Day 1 :
Keynote Forum
John Fox
Oxford University, UK
Keynote: ARTIFICIAL INTELLIGENCE IN MEDICINE: DATA SCIENCE MEETS KNOWLEDGE ENGINEERING
Time : 09:45-10:30
Biography:
John Fox is an interdisciplinary researcher with interests in computer science, AI, cognitive science and medical informatics. After degrees at Durham and Cambridge Universities he worked with AI founders Allen Newell and Herbert Simon at Carnegie-Mellon, and with Ulric Neisser at Cornell University in the USA. After a period with the MRC back in the UK he joined the ICRF (now Cancer Research UK) where his group made many theoretical and practical contributions in decision science, cognitive theory and medical informatics, and founded The Knowledge Engineering Review. He has published widely (see Researchgate.org) and has led the foundation of several medical AI companies (Expertech, InferMed, Deontics). A current passion is OpenClinical.net, an open access, open source knowledge
repository that uses AI and crowdsourcing techniques to capture and disseminate actionable knowledge of best medical practice.
Abstract:
Mathematical methods for supporting clinical decision making have a long history but have not achieved wide adoption because: (1) acquiring the large data sets that are needed to build decision models has been difficult and (2) the perception that mathematical models of decision-making are medically naive. Big data techniques address the problem of acquiring large data sets, but the problem remains that abstract data and algorithms are “black boxes” that are likely to be unintelligible and mistrusted by healthcare professionals and their patients. Artificial Intelligence and knowledge engineering offer many
techniques for clinical decision making and management of care that are complementary to analytic tools and there is good evidence that these methods have practical value and user appeal. The Dentists and OpenClinical approach is based on a naturalistic model of medical expertise formalised in first-order logic (Fox and Das 2000). The Perfume modelling language exploits this approach and the standard syntax and semantics are in the public domain (Sutton and Fox, 2003). Proforma has been used successfully in diverse applications and medical specialties, in some cases at scale and with high impact. Knowledge engineering has much to offer to our ever more pressured health services by increasing the quality and safety of care at reduced
cost, but more is needed to establish the “learning health systems” that are being widely discussed. Combining Proforma with the power of data mining and machine learning may offer a sound foundation for safe and clinically appropriate decision support services and a platform for implementing rapid learning systems in health care. Fox J and Das S, Safe and Sound: Artificial Intelligence in Hazardous Applications, AAAI and MIT Press, 2000. Sutton D and Fox, J “The syntax and semantics of the Proforma guideline modelling language” J Am. Medical Informatics Association, 2003.
Keynote Forum
Filiberto Di Prospero
Di Prospero’s Obstetrics & Gynecologic Center, Italy
Keynote: FETAL WELL-BEING ASSESSMENT IN OUTPATIENT PREGNANT WOMEN USING WEARABLE DEVICES
Time : 10:30-11:15
Biography:
Filiberto Di Prospero is Graduated as Medical Doctor in 1982 at Ancona University, Italy; Postgraduate School in Obstetrics and Gynecology, Endocrinology and Metabolism. Filiberto Di Prospero is a considered an expert in Medical Informatics and Computer Science with the passion to develop innovative solutions in health care. Some significant contributions: the first diagnostic support software for predicting human ovulation in 1990; the realization of one of the most important Italian websites on feminine health in 1999 (SaluteDonna.it); the first app for fetal well-being assessment, integrated into outpatient obstetric assistance in 2012; introduction
of the new concept of “proximity medicine” in 2013.
Abstract:
Intrauterine activity is one of the most ancient and important signs of fetal well-being. It’s perception reassures pregnant women and improves the fetus-maternal relationship. In 2012 Filiberto Di Prospero as a member of the Apple Developer Program published Fetal Activity Monitor (FAM), an app for smartphones and tablet PC that helps pregnant women in fetal movements, counting (kicks count); FAM for the first time introduced an interpretative support, integrable in more complex fetal surveillance systems. The new software (code name “Rose”) that will be presented, provides a better diagnostic accuracy considering fetal Biorhythms. The Rose project is a great step forward and is an example of modern proximity medicine, that merges advanced scientific knowledge in fetal life with informatics and the modern wearable electronic technology.
- Biomedical Informatics | Imaging Informatics | Health Informatics | Clinical Informatics | Telemedicine | Nursing | Health Information Technology | Health System
Chair
John Fox
Oxford University, UK
Co-Chair
Filiberto Di Prospero
Di Prospero’s Obstetrics & Gynecologic Center, Italy
Session Introduction
Yuh-Fong Hong
The University of Texas, USA
Title: INTEGRATING INFORMATICS AND TECHNOLOGY IN NURSING RESIDENCY PROGRAM
Time : 11:35-12:05
Biography:
Yuh-Fong Hong is an assistant professor of the School of Nursing at the University of Texas Health Science Center at Houston. Hong’s expertise includes health care quality improvement using health information technology (HIT) and applied health informatics. He leads the integration of applied informatics in nursing programs. