Thesis (Ph.D.)--University of California, San Francisco, 2004. / Bibliography: leaves 104-130. Also available online.

Thesis supervisor: W. N. Street. Includes bibliographic references (p. 111-118).

Thesis (Ph. D.)--State University of New York at Binghamton, Thomas J. Watson School of Engineering and Applied Science, Department of Systems Science and Industrial Engineering, 2009. / Includes bibliographical references.

Title from PDF of title page (University of Missouri--Columbia, viewed on February 24, 2010). The entire thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file; a non-technical public abstract appears in the public.pdf file. Dissertation advisor: Dr. James Laffey. Vita. Includes bibliographical references.

Thesis (M.S. in Biomedical Informatics)--Vanderbilt University, Dec. 2006. / Title from title screen. Includes bibliographical references.

The diffusion of information and communication technology (ICT) into healthcare has been generally low. This varies with application and setting, but at the point of care clinical level it has been particularly slow. The ICT niche in clinics has been recognised in numerous publications, where it potential benefits are proclaimed. A reoccurring factor identified with criticism of design i information systems research (ISR) is the difficulty in integrating the different human and technical elements. Activity Theory (AT) has been proposed as a means of overcoming this by providing single theoretical framework able to represent relevant factors across all levels of operational abstraction. In this work the (practical) operational functionality of AT is employed (tested) as a basis for design and evaluation of ICT, applied to integration at the clinical level of the National Health Service (NHS) healthcare organisation. Chronic wound healing is a complex activity, with a long history and strong dependence on data, as observed and recorded by clinicians, to treat and heal patients. Wound clinics that are part of the NHS, which is currently actively pursuing a strategy for information technology (IT) integration in healthcare, afford the opportunity to develop specific ICT for wound data and consider issues of diffusion at different levels of the organisation. An Action Research paradigm, using methods borrowed from soft systems methodology (SSM), is applied to the problem of producing ICT to manage wound data in participating NHS clinics. Data are collected via naturalistic (participant) observation, 'in-depth' interviews and focus groups, and are recorded using ethnographic field notes, a research logbook and diary, and digital and analogue voice recordings. Activity models are generated, to interpret the research process and represent the activity at the action level of the clinic, situating the analysis, both within the network of supporting activities, and the influence and constraints of the administrative and the organisational levels. Practical findings highlight the potential of ICT in participating clinics, showing how this can be expanded to the chronic wound healing activity in general, and reporting the implications that this has for the NHS IT strategy at the level of the clinics involved with regards to integration of ICT. Theoretical findings support the suitability of the Action Research strategy and the relevance of AT both as a descriptive framework for information systems development (!SD), and as an evaluative framework for ISR.

The wealth of medical-related information available today gives rise to a multidimensional source of knowledge. Research discoveries published in prestigious venues, electronic-health records data, discharge summaries, clinical notes, etc., all represent important medical information that can assist in the medical decision-making process. The challenge that comes with accessing and using such vast and diverse sources of data stands in the ability to distil and extract reliable and relevant information. Computer-based tools that use natural language processing and machine learning techniques have proven to help address such challenges. This current work proposes automatic reliable solutions for solving tasks that can help achieve a personalized-medicine, a medical practice that brings together general medical knowledge and case-specific medical information. Phenotypic medical observations, along with data coming from test results, are not enough when assessing and treating a medical case. Genetic, life-style, background and environmental data also need to be taken into account in the medical decision process. This thesis’s goal is to prove that natural language processing and machine learning techniques represent reliable solutions for solving important medical-related problems. From the numerous research problems that need to be answered when implementing personalized medicine, the scope of this thesis is restricted to four, as follows: 1. Automatic identification of obesity-related diseases by using only textual clinical data; 2. Automatic identification of relevant abstracts of published research to be used for building systematic reviews; 3. Automatic identification of gene functions based on textual data of published medical abstracts; 4. Automatic identification and classification of important medical relations between medical concepts in clinical and technical data. This thesis investigation on finding automatic solutions for achieving a personalized medicine through information identification and extraction focused on individual specific problems that can be later linked in a puzzle-building manner. A diverse representation technique that follows a divide-and-conquer methodological approach shows to be the most reliable solution for building automatic models that solve the above mentioned tasks. The methodologies that I propose are supported by in-depth research experiments and thorough discussions and conclusions.

