In the complex landscape of healthcare, patients navigate through various institutions from hospitals to long-term care facilities, and each step of their journey plays a crucial role in their disease progression and treatment plan. Traditional analyses often focus on individual transitions, offering limited insight into the broader picture of patient care and disease progression. This thesis aims to explore the entire sequence of patient transitions within the Canadian healthcare system to uncover meaningful patterns and commonalities. This research employs an innovative approach to leveraging the Canadian Institute for Health Information (CIHI) dataset, consisting of around 250,000 patient records after data cleaning and including approximately 10-11 variables. Extracting a diverse category of features, such as temporal, semantic, and clinical information, constructs a detailed profile for each patient journey. These profiles then undergo an parallel mini-batch average agglomerative hierarchical clustering process, grouping together patients with similar healthcare trajectories to identify prevailing pathways and transitions within the system. By understanding these patterns, healthcare providers and policymakers can gain insights into the patient experience, potentially revealing areas for improvement, optimization, and personalization of care. Key findings include uncovering transitions in the healthcare environment, identifying the most common pathways, and studying the alternate level of care length of stay for each scenario. Looking ahead, the research anticipates incorporating additional layers of data, such as specific interventions and medications, to enrich the analysis. This expansion aims to offer a more comprehensive view of patient journeys, further enhancing the ability to tailor healthcare services to meet individual needs effectively. / Thesis / Master of Computer Science (MCS)

Health Informatics has enlightened by the recent development in the internet of medical things 4.0. Healthcare services have seen greater acceptance of Information and Communications Technology (ICT) in recent years; in light of the increasing volume of patient data, the traditional way of storing data in physical files has eventually moved to a digital alternative such as Electronic Health Record (EHR). However, the conventional healthcare data systems are plagued with a single point of failure, security issues, mutable logging, and inefficient methods to retrieve healthcare records. Solid (Social Linked Data) has been developed as a decentralized technology to alter digital data sharing and ownership for its users radically. However, Solid alone cannot address all the security issues posed to data exchange and storage. This work combines two decentralized technologies, Solid ecosystem and Blockchain technology, to tackle potential security issues using Solidity-based Smart Contracts, thereby providing a secure patient centric design. This research evaluates a model solution for secure storage, emphasizing secure auditing of accessing the data stored. The architecture will also come with algorithms that will provide developers with logical instructions to implement the artefact.

No description available.

Dans cette thèse, nous traitons les systèmes d’information hospitaliers (SIH), nous analysons leurs problématiques de conception, d’interopérabilité et de communication, dans l’objectif de contribuer à la conception d’un SIH canonique, coopératif, et communicant, ainsi de modéliser les échanges entre ses composants et également avec les autres systèmes impliqués dans la prise en charge du patient dans un réseau de soin. Nous proposons une structure et un modèle de conception d’un SIH canonique en se basant sur trois concepts principaux responsables de la production de l’information médicale, à savoir, le cas pathologique, le Poste de Production de l’Information Médicale (PPIM) et l’activité médicale elle même. Cette dernière, étant modélisée sur la notion d’arbre, permettra une meilleure structuration du processus de soin.Autant, dans l’optique d'assurer la continuité de soins, nous fournissons un modèle d’échange de données médicales à base du standard XML. Ce modèle consiste en un ensemble de données pertinentes organisées autours de cinq catégories : les données du patient, les données sur les antécédents du patient, les données de l’activité médicale, les données des prescriptions médicales et les données sur les documents médicaux (images, compte rendu…).Enfin, nous décrivons une solution d’intégration des systèmes d’information hospitaliers. La solution est inspirée de l’ingénierie des systèmes d’information coopératifs et consiste en une architecture de médiation structurée en trois niveaux : le niveau système d’information, le niveau médiation, et le niveau utilisateur. L’architecture propose une organisation modulaire des systèmes d'information hospitaliers et contribue à satisfaire l’intégration des données, des fonctions et du workflow de l’information médicale. / In this thesis, we deal with hospital information systems (HIS), we analyze their design issues, interoperability and communication, with the aim of contributing to the design of a canonical, cooperative, and communicative HIS, and model the exchanges between its components and also with other systems involved in the management of patient in a healthcare network.We propose a structure and a conceptual model of a canonical HIS based on three main concepts involved in the production of healthcare data, namely, the pathological case, the Production Post of Healthcare Data (PPHD) and medical activity itself. The latter, being modeled as a tree, will allow better structuring of the care process.However, in view of ensuring continuity of care, we provide an XML-based model for exchanging medical data. This model consists of a set of relevant data organized around five categories: patient data, data on patient history, data of medical activity, data of medical prescriptions and medical records data (images, reporting ...).Finally, we describe a solution for integrating hospital information systems. The solution is inspired by the engineering of cooperatives information systems and consists of mediation-based architecture, structured into three levels: the level of information systems, the level of mediation, and the user level. The architecture offers a modular organization of hospital information systems and helps to insure data, function and workflow integration.

