Pharmacovigilance is defined as the science and activities relating to the detection, assessment, understanding, and prevention of adverse drug events (WHO 2004). Post-approval adverse drug events are a major health concern. They attribute to about 700,000 emergency department visits, 120,000 hospitalizations, and $75 billion in medical costs annually (Yang et al. 2014). However, certain adverse drug events are preventable if detected early. Timely and accurate pharmacovigilance in the post-approval period is an urgent goal of the public health system. The availability of various sources of healthcare data for analysis in recent years opens new opportunities for the data-driven pharmacovigilance research. In an attempt to leverage the emerging healthcare big data, pharmacovigilance research is facing a few challenges. Most studies in pharmacovigilance focus on structured and coded data, and therefore miss important textual data from patient social media and clinical documents in EHR. Most prior studies develop drug safety surveillance systems using a single data source with only one data mining algorithm. The performance of such systems is hampered by the bias in data and the pitfalls of the data mining algorithms adopted. In my dissertation, I address two broad research questions: 1) How do we extract rich adverse drug event related information in textual data for active drug safety surveillance? 2) How do we design an integrated pharmacovigilance system to improve the decision-making process for drug safety regulatory intervention? To these ends, the dissertation comprises three essays. The first essay examines how to develop a high-performance information extraction framework for patient reports of adverse drug events in health social media. I found that medical entity extraction, drug-event relation extraction, and report source classification are necessary components for this task. In the second essay, I address the scalability issue of using social media for pharmacovigilance by proposing a distant supervision approach for information extraction. In the last essay, I develop a MetaAlert framework for pharmacovigilance with advanced text mining and data mining techniques to provide timely and accurate detection of adverse drug reactions. Models, frameworks, and design principles proposed in these essays advance not only pharmacovigilance research, but also more broadly contribute to health IT, business analytics, and design science research.

Magister Commercii - MCom (IM) (Information Management) / In an ideal public health arena, scientific evidence should be incorporated in the health information practices of making management decisions, developing policies, and implementing programs. However, much effort has been spent in developing health information practices focusing mainly on data collection, data quality and processing, with relatively little development on the utilization side of the information spectrum. Although the South Africa Health National Indicator Dataset of 2013 routinely collects and reports on more than two hundred elements, the degree to which this information is being used is not empirically known. The overall aim of the study was to explore the dynamics of routine primary healthcare information utilization in the West Coast district while identifying specific interventions that could ultimately lead to the improved use of data to better inform decision making. The ultimate goal being to enable managers to better utilize their routine health information for effective decision making.

Background: Advances in information technology have paved the way to facilitate accessibility to population-level health data through web-based data query systems (WDQSs). Despite these advances in technology, US state agencies face many challenges related to the dissemination of their local health data. It is essential for the public to have access to high-quality data that are easy to interpret, reliable, and trusted. These challenges have been at the forefront throughout the COVID-19 pandemic. Objective: The purpose of this study is to identify the most significant challenges faced by state agencies, from the perspective of the Behavioral Risk Factor Surveillance System (BRFSS) coordinator from each state, and to assess if the coordinators from states with a WDQS perceive these challenges differently. Methods: We surveyed BRFSS coordinators (N=43) across all 50 US states and the District of Columbia. We surveyed the participants about contextual factors and asked them to rate system aspects and challenges they faced with their health data system on a Likert scale. We used two-sample t tests to compare the means of the ratings by participants from states with and without a WDQS. Results: Overall, 41/43 states (95%) make health data available over the internet, while 65% (28/43) employ a WDQS. States with a WDQS reported greater challenges (P=.01) related to the cost of hardware and software (mean score 3.44/4, 95% CI 3.09-3.78) than states without a WDQS (mean score 2.63/4, 95% CI 2.25-3.00). The system aspect of standardization of vocabulary scored more favorably (P=.01) in states with a WDQS (mean score 3.32/5, 95% CI 2.94-3.69) than in states without a WDQS (mean score 2.85/5, 95% CI 2.47-3.22). Conclusions: Securing of adequate resources and commitment to standardization are vital in the dissemination of local-level health data. Factors such as receiving data in a timely manner, privacy, and political opposition are less significant barriers than anticipated.

