The following half-day tutorials will take place during ICHI 2017.
Details of the tutorials are as follows.
Half-day Tutorial 1: Process Mining and Analytics for Care Pathways
This tutorial demonstrates the application of a range of process analytics techniques to the study and improvement of clinical care pathways. The past decade has seen increasing interest in care pathway design, documentation and dissemination but formal methods for describing, monitoring and assessing pathways have yet to be established. Outside of healthcare other industries have well established techniques for business processes and there is much scope for translating these to fit the unique nature of healthcare. In particular data analytics, data mining, and machine learning have converged on a set of technologies called process mining which has the potential to lead to a step-change in using e-health record data to mine and manage care pathways. The tutorial presents an iterative method developed with the UK NHS and the Connected Health Cities programme that combines process mining with other process analytics methods including process modeling, process simulation and business process improvement. The session is highly interactive and based on a series of hands-on exercises around a worked example supplemented by case studies of completed work. Links to further study are provided and the aim is encourage further research and build a global community of practice.
Owen Johnson, School of Computing, University of Leeds, UK
Bio: Owen Johnson is a computer science academic at the University of Leeds. His research interests are focused on the process mining of e-health record data to improve care pathways. He leads the e-health records research group at the University of Leeds, supervises seven PhDs in pathway mining (areas included cancer, musculoskeletal diseases, diabetes, elderly care, cardiovascular and emergency room) and is a principle or co-investigator on £14m of e-health record research grant funding. He is currently leading an international consortium developing novel methods for comparing care pathways and outcomes between different health systems. Owen held a joint post in Health Informatics in the School of Medicine for eight years during which time he set up the Leeds MSc in Health Informatics, the ResearchOne e-health records data service and the foundations for the now well established Leeds Institute for Data Analytics. He won the UK Research Councils award for the best contribution to the knowledge base in 2014 for his work creating ResearchOne. He has had a successful career as a software developer and in management roles at British Telecom, London Transport, Amoco, Thames Television, Forte and Sodexho. He is a founder and director of X-Lab, a successful e-health software company.
Half-day Tutorial 2: EasyCIE: A Development Platform to Support Quick and Easy, Rule-based Clinical Information Extraction
Clinical information extraction (IE) is a natural language processing (NLP) subtask leveraged aimed to identify, encode, and extract pertinent clinical variables from free-text notes into structured data, such that computers can utilize this data by itself or in conjunction with other structured data to support a task (e.g., quality improvement, deep phenotyping, clinical decision support, etc.). However, conducting clinical IE often requires NLP experts to create a customized NLP pipeline for specific tasks and variables, for example, identifying all mentions of fever from clinical notes. This tutorial will introduce and aid audience members by giving them hands-on experience following an NLP “cookbook” instructional to developing a rule-based clinical IE system leveraging a light-weight tool, EasyCIE, which reduces the clinical IE learning curve for non-NLP- experts.
Jianlin Shi, Department of Biomedical Informatics, University of Utah, USA
Jianlin Shi, MS, MD, has over 6 years of UX design, 4 years of clinical NLP research. Before he came to US, he practiced surgery in China for over 5 years. He has expertise in the development of smart, efficient, and general-purpose tools including rule-based NLP tools such as EasyCIE (an easy-to- use, rule-based information extraction system), RuSH (a highly accurate, Rule-based Sentence segmenter using Hash), and FastConText (an extremely efficient implementation of the ConText algorithm) to support clinical IE subtasks and system development.
Danielle Mowery, Department of Biomedical Informatics, University of Utah, USA
Bio: Danielle Mowery, PhD, is Director of NLP Service Line in the Department of Biomedical Informatics at the University of Utah. She has 10 years of NLP experience working with clinical domain experts to create domain-specific information models, to generate annotated reference datasets, to develop clinical IE systems, and to evaluate clinical IE systems that support biomedical research, population health, and other use clinical cases.