Artificial intelligence (AI) and machine learning (ML) present some application opportunities and challenges that can be framed as learning problems. The performance of machine learning models depends on algorithms and the data. Moreover, learning algorithms create a model of reality through learning and testing with data processes, and their performance shows an agreement degree of their assumed model with reality. ML algorithms have been successfully used in numerous classification problems. With the developing popularity of using ML models for many purposes in different domains, the validation of such predictive models is currently required more formally. Traditionally, there are many studies related to model evaluation, robustness, reliability, and the quality of the data and the data-driven models. However, those studies do not consider the concept of the applicability domain (AD) yet. The issue is that the AD is not often well defined, or it is not defined at all in many fields. This work investigates the robustness of ML classification models from the applicability domain perspective. A standard definition of applicability domain regards the spaces in which the model provides results with specific reliability. The main aim of this study is to investigate the connection between the applicability domain approach and the classification model performance. We are examining the usefulness of assessing the AD for the classification model, i.e. reliability, reuse, robustness of classifiers. The work is implemented using three approaches, and these approaches are conducted in three various attempts: firstly, assessing the applicability domain for the classification model; secondly, investigating the robustness of the classification model based on the applicability domain approach; thirdly, selecting an optimal model using Pareto optimality. The experiments in this work are illustrated by considering different machine learning algorithms for binary and multi-class classifications for healthcare datasets from public benchmark data repositories. In the first approach, the decision trees algorithm (DT) is used for the classification of data in the classification stage. The feature selection method is applied to choose features for classification. The obtained classifiers are used in the third approach for selection of models using Pareto optimality. The second approach is implemented using three steps; namely, building classification model; generating synthetic data; and evaluating the obtained results. The results obtained from the study provide an understanding of how the proposed approach can help to define the model’s robustness and the applicability domain, for providing reliable outputs. These approaches open opportunities for classification data and model management. The proposed algorithms are implemented through a set of experiments on classification accuracy of instances, which fall in the domain of the model. For the first approach, by considering all the features, the highest accuracy obtained is 0.98, with thresholds average of 0.34 for Breast cancer dataset. After applying recursive feature elimination (RFE) method, the accuracy is 0.96% with 0.27 thresholds average. For the robustness of the classification model based on the applicability domain approach, the minimum accuracy is 0.62% for Indian Liver Patient data at r=0.10, and the maximum accuracy is 0.99% for Thyroid dataset at r=0.10. For the selection of an optimal model using Pareto optimality, the optimally selected classifier gives the accuracy of 0.94% with 0.35 thresholds average. This research investigates critical aspects of the applicability domain as related to the robustness of classification ML algorithms. However, the performance of machine learning techniques depends on the degree of reliable predictions of the model. In the literature, the robustness of the ML model can be defined as the ability of the model to provide the testing error close to the training error. Moreover, the properties can describe the stability of the model performance when being tested on the new datasets. Concluding, this thesis introduced the concept of applicability domain for classifiers and tested the use of this concept with some case studies on health-related public benchmark datasets. / Ministry of Higher Education in Libya

Bien qu'il s'agisse d'une maladie évitable par la vaccination, la grippe cause annuellement environ 3 à 5 millions de cas de maladie grave et environ 290 000 à 650 000 décès dans le monde. Pour prévenir l'infection et ses complications, la vaccination antigrippale est généralement recommandée pour toutes les personnes de 6 mois et plus. La vaccination annuelle est nécessaire en raison des perpétuels changements antigéniques des virus de la grippe; par conséquent, les souches incluses dans les vaccins antigrippaux sont régulièrement mises à jour. Ainsi, l'efficacité de terrain des vaccins antigrippaux (EV) varie d'une saison à l'autre, ce qui nécessite une surveillance constante pour évaluer l'impact des programmes de vaccination contre la grippe saisonnière au fil du temps. Les données médico-administratives sont une riche source d'informations qui pourraient être exploitées pour estimer l'efficacité réelle des vaccins antigrippaux. De plus, des études récentes ont rapporté que la réponse immunitaire à l'infection grippale et à la vaccination peut être altérée par des expositions antérieures, ce qui pourrait affecter l’efficacité de terrain des vaccins antigrippaux. Cette thèse visait à déterminer si les données médico-administratives fournissent des estimations valables de l'EV et à évaluer l'impact de la vaccination répétée et d’une infection antérieure par les virus de la grippe sur l'EV contre le syndrome d’allure grippale (SAG). Nous avons d'abord effectué une revue systématique de la littérature pour évaluer l'impact de la vaccination antigrippale sur la réduction des visites médicales pour un SAG, des hospitalisations pour un SAG, des hospitalisations pour la grippe confirmée en laboratoire (LCI) et des hospitalisations toutes causes confondues. Nous avons identifié que la spécificité des résultats joue un rôle crucial dans l'estimation de l'EV et que la propension à utiliser des soins de santé peut introduire un biais dans les études d'EV si la propension à consulter pour un SAG est influencée par le statut vaccinal, mais également si la capacité à capturer le statut vaccinal et l’issue (SAG) est tributaire de la propension à consulter. Par la suite, nous avons constaté que les courbes d'incidence des consultations médicales liées à un SAG spécifique, dérivées des codes de diagnostic clinique spécifiques à l'infection grippale, étaient très similaires aux données de surveillance des Centers for Disease Control and Prevention (CDC) des États-Unis pour le LCI, suggérant ainsi qu'il est plus approprié d'utiliser la définition de SAG spécifique à l’infection grippale pour une surveillance des cas de grippe plutôt qu'une définition large du SAG, lorsque seuls les codes de diagnostic clinique sont disponibles pour l'évaluation de l’infection grippale. Ensuite, l'efficacité de terrain des vaccins antigrippaux de 2018-2019 à prévenir les consultations médicales pour un SAG (spécifique à l’infection grippale) a été évaluée dans une cohorte d'individus américains ayant au moins un enregistrement pertinent par année, entre 2015 et 2019, dans leur dossier de santé électronique (DSE). Les rapports de cotes ajustés (aOR) ont été dérivés de modèles de régression logistique multivariés et les EVs ajustées ont été calculées à l'aide de 100x(1-aOR). Les estimations d’EVs dérivées des données médico-administratives étaient toutes plus petites que celles rapportées par les CDC américains, suggérant ainsi que l’utilisation secondaire de ces données médico-administratives a mené à une sous-estimation de l’EV probablement due à des biais de détection et de mauvaise classification corrélés avec la propension à utiliser des soins de santé. Lorsque les EVs sont stratifiées sur le nombre de visites médicales, les estimations d'EVs et la couverture vaccinale augmentent avec le nombre de visites médicales, atteignant des estimations similaires à celles obtenues par les CDC américains et la couverture vaccinale nationale des États-Unis pour les personnes ayant au moins 6 visites médicales lors des 12 mois précédents. Les résultats suggèrent ainsi que l'utilisation secondaire des données médico-administratives ne permet pas de produire des estimations valables de l’efficacité de terrain des vaccins antigrippaux, et ce, en l’absence de données complètes sur la vaccination et l'infection grippale. Cependant, ces données médico-administratives ont le potentiel d'évaluer l'efficacité de terrain des vaccins antigrippaux dans les populations considérées à haut risque de complications à la suite de l'infection, ce qui est difficile à faire avec une surveillance active, ainsi que dans les populations ayant des conditions de santé nécessitant un suivi médical soutenu; car la probabilité que le statut vaccinal et/ou l’infection grippale soient déclarés dans les données médico-administratives augmente avec le nombre de contacts avec le système de santé. Enfin, l'efficacité de terrain des vaccins antigrippaux de 2018-2019 à prévenir le SAG a été estimée en fonction de l’historique de la vaccination antigrippale et des antécédents de SAG chez les utilisateurs fréquents de soins de santé. Bien que l'EV semble diminuer avec l’augmentation du nombre de vaccinations précédentes, la vaccination antigrippale lors de la saison en cours offre probablement une protection contre le SAG, quels que soient les antécédents de vaccination, en particulier chez les enfants. Néanmoins, les antécédents de SAG pourraient atténuer l'effet négatif d'une vaccination antérieure sur l'EV, probablement en raison d’une immunité naturelle liée à l’infection grippale. Même si une vaccination antérieure peut atténuer l'EV de la saison en cours dans certaines circonstances, cet effet d'interférence est imprévisible et les antécédents de vaccination ou d'infection ne devraient pas influencer la décision de se faire vacciner contre la grippe. Jusqu'à ce que des vaccins antigrippaux universels efficaces soient disponibles et éliminent la nécessité d'une vaccination annuelle, la recommandation actuelle de la vaccination annuelle contre la grippe reste un bon moyen de se protéger et ainsi protéger les autres de l’infection et de ses complications, principalement pour les personnes à haut risque de complications. / Although a vaccine-preventable disease, influenza causes annually approximately 3 to 5 million cases of severe illness and about 290 000 to 650 000 deaths worldwide. To prevent the infection and its complications, influenza vaccination is recommended for all individuals 6 months and older. Annual vaccination is necessary because of continual antigenic changes of influenza viruses; hence, vaccine compositions are regularly updated. Consequently, vaccine effectiveness (VE) varies between seasons requiring ongoing measurement to assess the impact of seasonal influenza vaccination programs over time. Administrative healthcare databases are a rich source of information that could be leveraged to estimate real-world influenza VE. Recent studies have reported that immunologic response to influenza infection and vaccination may be altered by previous exposures. This thesis aimed to determine if administrative healthcare data provide accurate VE estimates and to evaluate the impact of repeated vaccination and previous infection on VE against medically attended influenza-like illness (MA-ILI). We first performed a systematic review of the literature to evaluate the impact of influenza vaccination to reduce outpatient visits for influenza-like illness (ILI), hospitalization for ILI, hospitalization for lab-confirmed influenza (LCI) and all-cause hospitalization. We identified that outcome specificity plays a crucial role in VE estimate and healthcare seeking behaviour can bias VE estimates if the propensity to seek care for ILI is influenced by vaccination status, but also if the ability to capture patients’ vaccination status and/or ILI is dependent on their propensity to seek care. Subsequently, we found that the incidence curves of influenza-related medical encounters, derived from clinical diagnostic codesspecific to influenza infection, were very similar to the United States (U.S.) Centers for Disease Control and Prevention (CDC) surveillance data for LCI; suggesting that it is more appropriate to use influenza case definition for specific surveillance rather than a broad ILI definition, when only clinical diagnostic codes are available for the evaluation of influenza. Then, the 2018-2019 influenza vaccine effectiveness against medically attended influenza-like illness (MA-ILI) was evaluated in a cohort of U.S. individuals who had at least one relevant record per year between 2015 and 2019 in their electronic medical record (EMR). Adjusted odds ratios (aORs) were derived from multivariate logistic regression models and adjusted VE (aVEs) were calculated using 100x(1-aORs). Estimated aVEs derived from administrative healthcare data were all lower than CDC-reported VE; results suggested that the secondary use of these administrative healthcare data led to an underestimation of influenza VE, likely due to detection and misclassification biases, correlated with healthcare seeking behaviour. When stratified by the number of primary care visits, aVE estimates and vaccine coverage increased with the number of primary care visits, reaching estimates similar to those obtained by the U.S. CDC and U.S. national vaccination coverage among those with at least 6 primary care visits in the previous 12 months. Results suggested that the secondary use of these administrative healthcare data cannot produce accurate influenza VE without comprehensive influenza vaccination and infection data. However, these databases have the potential to assess influenza VE in populations considered at high risk of complications following the infection, which is not easily achievable with active surveillance, as well as in populations with health conditions requiring constant medical follow-up since probabilities that vaccination and/or infection status are reported in administrative healthcare data increase with the number of contacts with the healthcare system. Finally, the 2018-2019 aVE against MA-ILI was estimated by previous vaccination status and previous history of MA-ILI in frequent healthcare users. Although VE appeared to decrease in relation to increasing numbers of previous influenza vaccinations, current season vaccination likely provides protection against MA-ILI regardless of vaccination history, especially in children. Nevertheless, previous MA-ILI could mitigate the negative effect of prior influenza vaccination on VE likely via infection-induced immunity. Even if prior influenza vaccination may attenuate current season VE in some circumstances, this interference effect is unpredictable and previous vaccination or infection history should not influence the decision to get vaccinated against influenza. Until effective universal influenza vaccines are available and eliminate the need for annual vaccination, the current recommendation for annual influenza vaccination remains important to protect ourselves and others from influenza infection and its complications, particularly in at-risk populations.

Screening programs for cervical cancer are performed to diagnose the early stage of cancer and prevent its development. Historically, the screening test was taken every three years, but it gradually varied due to continuous changes in technology, guidelines, and a new type of test. These changes made the screening participants uncertain and worried about understanding and interpreting the meaning of various test results and frequent follow-ups. These test results are manual and time-consuming to date. The participants seek it as digital, allowing easy and fast access to information. The primary research challenge of the proposed study is to present simple and complex test history data to the participants in an easily understandable way. The proposed research helped to design an artifact (prototype) that can be utilized to develop a digital service that does not cause any uncertain worry for the participants in understanding the complex test history data. For this, the proposed thesis study used state-of-the-art visualization techniques following the guidelines of design science research.

Yes / Introduction Heart failure affects 26 million people globally, approximately 900 thousand people in the UK, and is increasing in incidence. Appropriate management of medicines for heart failure at the time of hospital discharge reduces readmissions, improves quality of life and increases survival. The Improving the Safety and Continuity Of Medicines management at Transitions (ISCOMAT) trial tests the effectiveness of the Medicines at Transition Intervention (MaTI), which aims to enhance self-care and increase community pharmacy involvement in the medicines management of heart failure patients. Methods and analysis ISCOMAT is a parallel-group cluster randomised controlled trial, randomising 42 National Health Service trusts with cardiology wards in England on a 1:1 basis to implement the MaTI or treatment as usual. Around 2100 patients over the age of 18 admitted to hospital with heart failure with at least moderate left ventricular systolic dysfunction within the last 5 years, and planned discharge to the geographical area of the cluster will be recruited. The MaTI consists of training for staff, a toolkit for participants, transfer of discharge information to community pharmacies and a medicines reconciliation/review. Treatment as usual is determined by local policy and practices. The primary outcome is a composite of all-cause mortality and heart failure-related hospitalisation at 12 months postregistration obtained from national electronic health records. The key secondary outcome is continued prescription of guideline-indicated therapies at 12 months measured via patient-reported data and Hospital Episode Statistics. The trial contains a parallel mixed-methods process evaluation and an embedded health economics study. / The study was funded as part of a National Institute for Health Research Programme Grant for Applied Research (RP-PG-0514-20009).

The healthcare sector is pivotal, offering life-saving services and enhancing well-being and community life quality, especially with the transition from paper-based to digital electronic health records (EHR). While improving efficiency and patient safety, this digital shift has also made healthcare a prime target for cybercriminals. The sector's sensitive data, including personal identification information, treatment records, and SSNs, are valuable for illegal financial gains. The resultant data breaches, increased by interconnected systems, cyber threats, and insider vulnerabilities, present ongoing and complex challenges. In this dissertation, we tackle a multi-faceted examination of these challenges. We conducted a detailed analysis of healthcare data breaches using the VERIS (Vocabulary for Event Recording and Incident Sharing) dataset. We delve into the trends of these breaches, investigate the attack vectors, and identify patterns to inform effective mitigation strategies. We conducted a spatiotemporal analysis of the VERIS and the Office of Civil Rights (OCR) datasets. We explored the geographical and temporal distribution of breaches and focused on the types of targeted assets to decipher the attackers' motives. Additionally, we conducted a detailed analysis of hospitals' online presence, focusing on their security and performance features. By comparing government, non-profit, and private hospitals in the U.S., we examined their security practices, content, and domain attributes to highlight the differences and similarities in the digital profiles of these hospital types. Furthermore, we expand our scope to include a comparative sector-based study investigating data breaches across various critical sectors. This broader view provides a contextual understanding of the healthcare sector's unique vulnerabilities compared to other sectors. Overall, this dissertation contributes fundamental insights into healthcare data breaches and hospitals' digital presence and underscores the urgent need for enhanced understanding and implementation of robust security measures in this vitally important sector, striving for a balance between technological advancement and data security.

This Diploma thesis describes issues of big data in healthcare focus on picture archiving and communication system. DICOM format are store images with header where it could be other valuable information. This thesis mapping data from 1215 studies.