Complex patient are becoming more and more of a challenge to the health care system given the amount of care they require and the amount of documentation needed to keep track of their state of health and treatment. Record keeping using the EHR makes this easier but mounting amounts of patient data also means that clinicians are faced with information overload. Information overload has been shown to have deleterious effects on care, with increased safety concerns due to missed information. Patient record summarization has been a promising mitigator for information overload. Subsequently, a lot of research has been dedicated to record summarization since the introduction of EHRs. In this dissertation we examine whether unsupervised inference methods can derive patient problem-oriented summaries, that are robust to different patients. By grounding our experiments with HIV patients we leverage the data of a group of patients that are similar in that they share one common disease (HIV) but also exhibit complex histories of diverse comorbidities. Using a user-centered, iterative design process, we design an interactive, longitudinal patient record summarization tool, that leverages automated inferences about the patient's problems. We find that unsupervised, joint learning of problems using correlated topic models, adapted to handle the multiple data types (structured and unstructured) of the EHR, is successful in identifying the salient problems of complex patients. Utilizing interactive visualization that exposes inference results to users enables them to make sense of a patient's problems over time and to answer questions about a patient more accurately and faster than using the EHR alone.

Electronic health records (EHRs) led to improvements in patient safety, care delivery, and efficiency; however, they have also resulted in significant increases in documentation time. EHR documentation burden, defined as “added work (e.g., documentation) or extraneous actions (e.g., clicks) performed in the EHR beyond that which is required for good clinical care”, has been linked to increased medical errors, poorer patient outcomes, reduced care quality, cognitive overload, and ultimately, burnout among clinicians. Relative to other clinical practice settings where patient flows are more predictable and of lower intensity, emergency department (ED) clinicians report markedly higher workload. Furthermore, EHR implementation research in the ED indicates that incongruities between EHR design and usability and the clinical workflow may intensify clinician workflow fragmentation. In our prior work, we identified workflow fragmentation, which we define as task switching, as one potential approach for evaluating documentation burden in ED practice settings. Yet, no standardized, scalable measures of documentation burden have been developed. Despite shortcomings, there have been increasing efforts to leverage information from EHR event logs as an alternative to direct clinical observation methods in evaluating user-centric behaviors and interactions with health information technology systems. Using EHR event logs, this dissertation aims to advance the study of evaluating burden by investigating EHR-mediated workflow fragmentation as a measure of EHR documentation burden among physicians and registered nurses (hereinafter interchangeably referred to as “clinicians”) in the ED. First, I review the literature on the existing quantitative approaches employed for measuring clinician documentation burden in clinical practice settings. Next, I explore EHR factors perceived to contribute to clinician documentation burden as well as the perceived role of workflow fragmentation on clinician documentation burden in the ED. Lastly, I investigate data-driven approaches to abstract clinically relevant concepts from EHR event logs for studying EHR documentation burden—culminating into a computational framework to evaluate ED clinician documentation burden in the context of cognitive burden. Collectively, the work conducted in this dissertation contributes computational methods that are foundational for investigating clinician documentation burden measurement at scale using EHR event logs, informed by current evidence and clinician perspectives, and grounded in theory.

Innovative primary care models that deliver comprehensive primary care to address medical and social needs are an established means of improving health outcomes and reducing healthcare costs among persons living with chronic disease. Care management is one such approach that requires providers to monitor their respective patient panels and intervene on patients requiring care. Health information technology (IT) has been established as a critical component of care management and similar care models. While there exist a plethora of health IT systems for facilitating primary care, there is limited research on their ability to support care management and its emphasis on monitoring panels of patients with complex needs. In this dissertation, I advance the understanding of how computational methods can better support clinicians delivering care management, and use the management of human immunodeficiency virus (HIV) as an example scenario of use. The research described herein is segmented into 3 aims; the first was to understand the processes and barriers associated with care management and assess whether existing IT can support clinicians in this domain. The second and third aim focused on informing potential solutions to the technological shortcomings identified in the first aim. In the studies of the first aim, I conducted interviews and observations in two HIV primary care programs and analyzed the data generated to create a conceptual framework of population monitoring and identify challenges faced by clinicians in delivering care management. In the studies of the second aim, I used computational methods to advance the science of extracting from the patient record social and behavioral determinants of health (SBDH), which are not easily accessible to clinicians and represent an important barrier to care management. In the third aim, I conducted a controlled experimental evaluation to assess whether data visualization can improve clinician’s ability to maintain awareness of their patient panels.

