Advanced information technologies promise a massive influx of individual-specific medical data. These rich sources offer great potential for an increased understanding of disease mechanisms and for providing evidence-based and personalized clinical decision support. However, the size, complexity, and biases of the data pose new challenges, which make it difficult to transform the data to useful and actionable knowledge using conventional statistical analysis. The so-called “Big Data” era has created an emerging and urgent need for scalable, computer-based data mining methods that can turn data into useful, personalized decision support knowledge in a flexible, cost-effective, and productive way. The goal of my Ph.D. research is to address some key challenges in current clinical deci-sion support, including (1) the lack of a flexible, evidence-based, and personalized data mining tool, (2) the need for interactive interfaces and visualization to deliver the decision support knowledge in an accurate and effective way, (3) the ability to generate temporal rules based on patient-centric chronological events, and (4) the need for quantitative and progressive clinical predictions to investigate the causality of targeted clinical outcomes. The problem statement of this dissertation is that the size, complexity, and biases of the current clinical data make it very difficult for current informatics technologies to extract individual-specific knowledge for clinical decision support. This dissertation addresses these challenges with four overall specific aims: Evidence-Based and Personalized Decision Support: To develop clinical decision support systems that can generate evidence-based rules based on personalized clinical conditions. The systems should also show flexibility by using data from different clinical settings. Interactive Knowledge Delivery: To develop an interactive graphical user interface that expedites the delivery of discovered decision support knowledge and to propose a new visualiza-tion technique to improve the accuracy and efficiency of knowledge search. Temporal Knowledge Discovery: To improve conventional rule mining techniques for the discovery of relationships among temporal clinical events and to use case-based reasoning to evaluate the quality of discovered rules. Clinical Casual Analysis: To expand temporal rules with casual and time-after-cause analyses to provide progressive clinical prognostications without prediction time constraints. The research of this dissertation was conducted with frequent collaboration with Children’s Healthcare of Atlanta, Emory Hospital, and Georgia Institute of Technology. It resulted in the development and adoption of concrete application deliverables in different medical settings, including: the neuroARM system in pediatric neuropsychology, the PHARM system in predictive health, and the icuARM, icuARM-II, and icuARM-KM systems in intensive care. The case studies for the evaluation of these systems and the discovered knowledge demonstrate the scope of this research and its potential for future evidence-based and personalized clinical decision support.

Background: Tobacco smoke exposure (TSE) is unequivocally harmful to children's health, yet up to 48% of children who visit the pediatric emergency department (PED) and urgent care setting are exposed to tobacco smoke. The incorporation of clinical decision support systems (CDSS) into the electronic health records (EHR) of PED patients may improve the rates of screening and brief TSE intervention of caregivers and result in decreased TSE in children. Objective: We propose a study that will be the first to develop and evaluate the integration of a CDSS for Registered Nurses (RNs) into the EHR of pediatric patients to facilitate the identification of caregivers who smoke and the delivery of TSE interventions to caregivers in the urgent care setting. Methods: We will conduct a two-phase project to develop, refine, and integrate an evidence-based CDSS into the pediatric urgent care setting. RNs will provide input on program content, function, and design. In Phase I, we will develop a CDSS with prompts to: (1) ASK about child TSE and caregiver smoking, (2) use a software program, Research Electronic Data Capture (REDCap), to ADVISE caregivers to reduce their child's TSE via total smoking home and car bans and quitting smoking, and (3) ASSESS their interest in quitting and ASSIST caregivers to quit by directly connecting them to their choice of free cessation resources (eg, Quitline, SmokefreeTXT, or SmokefreeGOV) during the urgent care visit. We will create reports to provide feedback to RNs on their TSE counseling behaviors. In Phase II, we will conduct a 3-month feasibility trial to test the results of implementing our CDSS on changes in RNs' TSE-related behaviors, and child and caregiver outcomes. Results: This trial is currently underway with funding support from the National Institutes of Health/National Cancer Institute. We have completed Phase I. The CDSS has been developed with input from our advisory panel and RNs, and pilot tested. We are nearing completion of Phase II, in which we are conducting the feasibility trial, analyzing data, and disseminating results. Conclusions: This project will develop, iteratively refine, integrate, and pilot test the use of an innovative CDSS to prompt RNs to provide TSE reduction and smoking cessation counseling to caregivers who smoke. If successful, this approach will create a sustainable and disseminable model for prompting pediatric practitioners to apply tobacco-related guideline recommendations. This systems-based approach has the potential to reach at least 12 million smokers a year and significantly reduce TSE-related pediatric illnesses and related costs.