M.Comm. / The present study was undertaken in a bid to meet an urgent need uncovered in medical-information systems (MIS) for a formal process whereby risks posing a threat to patients in medical institutions could be identified and controlled by means of the appropriate security measures. At the time of the study, however, no such formal risk-analysis model had yet been developed specifically for application in MIS. This gave rise to the development of RAMMO, a riskanalysis model specifically aimed at the identification of risks threatening the patient in his or her capacity as an asset in a medical institution. The author, therefore, managed to achieve her object with the study, namely to initiate a riskanalysis model that could be applied to medical environments. Following, an overview of the research method used in order to achieve the objectives of the study: Firstly, background information regarding the issues and problems to be addressed was obtained, and they provided the well-founded motivation for the study. Secondly, the development and importance of MIS in medical environments came under consideration, as well as the applicability of information security in an MIS. In the third instance, general terms and concepts used in the risk-management process were defined, by means of which definitions existing risk-analysis models were investigated and critically evaluated in a bid to identify a model that could be applied to a medical environment. Fourthly, a conceptual or draft design was suggested for a risk-analysis model developed specifically for medical environments. In doing so, the first two stages of the model, namely risk identification and risk assessment, were given special emphasis. The said model was then illustrated by means of a practical application in a general hospital in South Africa. The study culminated in a summation of the results of and the conclusions reached on the strength of the research. Further problem areas were also touched upon, which could become the focus of future research projects.

This study involves the investigation of the basis and validity of considering health information as therapeutic, the definition of Information Therapy, and whether the therapeutic nature of information can be measured empirically. The purpose of the study is to determine if there are any significant differences in the therapeutic effect of Information Therapy through the different delivery modes of support groups communicating face-to-face and those utilizing computer-mediated communication on the Internet. The comparison of these groups revealed no significant differences on three measures of health: physical, mental, and social support. Because one communication medium is not found to be advantageous over the other, the use of the computer can extend the benefits of Information Therapy to the home-bound, to those in remote areas, to people with time restraints, and those who may be shy. The validity of the therapeutic nature of information was verified by participant report of the effect of a health information search. Results demonstrated that the primary source for information is the physician, followed by the Internet, and 77% of participants reported a positive or therapeutic effect when health information was found. These results are significant because individuals who are in positions to deliver Information Therapy can better meet needs by identification of the sources to which people look for information and can have a major impact on patient care and the general health of the population. Providing people with information can empower them to take an active role in their health, can increase confidence in self-care, and should provide coping and disease management skills thus decreasing the utilization of healthcare resources and preventing costly acute and chronic health complications.

Learning representations of medical concepts from the Electronic Health Records (EHRs) has been shown effective for predictive analytics in healthcare. The learned representations are expected to preserve the semantic meanings of different medical concepts, which can be treated as features and thus benefit a variety of applications. Medical ontologies have also been explored to be integrated with the EHR data to further enhance the accuracy of various prediction tasks in healthcare. Most of the existing works assume that medical concepts under the same ontological category should share similar representations, which however does not always hold. In particular, the categorizations in the categorical medical ontologies were established with various factors being considered. Medical concepts even under the same ontological category may not follow similar occurrence patterns in the EHR data, leading to contradicting objectives for the representation learning. In addition, these existing works merely utilize the categorical ontologies. Actually, it has been noticed that ontologies containing multiple types of relations are also available. However, studies rarely make use of the diverse types of medical ontologies. In this thesis research, we propose three novel representation learning models for integrating the EHR data and medical ontologies for predictive analytics. To improve the interpretability and alleviate the conflicting objective issue between the EHR data and medical ontologies, we propose techniques to learn medical concepts embeddings with multiple ontological representations. To reduce the reliance on labeled data, we treat the co-occurrence statistics of clinical events as additional training signals, which help us learn good representations even with few labeled data. To leverage the various domain knowledge, we also consider multiple medical ontologies (CCS, ATC and SNOMED-CT) and propose corresponding attention mechanisms so as to take the best advantage of the medical ontologies with better interpretability. Our proposed models can achieve the final medical concept representations which align better with the EHR data. We conduct extensive experiments, and our empirical results prove the effectiveness of the proposed methods. Keywords: Bio/Medicine, Healthcare-AI, Electronic Health Record, Representation Learning, Machine Learning Applications