The expectation of graduating nurses today is to be knowledgeable and responsive to rapidly changing technology in the health care environment. Although federal mandates, Institute of Medicine (IOM) recommendations, and nursing program accreditation initiatives are pushing an "informatics" healthcare agenda by promoting the implementation of electronic health record (EHR) systems by 2014 in all healthcare facilities, very few US nursing schools provide students with access and training in, EHR systems. In addition, nursing faculty may not have a clear understanding of healthcare informatics; the use of information and technology to communicate, manage knowledge, mitigate error, and support decision-making. Nursing education must address faculty issues related to this innovative paradigm in order to keep pace and participate as co-creators of relevant informatics technology curriculum that prepares graduates for real life workforce. Understanding the challenges, concerns, and successes in implementing informatics may help nurse educators as they develop curriculum and teach in this environment. This case study explores and describes, with nursing faculty of a small liberal arts college, faculty knowledge, skills, and attitudes (KSAs) as they participate in an action research framed curriculum development program for informatics academic EHR (AEHR) integration. The research question:What is the experience of nursing educators and nursing faculty members involved in the integration of an AEHR project framed in the Learning by Developing model at a small liberal arts college school of nursing? Significant insights as participants in the study influenced nurse educators' ideas regarding collaborative curricular design, meaningful assignments, and the importance of feedback.

The overarching goal of this dissertation is to integrate heterogeneous data for the estimation of disease coheritability and subtyping. Chapter 2 focuses on the significance and estimation of heritability and coheritability, which quantify the proportion of phenotypic variation attributable to genetic factors and the genetic correlations between different traits, respectively. To achieve this, we develop robust statistical methods based on estimating equations that account for familial correlations and the computational challenges posed by large pedigrees and extensive datasets. Our methods are evaluated through simulations, demonstrating satisfactory consistency and robust inference properties. Compared to simpler methods performing separate trait analysis, our approaches show a greater power through joint analysis of multiple traits. An application to the analysis of heritability and coheritability in electronic health record (EHR) data reveals substantial genetic correlations between mental disorders and metabolic/endocrine measurements, suggesting shared genetic influences that warrant further investigation. These findings have implications for understanding these conditions' etiology, diagnosis, and treatment. Chapters 3 and 4 focus on the importance of patient subtyping for personalized mental health care, particularly relevant to the substantial variability observed in mental disorders. Chapter 3 develops methods for subtyping patients with mental disorders using various data modalities and variational inference. We propose latent mixture models inspired by the Item Response Theory to handle both binary and continuous data. We also introduce Black Box Variational Inference (BBVI) algorithms to overcome the challenges of numeric integration in nonlinear models. Our numerical experiments validate the proposed methods, demonstrating that variance-controlling techniques improve convergence speed and reduce iteration variance. However, the proposed algorithm encounters limitations with latent mixture models containing binary modalities due to approximations used in non-conjugate posterior distributions resulting from the non-exponential family likelihood function. Chapter 4 investigates multi-modal integration techniques for subtyping patients using data from the Adolescent Brain Cognitive Development (ABCD) study. We introduce a Bayesian hierarchical joint model with latent variables and utilize Pólya-Gamma augmentation for posterior approximation, which enables efficient Gibbs sampling and accurate estimation of model parameters. Extensive simulations confirm the consistency of estimators and the prediction accuracy of our method. Applying these methods to patient clustering in the ABCD study provides information for identifying potential clinical subtypes within mental health, which can inform the development of targeted psychological and educational interventions, ultimately improving mental health outcomes. Keywords: latent mixture model, integrative analysis, coheritability, multi-modality, disease subtyping, variational inference, Pólya-Gamma

Indiana University-Purdue University Indianapolis (IUPUI) / Healthcare data visualization presents challenges due to its non-standard organizational structure and disparate record formats. Epidemiologists and clinicians currently lack the tools to discern patterns in large-scale data that would reveal valuable healthcare information at the granular level of individual patients and populations. Integrating geospatial and temporal healthcare data within a common visual context provides a twofold benefit: it allows clinicians to synthesize large-scale healthcare data to provide a context for local patient care decisions, and it better informs epidemiologists in making public health recommendations. Advanced implementations of the Scalable Vector Graphic (SVG), HyperText Markup Language version 5 (HTML5), and Cascading Style Sheets version 3 (CSS3) specifications in the latest versions of most major Web browsers brought hardware-accelerated graphics to the Web and opened the door for more intricate and interactive visualization techniques than have previously been possible. We developed a series of new geotemporal visualization techniques under a general healthcare data visualization framework in order to provide a real-time dashboard for analysis and exploration of complex healthcare data. This visualization framework, HealthTerrain, is a concept space constructed using text and data mining techniques, extracted concepts, and attributes associated with geographical locations. HealthTerrain's association graph serves two purposes. First, it is a powerful interactive visualization of the relationships among concept terms, allowing users to explore the concept space, discover correlations, and generate novel hypotheses. Second, it functions as a user interface, allowing selection of concept terms for further visual analysis. In addition to the association graph, concept terms can be compared across time and location using several new visualization techniques. A spatial-temporal choropleth map projection embeds rich textures to generate an integrated, two-dimensional visualization. Its key feature is a new offset contour method to visualize multidimensional and time-series data associated with different geographical regions. Additionally, a ring graph reveals patterns at the fine granularity of patient occurrences using a new radial coordinate-based time-series visualization technique.