This thesis presents the framework of a complex business process driven event stream processing system to produce meaningful output with direct implications to the business objectives of an organization. This framework is demonstrated using a case study instantiating the management of a newborn infant with hypoglycaemia. Business processes defined within guidelines, are defined at build-time while critical knowledge found in the definition of business processes are used to support their enactment for stream analysis. Four major research contributions are delivered. The first contribution enables the definition and enactment of complex business processes in real-time. The second contribution supports the extraction of business process using knowledge found within the initial expression of the business process. The third contribution allows for the explicit use of temporal abstraction and stream analysis knowledge to support enactment in real-time. Finally, the last contribution is the real-time integration of heterogeneous streams based on Service-Oriented Architecture principles. / UOIT

Osteoporosis is a major public health concern, affecting over 200 million people worldwide. There is valid evidence outlining how osteoporosis can be diagnosed and managed, but gaps exist between evidence and practice. Graham’s “Knowledge to Action” (KTA) process for knowledge translation and the Medical Research Council (MRC) framework for complex interventions were used to address these gaps. The first 4 KTA steps were collapsed into 3 phases of the PhD research plan. In PhD Phase 1, a systematic review was conducted to identify tools that facilitate decision making in osteoporosis disease management (DM). Results showed that few DM tools exist, but promising strategies were those that incorporated reminders and education and targeted physicians and patients. PhD Phase 2 used the findings from the systematic review and consultation with clinical and human factors engineering experts to develop a conceptual design of the tool. Multiple components targeted to both physicians and patients at the point of care, and which could be used as a standalone system or modifiable for integration with electronic health record systems were outlined. PhD Phases 3a and 3b were devoted to the assessment of the barriers to knowledge. In Phase 3a, a qualitative study of focus groups was conducted with physicians to identify attitudes and perceived barriers to implementing decision support tools in practice, and to identify the features that should be included in the design. Findings from 4 focus groups combined with aging research, and input from design and information experts were used to transform the conceptual design into a functional prototype. In Phase 3b, each component of the prototype was tested in 3 usability evaluation studies using an iterative, participant-centered approach to assess how well the prototype met end users’ needs. Findings from the usability study informed the final prototype, which is ready for implementation as part of the post PhD plan to fulfill the requirements of the remaining steps of the KTA and MRC frameworks.

Osteoporosis is a major public health concern, affecting over 200 million people worldwide. There is valid evidence outlining how osteoporosis can be diagnosed and managed, but gaps exist between evidence and practice. Graham’s “Knowledge to Action” (KTA) process for knowledge translation and the Medical Research Council (MRC) framework for complex interventions were used to address these gaps. The first 4 KTA steps were collapsed into 3 phases of the PhD research plan. In PhD Phase 1, a systematic review was conducted to identify tools that facilitate decision making in osteoporosis disease management (DM). Results showed that few DM tools exist, but promising strategies were those that incorporated reminders and education and targeted physicians and patients. PhD Phase 2 used the findings from the systematic review and consultation with clinical and human factors engineering experts to develop a conceptual design of the tool. Multiple components targeted to both physicians and patients at the point of care, and which could be used as a standalone system or modifiable for integration with electronic health record systems were outlined. PhD Phases 3a and 3b were devoted to the assessment of the barriers to knowledge. In Phase 3a, a qualitative study of focus groups was conducted with physicians to identify attitudes and perceived barriers to implementing decision support tools in practice, and to identify the features that should be included in the design. Findings from 4 focus groups combined with aging research, and input from design and information experts were used to transform the conceptual design into a functional prototype. In Phase 3b, each component of the prototype was tested in 3 usability evaluation studies using an iterative, participant-centered approach to assess how well the prototype met end users’ needs. Findings from the usability study informed the final prototype, which is ready for implementation as part of the post PhD plan to fulfill the requirements of the remaining steps of the KTA and MRC frameworks.

Clinical decision support systems (CDSS) and the Picture Archiving and Communication System (PACS) have been adopted by large healthcare organization to support stroke diagnosis to reduce the level of misdiagnosis occurrence. This research presents a methodology for constructing a stroke decision support system (Stroke DSS) which integrates basic information, physical and image stroke assessment criterions, constructs ischemic, hemorrhage and subarachnoid hemorrhage of stroke diagnosis flow. A prototype embedded methodology was built to support stroke diagnosis in healthcare organization. Using a design science approach, we embed the constructs of our methodology in a prototype and perform a usability evaluation to demonstrate the utility of our approach. The usability evaluation demonstrates the effectiveness of our approach in terms of efficiency, effectiveness and satisfaction. The resulting system allowed flexible knowledge model and representation that are useful for stroke diagnosis.