The ongoinging demographic change will increase the demands on health care with an increase of old people suffering from long-term conditions in need of care. One way of meeting this increasing demand is to combine the use of modern technology and the involvement of patients by letting patients monitor their health themselves.  The use of patient-generated health data (PGHD) can benefit both patients and the health care and is an area that is being studied internationally. In this thesis work, two parts of the area of using PGHD in health care have been studied. First eleven professionals from the field of health care (both medical professionals, strategists and project leaders) were interviewed regarding their opinions on PGHD. They were generally positive to the phenomena and mentioned several types of measurements they found to be suited for PGHD. Among these measurement types were blood pressure, weight, blood glucose, electrocardiogram (ECG), peak expiratory flow (PEF), blood oxygen saturation and a variety of blood tests. The professionals found the greatest benets of PGHD to be the increased freedom and quality of life it offers to patients and the increased engagement to their own care it may lead to. The greatest concerns were related to technology problems and the patients using the measurement devices incorrectly. The second part of the work investigated how measurements of weight, blood pressure and ECG would be transferred from the device used by the patient to the electronic health record (EHR) in a standardized way. For the transfer being standardized, the study followed the Continua design guidelines (CDG), which are based on international standards and aim to achieve plug-and-play interoperability among health care devices and systems. This part of the study was carried out by studying the CDG documents as well as the standards to which they refer. The measurement data from the three device types were all described to be handled as numeric values, but in different formats. The weight is a single value, the blood pressure is a compound of values: systolic, diastolic and mean arterial pressure, and the ECG is an array of values. All measurement data is contained in a specic message together with additional data such as device type, device manufacturer, a reference ID and the date and time of the measurement. The data message is transferred from the measurement device to an application hosting device (AHD) with either a touch area network interface (using Near-Field Communications), personal area network interface (using USB or Bluetooth communication) or local area network interface (using ZigBee communication). From the AHD the data transfer chain continues until the data reaches the EHR.

Healthcare utilizations are typically measured independently of each other; neglecting the interdependencies between services. An episode of care is suitable for measuring healthcare utilizations of patients with complex health conditions because it tracks all contacts throughout the healthcare system. The overall goal of this research was to construct an episode of care data system to study healthcare utilizations and costs of chronic obstructive pulmonary disease (COPD) exacerbations. To achieve this goal, four related studies were undertaken. The first study (Chapter 2) evaluated the agreement between emergency department (ED) data and hospital records for capturing transitions between the two care settings. Using the κ statistic as a measure of concordance, we found good agreement between the two data sources for intra-facility transfers; but only fair agreement for inter-facility transfers. The results show that linking multiple data sources would be important to identify all related healthcare utilization across care settings. The second study (Chapter 3) linked hospital data, ED data, physician billing claims, and outpatient drug records to construct an episode of care data system for COPD patients. Latent class analysis was used to identify COPD patient groups with distinct healthcare pathways. Pathways were associated with outcomes such as mortality and costs. A few individuals followed complex pathways and incurred high costs. Building on the previous study, the next one (Chapter 4) predicted whether high-cost patients in one episode also incurred high costs in subsequent episodes. Using logistic regression models, we found that patient information routinely collected in administrative health data could satisfactorily predict those who become persistent high users. The final study (Chapter 5) used a cross-validation approach to compare the performance of eight alternative linear regression models for predicting costs of episodes of COPD exacerbations. The results indicate that the robust regression model, a model not often considered for cost prediction, was among the best models for predicting episode-based costs. Overall, this research demonstrated how population-based administrative health databases could be linked to construct an episode of care data system for a chronic health condition. The resulting data system supported novel investigations of healthcare system-wide utilizations and costs. / May 2016

The study was user survey method based, performed to get the overview of use and perceptions of health caremanagers towards Lithuanian computerized health information system as a tool for decision – making. Aims of the study were to describe LCHIS, its inputs and potential use; to account for a surveyofpotential users, health care administrators; to discuss the need for improvement of the system and itsuse. Methods. User survey method was applied. Literature search was performed and the questionnaire was constructed after interview with four respondents and clarification of questions. Totally 100 ofrespondents from different health care institutions were interviewed. Data was analysed using normal statistical methods, using MS Excel 2000 and statistical package SPSS 10.0 as tools. Main results. Concerning the awareness about the existing of LCHIS, 68% of the respondents saidtheyhave heard about it and 15% said theyhave been using this system daily. As many as 68% of respondents didn’t really take care about the existence of LCHIS, while the size of respondents being satisfied and not was pretty the same. The number of satisfied with the structure was rather small ifcomparing with those partially satisfied. As many as 76% of the respondents said they haven’t been using the system at all. 24% of the respondents were satisfied with the certain groups of healthindicators within the system. Group of morbidity indicators and group of hospital activity indicatorswere among the mostly used (17% together). Almost 20% of the respondents said it was easy for them to use LCHIS; the same number of health care administrators trusted the information comingfrom LCHISand they have experienced the situation, where they have used LCHIS for planning ormanagement in current situation. As many as 82% of health care managers agreed heads or administrative staff of hospitals supposed to be the key members, who must encourage them to use the system. Conclusions. About two thirds of health care administrators interviewed knew about LCHIS and the rest had been or were users. In the comments this group claimed they were supporting their decisions by using the systemand indicators in it. As many as 96% of the respondents stated there was a needfor statistical information and skills for dailydecision - making and managerial activities. The respondents, who used LCHIS, trusted the information in the system and found it useful in their dailywork as health managers. The main comments, why respondents didn’t use the system or didn’tknow about it, was lack of information technologies in work place, lack of computer skills and lackof support from hospital authorities / <p>ISBN 91-7997-097-4</p>

Le vieillissement de la population s'accompagne de l'augmentation du nombre de patients souffrant de maladies chroniques. Ceci entraîne une augmentation du nombre de visites et d'examens à l'hôpital. La télémédecine peut apporter un bénéfice tant en matière de qualité et de sécurité des soins qu'en matière de réduction des dépenses de santé. Elle apparaît comme une piste prometteuse, mais encore insuffisamment déployée. Nous avons déterminé les pathologies chroniques les plus fréquentes compatibles avec la télémédecine, à savoir l'insuffisance cardiaque, le diabète, l'insuffisance respiratoire et l'insuffisance rénale. Cette étude a mis en exergue la présence d'un ensemble de comorbidités associées à ces quatre pathologies et montré la nécessité d'une prise en charge globale du patient. Un état de l'art des différentes expériences de télémédecine dans le monde pour ces maladies a mis en évidence que les différentes applications proposées sont partiellement redondantes et ne sont pas interopérables entre elles. Ainsi, les patients peuvent réaliser deux fois la même mesure pour le même examen médical, mais pour deux pathologies différentes. Ces deux problèmes peuvent induire des développements redondants pour chaque application de télémédecine, des risques de diminution de l'efficience des applications de télémédecine lors de leur déploiement, ainsi que des risques d'aggravation de la santé du patient dès lors que l'action d'un professionnel de santé sur une pathologie peut avoir des répercussions sur une autre pathologie. Par ailleurs, cette étude a fait apparaître les besoins communs à ces pathologies. Nos travaux ont donc consisté à développer une architecture générique permettant à différentes applications de télémédecine spécifiques à une pathologie chronique de partager un plateau technique commun. L'originalité de ce travail porte d'une part sur l'étude des normes et des standards de communication nécessaires à l'interopérabilité de l'infrastructure envisagée, et d'autre part sur une modélisation des données relatives aux signes vitaux analysés et à leur contexte. En effet, ces dernières contiennent toutes les informations pouvant influer sur l'interprétation des résultats, telles que la date et l'horaire de la mesure réalisée, la nature de la donnée acquise et les caractéristiques des capteurs utilisés. Pour valider notre modèle d'application de télésurveillance des maladies chroniques, nous avons réalisé deux expérimentations. La première, menée en collaboration avec la société AZNetwork, a consisté à mettre en œuvre une plate-forme digitale de recueil et d'archivage des données médicales pour les seniors dans le cadre du projet Silver@Home. La seconde expérimentation réalisée en partenariat avec le réseau de soins TELAP sur le projet Domoplaies a permis d'étendre notre modèle à un système d'échanges d'information médicale entre les professionnels de santé. Ces travaux constituent une proposition de modèle d'application de télémédecine qui est non seulement conforme au Cadre d'Interopérabilité des Systèmes d'Information de Santé (CI-SIS) de l'Agence des Systèmes d'Information Partagés de Santé (ASIP Santé), mais qui constitue une proposition d'extension de ce dernier à l'acquisition des données au domicile du patient. / The current trend in aging population leads to an increasing number of chronic diseases cases and consequently, to an increase of the number of medical examinations and hospital stays. Telemedicine system can contribute to both increase or maintain care quality and safety, as well as to reduce costs. In spite of this potential, telemedicine deployment is currently limited. We identified the most frequent chronic diseases consistent with telemedicine, namely heart failure, diabetes, respiratory failure and kidney failure. This study highlighted a number of comorbidities associated to these four diseases, reflecting the need for overall patient care. A state of the art report on worldwide Telemedicine experiments for the four chronic diseases showed that the current applications are partially redundant and hardly interoperable. Thus, the same measure can be performed twice for the same medical examination, but for two different diseases. These two problems can induce redundant developments, a risk of a decreased efficiency of a telemedicine application during its deployment, as well as risks of making the patient health worse when the intervention of a healthcare professional can have an impact on another chronic disease. Furthermore, this study revealed common requirements for these chronic diseases and their specific features. We developed a generic architecture that allows different telemedicine applications associated with specific diseases to share a common technical platform. The original aspects of this work are first, a study of communication standards to achieve an interoperable system, and, on the other hand second, a health data model for the patient's vital signs. This model contains all the necessary information to interpret the results, such as the date and time of the measurement, the acquired data format and the sensor characteristics. To validate our telemedicine application model, we conducted two experiments. The first one was a collaboration with AZNetwork company. It consisted in the development of a digital platform to collect and archive seniors' data in the context of the Silver@Home project. The second experiment was a partnership with the TELAP network on the Domoplaies project. It allowed us to extend our telemedicine model to a medical data exchange system among healthcare providers. This led us to propose a telemedicine application model, which is not only in conformity with the Health Information Systems Interoperability Framework (HIS-IF) of the "Agence des Systèmes d'Information Partagés de Santé" (ASIP Santé), but also constitutes a proposed extension of this framework to the patient's home.