Indiana University-Purdue University Indianapolis (IUPUI) / There has been vast and growing amount of healthcare data especially with the rapid adoption of electronic health records (EHRs) as a result of the HITECH act of 2009. It is estimated that around 80% of the clinical information resides in the unstructured narrative of an EHR. Recently, natural language processing (NLP) techniques have offered opportunities to extract information from unstructured clinical texts needed for various clinical applications. A popular method for enabling secondary uses of EHRs is information or concept extraction, a subtask of NLP that seeks to locate and classify elements within text based on the context. Extraction of clinical concepts without considering the context has many complications, including inaccurate diagnosis of patients and contamination of study cohorts. Identifying the negation status and whether a clinical concept belongs to patients or his family members are two of the challenges faced in context detection. A negation algorithm called Dependency Parser Negation (DEEPEN) has been developed in this research study by taking into account the dependency relationship between negation words and concepts within a sentence using the Stanford Dependency Parser. The study results demonstrate that DEEPEN, can reduce the number of incorrect negation assignment for patients with positive findings, and therefore improve the identification of patients with the target clinical findings in EHRs. Additionally, an NLP system consisting of section segmentation and relation discovery was developed to identify patients' family history. To assess the generalizability of the negation and family history algorithm, data from a different clinical institution was used in both algorithm evaluations.

Indiana University-Purdue University Indianapolis (IUPUI) / Living donors have contributed about 6,000 kidneys per year in the past 10 years, but more than 100,000 individuals are still waiting for a kidney transplant. Living kidney donors undergo a major surgical procedure without direct medical benefit to themselves, but comprehensive follow-up information on living donors’ health is unfortunately limited. Expert recommendations suggest capturing clinical information beyond traditional sources to improve surveillance of co-morbid conditions from living kidney donors. Currently the United Network for Organ Sharing is responsible for collecting and reporting follow-up data for all living donors from U.S. transplant centers. Under policy implemented in February of 2013, transplant centers must submit follow-up date for two years after donation, but current processes often yield to incomplete and untimely reporting. This dissertation uses a statewide Health Information Exchange as a new clinical data source to 1) retrospectively identify a cohort of living kidney donors, 2) understand their follow-up care patterns, and 3) observe selected clinical outcomes including hypertension, diabetes and post-donation renal function.

Indiana University-Purdue University Indianapolis (IUPUI) / Identifying spatial and temporal patterns of disease occurrence by mapping the residential locations of affected people can provide information that informs response by public health practitioners and improves understanding in epidemiological research. A common method of locating patients at the individual level is geocoding residential addresses stored in electronic medical records (EMRs) using address matching procedures in a geographic information system (GIS). While the process of geocoding is becoming more common in public health studies, few researchers take the time to examine the effects of using different address databases on match rate and positional accuracy of the geocoded results. This research examined and compared accuracy and match rate resulting from four commonly-used geocoding databases applied to sample of 59,341 subjects residing in and around Marion County/ Indianapolis, IN. The results are intended to inform researchers on the benefits and downsides to their selection of a database to geocode patient addresses in EMRs.

The purpose of the study was to describe the effects of an EMR on patient management in selected HIV clinics in Johannesburg. A quantitative, descriptive, cross-sectional study was undertaken in four HIV clinics in Johannesburg. The subjects (N=44) were the healthcare workers selected by stratified random sampling. Consent was requested from each subject and from the clinics in Johannesburg. Data was collected using structured questionnaires. Median age of subjects was 36, 82% were female. 86% had tertiary qualifications. 55% were clinicians. 52% had 2-3 years work experience. 80% had computer experience, 86% had over one year EMR experience. 90% used the EMR daily, 93% preferred EMR to paper. 93% had EMR training, 17% used EMR to capture clinical data. 87% perceived EMR to have more benefits; most felt doctor-patient relationship was not interfered with. 89% were satisfied with the EMR’s overall performance. The effects of EMR benefit HIV patient management. / Health Studies / MA (Public Health)