmaning. Konventionella signalbehandlingslösningar (analys av effektspektraltäthet) har begränsad generaliseringsförmåga eftersom de är specifika för vissa typer av rörelser, vilket betonar intresset för maskininlärningsverktyg, särskilt djupinlärning (DL). Eftersom DL-modeller i litteraturen tränas på data från en annan sensor än den interna sensorn kan överföringsinlärning vara misslyckad. Detta arbete föreslår en DL-ansats för att uppskatta HR från PPG-signaler på handleden. Modellen tränas på interna data med en större demografisk mångfald bland deltagarna. Detta projekt illustrerar även bidraget från flervägs- och flervågs-PPG istället för den konventionella enkla gröna PPG-lösningen. Detta arbete presenterar flera modeller, kallade DeepTime, med utvalda ingångskanaler och våglängder: Mono_Green, Multi_Green, Multi_Wavelength och Multi_Channel_Multi_Wavelength. De tar in temporära PPG-signaler tillsammans med 3D-acceleration och ger HR-uppskattning var 2:a sekund med en initialisering på 8 sekunder. Detta konvolutionella neurala nätverk, som består av flera ingångsgrenar, förbättrar den totala medelabsoluta felet (MAE) från 9,9 BPM (befintlig intern metod) till 6,9 BPM på testuppsättningen. Detta arbete kan kompletteras och förbättras genom att integrera den temporala historiken hos signalen med hjälp av återkommande lager (som LSTM), träna modellen på mer data, förbättra förbehandlingsstegen eller välja en mer sofistikerad förlustfunktion som inkluderar ett konfidensvärde.

Diss. (sammanfattning) Umeå : Umeå universitet, 2009. / Härtill 4 uppsatser. Även tryckt utgåva.

Το αντικείμενο της διατριβής αυτής είναι η εισαγωγή ενός νέου πρωτοκόλλου (e-SCP-ECG+) με στόχο την μεταφορά και διαχείριση πολλαπλών τύπων πληροφορίας που προέρχονται από ιατρικές συσκευές συλλογής ζωτικών σημάτων, δεδομένα που αφορούν τις αλλεργίες από τις οποίες υποφέρει ο ασθενής, στοιχεία γεωτοποθεσίας, καθώς επίσης και δημογραφικών στοιχείων, από τους ασθενείς σε υπολογιστικούς σταθμούς επεξεργασίας, διαχείρισης και αποθήκευσής της. Ορίζεται επίσης η αρχιτεκτονική ενός Συστήματος Τηλεπαρακολούθησης Υγείας Ασθενούς (ΣΤΥΑ), το οποίο χρησιμοποιεί το πρωτόκολλο e-SCP-ECG+ για τη μεταφορά, τη διαχείριση και την αρχειοθέτηση της συλλεγόμενης πληροφορίας. Η αρχιτεκτονική περιλαμβάνει, επίσης, τη δημιουργία ενός Δικτύου από ΣΤΥΑ, με στόχο την δικτυακή αναζήτηση πληροφορίας σχετικής με τον ασθενή, εξασφαλίζοντας έτσι τη δυνατότητα του ελέγχου της πορείας της υγείας ενός ασθενούς. Το ΣΤΥΑ πέρα από την λειτουργία του σε εργαστηριακό επίπεδο, δοκιμάστηκε πιλοτικά σε πραγματικές συνθήκες. / This dissertation introduces a new protocol named e-SCP-ECG+, which permits the transport and management of multiple information types collected from patients (vital signs, citizen demographic data, other information relative with the treated incident, allergy data, geolocation data, etc.), through a communication network to a Health Reception Center. The dissertation also defines the architecture of a Health Tele-monitoring System (HTS) aiming to protocol’s application and evaluation. The pilot HTS, uses the protocol e-SCP-ECG+, in order to transmit, manage and archive the collected information. The creation of an HTS’s Network is also included in this architecture. This network supports health continuity and gives doctor the ability to search information relative to the patient between different networked HTSs. The pilot HTS, has been tested both on laboratory conditions and in real-world operation.