The Kimball methodology, often referred to as dimensional modelling, is well established in data warehousing and business intelligence as a highly successful means for turning data into information. Yet weaknesses exist in the Kimball approach that make it difficult to rapidly extend or interrelate dimensional models in complex business areas such as Health Care. This Thesis looks at the development of a methodology that will provide for the rapid extension and interrelation of Kimball dimensional models. This is achieved through the use of techniques similar to those employed in the semantic web. These techniques allow for rapid analysis and insight into highly variable data which previously was difficult to achieve. / Graduate

Data collection is one of the important components of public health systems. Decision makers, policy makers and health service providers need accurate and timely data in order to improve the quality of their services. The rapidly growing use of mobile technologies has increased pressure on the demand for mobile-based data collection solutions to bridge the information gaps in the health sector of the developing world. This study reviews existing health data collection systems and the available open source tools that can be used to improve these systems. We further propose a prototype using open source data collection frameworks to test their feasibility in improving the health data collection in the developing world context. We focused on the statistical health data, which are reported to secondary health facilities from primary health facilities. The proposed prototype offers ways of collecting health data through mobile phones and visualizes the collected data in a web application. Finally, we conducted a qualitative study to assess challenges in remote health data collection and evaluate usability and functionality of the proposed prototype. The evaluation of the prototype seems to show the feasibility of mobile technologies, particularly open source technologies, in improving the health data collection and reporting systems for the developing world.

Over the past 10 years, smartwatches have become increasingly popular for commercial use. Their ever-increasing capabilities, accuracy, and sophistication of smartwatches is making them increasingly appealing to physical activity researchers as a valuable research tool. The non-invasive nature, prevalence, and versatility of smartwatches is being utilized to track heart rate, blood-oxygen levels, activity and movement, and sleep. However, the current state of the art lacks a uniform method to extract, organize, and analyze data collected from these devices. The objective of this research was to develop a Python software library that is widely available, highly capable, and easy to use with the data collected by the Apple Watch. The library was designed to offer data science, visualization, and mining features that help physical activity research find and communicate patterns in the Apple Health data. The custom-built caching system of the library provides near-instant runtime to parse and analyze large files without trading off on memory usage. The Wanjara Smartwatch Library has significantly better performance, proven reliability and robustness, and improved usability than the alternatives discovered in the review of the literature. / Master of Science / Over the past 10 years, smartwatches have become increasingly popular for commercial use. Their ever-increasing capabilities, accuracy, and sophistication of smartwatches is making them increasingly appealing to physical activity researchers as a valuable research tool. The non-invasive nature, prevalence, and versatility of smartwatches is being utilized to track heart rate, blood-oxygen levels, activity and movement, and sleep. However, the current state of the art lacks a uniform method to extract, organize, and analyze data collected from these devices. The objective of this research was to develop a Python software library that is widely available, highly capable, and easy to use with the data collected by the Apple Watch. The library was designed to offer data science, visualization, and mining features that help physical activity research find and communicate patterns in the Apple Health data. The custom-built caching system of the library provides near-instant runtime to parse and analyze large files without trading off on memory usage. The Wanjara Smartwatch Library has significantly better performance, proven reliability and robustness, and improved usability than the alternatives discovered in the review of the literature.