Gait analysis is the process of collecting and analyzing quantitative information about walking patterns of the people. Gait analysis enables the clinicians to differentiate gait deviations objectively. Diagnostic decision making from gait data only requires high level of medical expertise of neuromusculoskeletal system trained for the purpose. An automated system is expected to decrease this requirement by a &lsquo / transformed knowledge&rsquo / of these experts. This study presents a clinical decision support system for the detecting and scoring of a knee disorder, namely, Osteoarthritis (OA). Data used for training and recognition is mainly obtained through Computerized Gait Analysis software. Sociodemographic and disease characteristics such as age, body mass index and pain level are also included in decision making. Subjects are allocated into four OA-severity categories, formed in accordance with the Kellgren-Lawrence scale: &ldquo / Normal&rdquo / , &ldquo / Mild&rdquo / , &ldquo / Moderate&rdquo / , and &ldquo / Severe&rdquo / . Different types of classifiers are combined to incorporate the different types of data and to make the best advantages of different classifiers for better accuracy. A decision tree is developed with Multilayer Perceptrons (MLP) at the leaves. This gives an opportunity to use neural networks to extract hidden (i.e., implicit) knowledge in gait measurements and use it back into the explicit form of the decision trees for reasoning. Individual feature selection is applied using the Mahalanobis Distance measure and most discriminatory features are used for each expert MLP. Significant knowledge about clinical recognition of the OA is derived by feature selection process. The final system is tested with test set and a success rate of about 80% is achieved on the average.

Clinical guidelines are developed to assist healthcare practitioners to make decisions on a patient&#039 / s medical problems and as such they communicate with external applications to retrieve patient data, to initiate medical actions through clinical workflows and to transmit information to alert/reminder systems. The interoperability problems in the healthcare IT domain for interacting with heterogeneous clinical workflow systems and Electronic Healthcare Record (EHR) Systems prevent wider deployment of clinical guidelines because each deployment requires a tedious custom adaptation phase. In this thesis, we provide machine processable mechanisms that express the semantics of clinical guideline interfaces so that automated processes can be used to access the clinical resources for guideline deployment and execution. For this purpose, we propose a semantically enriched clinical guideline representation formalism by extending one of the computer interpretable guideline representation languages, GuideLine Interchange Format (GLIF). To be able to deploy the semantically extended guidelines to healthcare settings semi-automatically, the underlying application&#039 / s semantics must also be available. We describe how this can be achieved based on two prominent implementation technologies in use in the eHealth domain: Integrating Healthcare Enterprise (IHE) Cross Enterprise Document Sharing Integration Profile (XDS) for discovering and exchanging EHRs and Web service technology for interacting with the clinical workflows and wireless medical sensor devices. Since the deployment and execution architecture should be dynamic, and address the heterogeneity of underlying clinical environment, the deployment and execution is coordinated by a multi-agent system. The system described in this thesis is realized within the scope of the SAPHIRE Project.

Despite considerable effort considerable cost in both time and money, as many as two out of three persons with type 1 diabetes are not in control of their disease. As a result, 40% of these individuals will go on to develop at least one serious complication including retinopathy, nephropathy, neuropathy and cardiomyopathy. It is further estimated that as much as $4 billion could be saved annually if all persons with type 1 diabetes in the US were properly controlled. Adequate treatment of type 1 diabetes is predicated on the estimation of three clinical insulin therapy parameters: the basal dose, the insulin sensitivity factor and the insulin-to-carbohydrate ratio. Currently, these therapy parameters are determined by iterative titration procedures based on expert opinion. Unfortunately, there is evidence suggesting that for the majority of individuals, these titration protocols do not provide good results. In this work we develop an alternative to traditional insulin titration protocols that allows clinical insulin therapy parameters to be estimated directly from a set of easily acquired measurements. First, a simple model of type 1 diabetes is used to derive a series of equations connecting the model's parameters to the clinically important insulin therapy parameters of insulin sensitivity factor, insulin-to-carbohydrate ratio and basal insulin dose. The simplifying assumptions used to derive these equations are tested and shown to be valid and the Fisher Information Matrix is used to demonstrate parameter identifiability. Parameter estimation is then performed on two cohorts of virtual subjects, as well as two segments of real continuous glucose monitoring data from a person with type 1 diabetes. Identification of the true insulin therapy parameters is successful under most conditions for both cohorts of virtual subjects. Parameter estimation for one of the two segments of real continuous glucose monitoring data is also successful. Finally, because continuous glucose monitors are instrumental to successful implementation of our insulin therapy framework, the physiological environment in which continuous glucose monitoring takes place is modeled and a fundamental limitation on measurement precision is shown to exist. An examination of physiological variability in the parameters indicates that many of the challenges observed in real world continuous glucose monitoring may have a relationship to changes in capillary bed perfusion. A rationale for anecdotally reported sensor faults is also proposed based on the physical mechanisms explored. / text

<p>Diss. (sammanfattning) Umeå : Umeå universitet, 2003</p> / digitalisering@umu