In Sweden, there is currently no consensus addressing the screening, diagnostics and treatment of gestational diabetes mellitus (GDM). In addition, there is little knowledge on the impact of GDM on the daily life of pregnant women and the experiences of health care professionals providing maternal health care to women with GDM. Using different perspectives, this thesis examines the experiences of GDM and the performance of screening for GDM in a regional context in Sweden. The studies used qualitative and quantitative methods. In the qualitative studies, grounded theory was applied in two studies and qualitative content analysis in one study. In the quantitative study, a combination of questionnaire data and data from medical records of pregnancy and birth were processed. Surprisingly, screening for GDM was reduced despite local clinical guidelines stipulating the risk factors indicating an OGTT. Furthermore, the prevalence of the risk factors for GDM in the population investigated was almost doubled compared to previous Swedish studies. Pregnant women developing risk factors for GDM during pregnancy were found to be at substantially increased risk of giving birth to an infant with macrosomia. The experiences of pregnant women with GDM revealed that being diagnosed with and living with GDM during pregnancy might be understood as a process ‘from stun to gradual balance’. The experience comprised both negative and positive dimensions. Despite the challenges, the inconveniences and the changes involved, gradually adapting to an altered lifestyle and finding their balance in daily life was ‘the prize’ the women ‘were willing to pay’ to secure optimal maternal and foetal health. The experiences of midwives comprised managing conflicting demands providing antenatal care to pregnant women diagnosed with GDM. Most midwives felt the obligation to control and monitor the complicated pregnancy, to initiate and motivate the recommended changes in life style together with providing an empowering and caring relation with the women. These assignments disclosed complex conflicting situations and the midwives appeared to choose strategy for managing the situation depending on their perception of the circumstances. The experiences of the obstetricians were understood as ‘dealing with ambiguity’. The ambiguity permeated all aspects of working as an obstetrician within the maternal health care counselling women with GDM: the role of the obstetrician, the context of the organization, balancing the multifaceted interests of the maternal and foetal conditions and the lack of consensus, recommendations and evidence-based knowledge.   The studies revealed the complexity of the situation for the affected pregnant women as well as for the health care professionals providing antenatal care to women diagnosed with GDM. Furthermore, the performance of screening of GDM in pregnant women with risk factors for GDM was insufficient in the investigated region. The findings in this thesis may be useful to increase knowledge of the experiences of pregnant women living with or managing GDM. The findings may also be useful when planning for improvements of maternal health care directed to pregnant women diagnosed with GDM during pregnancy.

A documentação clínica requer a representação de situações complexas como pareceres clínicos, imagens e resultados de exames, planos de tratamento, dentre outras. Entre os profissionais da área de saúde, a linguagem natural é o meio principal de documentação. Neste tipo de linguagem, caracterizada por uma elevada flexibilidade sintática e léxica, é comum a prevalência de ambigüidades em sentenças e termos. O objetivo do presente trabalho consiste em mapear informações codificadas em narrativas clínicas para uma ontologia de domínio (SNOMED CT). Para sua consecução, aplicaram-se ferramentas processamento de linguagem natural (PLN), assim como adotaram-se heurísticas para o mapeamento de textos para ontologias. Para o desenvolvimento da pesquisa, uma amostra de sumários de alta foi obtida junto ao Hospital das Clínicas de Porto Alegre, RS, Brasil. Parte dos sumários foi manualmente anotada, com a aplicação da estratégia de Active Learning, visando a preparação de um corpus para o treinamento de ferramentas de PLN. Paralelamente, foram desenvolvidos algoritmos para o pré-processamento dos textos ‘sujos’ (com grande quantidade de erros, acrônimos, abreviações, etc). Com a identificação das frases nominais, resultado do processamento das ferramentas de PLN, diversas heurísticas (identificação de acrônimos, correção ortográfica, supressão de valores numéricos e distância conceitual) para o mapeamento para a SNOMED CT foram aplicadas. A versão atual da SNOMED CT não está disponível em português, demandando o uso de ferramentas para processamento multi-lingual. Para tanto, o pesquisa atual é parte da iniciativa do projeto MorphoSaurus, por meio do qual desenvolve-se e disponibiliza-se um thesaurus multi-língue (português, alemão, inglês, espanhol, sueco, francês), bem como componentes de software que permitem o processamento inter-lingual. Para realização da pesquisa, 80% da base de sumários foi analisada e manualmente anotada, resultando em um corpus de domínio (textos médicos e em português) que permitiu a especialização do software OpenNLP (baseado no modelo estatístico para o PLN e selecionado após a avaliação de outras soluções disponíveis). A precisão do etiquetador atingiu 93.67%. O thesaurus multi-língue do MorphoSaurus foi estendido, reestruturado e avaliado (automaticamente com a comparação por meio de textos comparáveis – ‘traduções de um mesmo texto para diferentes idiomas’) e sofreu intervenções objetivando a correção de imperfeições existentes, resultando na melhoria da cobertura lingüística, no caso do português, em 2%; e 50% para o caso do espanhol, medidas obtidas por meio do levantamento das curvas de precisão e revocação para a base do OHSUMED. Por fim, a codificação de informações de narrativas clínicas para uma ontologia de domínio é uma área de elevado interesse científico e clínico, visto que grande parte dos dados produzidos quando do atendimento médico é armazenado em texto livre e não em campos estruturados. Para o alcance deste fim, adotou-se a SNOMED CT. A viabilidade da metodologia de mapeamento foi demonstrada com a avaliação dos resultados do mapeamento automático contra um padrão ouro, manualmente desenvolvido, indicando precisão de 83,9%. / Clinical documentation requires the representation of fine-grained descriptions of patients' history, evolution, and treatment. These descriptions are materialized in findings reports, medical orders, as well as in evolution and discharge summaries. In most clinical environments natural language is the main carrier of documentation. Written clinical jargon is commonly characterized by idiosyncratic terminology, a high frequency of highly context-dependent ambiguous expressions (especially acronyms and abbreviations). Violations of spelling and grammar rules are common. The purpose of this work is to map free text from clinical narratives to a domain ontology (SNOMED CT). To this end, natural language processing (NLP) tools will be combined with a heuristic of semantic mapping. The study uses discharge summaries from the Hospital das Clínicas de Porto Alegre, RS, Brazil. Parts of these texts are used for creating a training corpus, using manual annotation supported by active learning technology, used for the training of NLP tools that are used for the identification of parts of speech, the cleansing of "dirty" text passages. Thus it was possible to obtain relatively well-formed and unambiguous noun phrases, heuristics was implemented to semantic mapping between these noun phrases (in Portuguese) and the terms describing the SNOMED CT concepts (English and Spanish) uses the technology of morphosemantic indexing, using a multilingual subword thesaurus, provided by the MorphoSaurus system, the resolution of acronyms, and the identification of named entities (e.g. numbers). In this study, 80 per cent of the summaries are analyzed and manually annotated, resulting in a domain corpus that supports the specialization of the OpenNLP system, mainly following the paradigm of statistical natural language processing (the accuracy of the tagger obtained was 93.67%). Simultaneously, several techniques have been used for validating and improving the subword thesaurus. To this end, the semantic representations of comparable test corpora from the medical domain in English, Spanish, and Portuguese were compared with regard to the relative frequency of semantic identifiers, improving the corpus coverage (2% to Portuguese, and 50% to Spanish). The result was used as an input by a team of lexicon curators, which continuously fix errors and fill gaps in the trilingual thesaurus underlying the MorphoSaurus system. The progress of this work could be objectified using OHSUMED, a standard medical information retrieval benchmark. The mapping of text-encoded clinical information to a domain ontology constitutes an area of high scientific and practical interest due to the need for the analysis of structured data, whereas the clinical information is routinely recorded in a largely unstructured way. In this work the ontology used was SNOMED CT. The evaluation of mapping methodology indicates accuracy of 83.9%.