The specific time and date of all sessions, speakers names and abstracts are available in the Itinerary Planner. It is an excellent tool for building your AMIA 2012 Symposium experience.
- Regenstrief Institute’s Next-Generation Clinical Decision Support System
- HDD Terminology and Information Model Browsing Tools
- Discharge Decision Support to Identify High Risk Patients and Reduce Readmissions
- The CER-HUB: A Collaborative Platform for Multi-Institutional Comparative Effectiveness Research
- The Chicago Health Atlas: A Public Resource to Visualize Health Conditions and Resources in Chicago
- TURFS: A Comprehensive Tool Suite for Usability Evaluation and Redesign
- Development of Virtual Reality based Advanced Cardiac Life Support Training Simulator in Unreal Development Kit®
- Visual Analytics for Accelerating Discoveries in Translational Science
- MedCafe: Connecting EHRs and Creating a Personal Patient View for Clinicians
Jon D. Duke, Burke W. Mamlin, Doug K. Martin, Regenstrief Institute, Indiana University School of Medicine
The Regenstrief Institute, a pioneer in physician order entry and clinical decision support systems, has maintained a functioning CPOE since 1984. In 2011, Regenstrief unveiled a new order entry system designed to support advanced research in decision support, human-computer interaction, and clinical workflow. In this demonstration, we will be focusing on the CDS capabilities of this new system, including support for context-driven dynamic alerting, visual aggregation of multiple alerts, real-time natural language processing of notes for generation of safety and study reminders, multi-media alert content, and “learning” behaviors based on user experience with the CPOE. We will also be demonstrating our “lightning search” functionality for rapid chart review. We will conclude with a discussion of opportunities for cross-institutional collaboration, future CDS research and development plans, and a path towards open-source release of this new CPOE framework.
Senthil K. Nachimuthu, Susan A. Matney, 3M Health Information Systems, Inc.;
Mark G. Weiner, University of Pennsylvania; Stanley M. Huff, Intermountain Healthcare
The presentation will describe development, implementation and selected uses of two terminology browsing tools developed by 3M Health Information Systems and Intermountain Healthcare. Such tools advance the use of standardized terminology by supporting users’ ability to browse and select the correct terminology content. The Healthcare Data Dictionary (HDD) browser is a web-based tool that can be used by all 3M HDD customers. The ASN.1 model browser is used within Intermountain Healthcare to browse clinical information models and the associated terminology content bindings.
Eric M. Heil, Mrinal Bhasker, Matt Tanzer, RightCareSolutions LLC;
Kathryn Bowles, University of Pennsylvania
The Readmission Problem. Over 39 million hospital discharges occur each year in the United States. Failing to provide the appropriate post-acute care reduces the quality of life for patients and frequently results in potentially preventable readmissions. Nationally, the Agency for Healthcare Research and Quality (AHRQ) estimated that approximately $30 billion is spent annually on potentially preventable hospital readmissions1. Additionally according to a 2007 MedPAC report, Medicare alone currently spends approximately $12 billion per year on potentially preventable readmissions2. Discharging a patient from the hospital requires a careful, comprehensive assessment to adequately determine the patient's present and future needs, to make appropriate referral decisions, and to coordinate follow-up services; however, there is often little time for the multi-disciplinary collaboration and the careful deliberation needed. Despite the large number of discharge decisions made each year, there are no national, empirically derived decision support tools to assist in making these important decisions. To improve decision-making and prevent the high rate of readmission, health care providers need an empirically-derived evidence-based decision tool to assist clinicians, and discharge planners, as well as patients and their families.
Brian Hazlehurst, Kaiser Permanente Northwest Center for Health Research
Meeting the goals of comparative effectiveness research (CER) requires large, linked databases combining complex clinical events captured in the electronic data of multiple and diverse care organizations. Informatics challenges arise because CER requires aggregation and analysis of disparate data sources held by different institutions, each with its own representation of relevant events and concerns for protecting data. As part of the American Recovery and Reinvestment Act, the Agency for Healthcare Research and Quality invested over $100M in 11 projects to build infrastructure for CER. The CER-HUB is one of those projects, and aims to provide infrastructure for flexible, rapid, and efficient utilization of distributed, heterogeneous, electronic clinical data for CER analyses. The CER-HUB provides a means for collaborative development and validation of standardized, study-specific processors of distributed clinical data, including the capacity to extract clinical events from narrative in electronic medical records. Within this framework, study-specific data processors extend a system (called MediClass) that classifies both free-text and coded data, enabling comprehensive capture of relevant clinical events. A central website (www.cerhub.org) hosts tools for developing the data processor needed to pursue the study, and provides a library of processors that can be refined, validated, and reused by any end-user researcher.