Half-day Tutorial 3: An Integrative Model of the Sociotechnical Lifecycle Methods to Design, Develop or Optimize Clinical Systems
Healthcare is a complex sociotechnical system, involving the interaction of both people and technology. Designing technical systems that are effective, are adopted and that function seamlessly with larger organizational structures requires a thorough understanding of both the human and technology requirements. In this tutorial, we present the major components of the sociotechnical lifecycle, specifically identification of functional requirements, system design, implementation and evaluation. We will also present various methods that can be used to support each of the components. The first methods covered will be cognitive task analysis (CTA) and ethnographic observation, which are tools commonly used to extract information about the users’ needs and the development of functional requirements. Then, we will discuss methods of participatory and contextual design. In a third section of the tutorial we will describe the principles of formal heuristic evaluation and usability testing. Finally, we will present methods for program evaluation and implementation, such as tools for evaluating system usefulness, methods for assessing impact on user’s decision-making, and the use of simulation testing. In each section, we will engage participants in hands-on activities, consult with participants in designing their own sociotechnical study, and supply an annotated reference file. Qualitative and experimental methods will be covered and discussed. In the last section, we will discuss the overall challenges of implementing systems in clinical environments, the importance of establishing relationships with clinical and IT operations and the new domain of commercial clinical system optimization.
Charlene Weir, Department of Biomedical Informatics, University of Utah, USA
Bio: Charlene Weir is Associate Professor in the Department of Biomedical Informatics at the University of Utah and the Associate Director for Education and Evaluation for the SLC GRECC as well as an investigator in the SLC IDEAS VA HSR&D center. Dr. Weir has a PhD in social cognitive psychology as well as a MS in Nursing. Her area of interest is in the application of motivation and cognitive theories as applied to the design and use of health information technologies. Dr. Weir will focus her presentation on cognitive task analysis (CTA), and ethnographic observations to extract useful data in the system design process.
Frank Drews, Department of Psychology, University of Utah, USA
Bio: Frank Drews is Professor of Cognitive Psychology at the University of Utah. Frank's work has focused on Applied Cognitive Psychology and Human Factors in Patient Safety. Dr. Drews spent significant time investigating Human Factors issues such as the limits of human performance and the impact of performance breakdowns on provider performance. He will discuss the main concept related to participatory and contextual design, iterative design process.
Guilherme Del Fiol, Department of Biomedical Informatics, University of Utah, USA
Bio: Guilherme Del Fiol is Assistant Professor in the Department of Biomedical Informatics at the University of Utah. Dr. Del Fiol has over 15 years of experience in clinical informatics, especially in clinical decision support (CDS) tools. Dr. Del Fiol is a co-chair of the Health Level Seven (HL7) CDS Work Group and has authors several CDS standards, particularly the HL7 Infobutton Standard. He is also the project lead for OpenInfobutton, an open source suite of infobutton tools and Web services, which is in production use at several healthcare organizations in the United States. Dr. Del Fiol will present an example of a research project that applied user centered design and simulation studies as formative evaluation methods.
Heidi Kramer, Department of Biomedical Informatics, University of Utah, USA
Bio: Heidi Kramer is Research Associate in the Department of Biomedical Informatics at the University of Utah. Dr. Kramer spent over 20 years designing and implementing a variety of software applications before returning to school to study cognitive psychology. Her interests are in applying knowledge of cognitive processes to improve computer-assisted data visualization, information search and decision making. She will talk about formal heuristic evaluations and usability studies.
Damian Borbolla, Department of Biomedical Informatics, University of Utah, USA
Bio: Damian Borbolla is Assistant Professor in the Department of Biomedical Informatics at the University of Utah. Damian is a physician with an MS in Medical Informatics. He has experience in the development and implementation of clinical systems in different countries and cultures. During his talk, Dr. Borbolla will describe the main challenges in system implementation and optimization strategies. He will also describe some of the Department of Biomedical Informatics Sociotechnical service line’s evaluation projects.