Abel N. Kho, and John P. Cashy, Northwestern University; Bala N. Hota, Cook County Health and Hospital Systems; Shannon A. Sims, Rush University; Brad A. Malin, Vanderbilt University; William L. Galanter, MD, University of Illinois
We are creating a shared data resource to provide policy makers, researchers, community advocates and public health leaders insight into the health of the Chicago community, and identify opportunities to improve care. We specifically focused on developing tools which balanced the need for anonymity of patients and providers, while preserving uniqueness of patients. We created a stand alone Java application to perform standardized data cleaning and pre-processing, hashing of patient identifiers to remove all PHI using the HIPAA compliant SHA-512 algorithm, and measure effective match rate. Initial analysis indicates a 92-99% match rate using this approach as compared with an operational master patient index. To date we have IRB approvals and data extractions underway at three large healthcare institutions throughout Chicago, with three more pending, and parallel development of the data visualization platform. We are extracting diagnoses, medications, and laboratory tests for all patients seen at participating institutions for linking with publicly available citywide data. In this demonstration we will review the design and privacy considerations, and demonstrate use cases of clinical data aggregated across multiple institutions in Chicago and visualized across the Chicago metropolitan area.
Min Zhu MD, Deevakar Rogith, and Jiajie Zhang, National Center for Cognitive Informatics and Decision Making in Healthcare and University of Texas School of Biomedical Informatics at Houston; Muhammad F. Walji, National Center for Cognitive Informatics and Decision Making in Healthcare and University of Texas School of Dentistry at Houston
To facilitate EHR adoption and meaningful use, TURF was proposed as an EHR-specific usability framework. Under TURF, a suite of usability assessment and redesign methods are under development, such as heuristic evaluation, predictive cognitive model computation, survey assessment, concept coding, modeling and prototyping. In this presentation we will demonstrate how the TURF Suite (TURFS) can be used to conduct comprehensive usability assessments.
Development of Virtual Reality based Advanced Cardiac Life Support Training Simulator in Unreal Development Kit®
Akshay Vankipuram, Prabal Khanal, and Aaron Ashby, Arizona State University; Karen Josey and Marshall Smith, Banner Health SimET Center
Advanced Cardiac Life Support (ACLS) is a team based medical intervention performed during emergency medical resuscitation of an individual in a state of cardiac and/or respiratory failure. Teams are comprised of multiple roles with specialized tasks. The current state-of-the-art in ACLS team training involves learning to apply knowledge of American Heart Association (AHA) guidelines through in-situ mock code scenarios on a manikin-based patient simulators. We present a virtual reality (VR) team training platform as a way of augmenting the state-of-the-art in order to: Increase the frequency of ACLS training sessions, train a team of users conveniently from remote locations and reduce the associated costs. The VR team trainer has been developed on the Unreal Engine platform using the Unreal Development Kit®. The system consists of an interactive graphical simulation coupled with real time visual and auditory feedback to facilitate on screen guided messaging and team-wide voice communication. All information is stored in a remote MySQL database server and retrievable for detailed team evaluation report.
Suresh K. Bhavnani, Inst. for Translational Sciences, Univ. of Texas Medical Branch, Vickie McMicken, Management and Information Science, UHCL; Rohit Divekar, Div. of Allergy, Dept. of Med. Univ. of Texas Medical Branch
Translational science increasingly involves the discovery of complex relationships among different granularities of information such as at the molecular, phenotype, and environmental levels. However, few visual analytical systems enable translational researchers to imultaneously represent, and intuitively interact which such complex data. Here we demonstrate a novel system that was designed and implemented by integrating (1) user needs for representing
and interacting with complex biomedical data, (2) design heuristics from usability engineering and visual analytics, and (3) agile programming using HTML5 and scalar vector graphics (SVG). A formative evaluation of the system’s functionality in two disease datasets demonstrated the efficacy of the system for discovering patterns in data, but required addressing design and performance limitations.
Gail Hamilton, Jeff C. Hoyt
medCafe, developed at MITRE, is a prototype system that explores the feasibility of using a composable architecture to access medical records. A composable architecture gives the clinician the ability to rapidly build a personal view of a patient health record by use of stand alone components, each of which provides a specific capability. By isolating each component’s capability, features can be added or removed from the system without impacting the functionality of the system as a whole. The components display and interact with the data via an HL7 standard. The data itself is accessible through RESTful web services. Thus, a clinician can rapidly build a composite view of the patient’s health information, and easily adapt the interface to the patient’s particular condition. As medCafe is completely open source, with a pluggable API, we showcase a framework that allows for new components to be built and quickly added to existing system by any technically skilled person. We will also demonstrate how this composable approach facilitates modularity and agility and how using this architectural approach enables a system to evolve as needs change and new capabilities are envisioned.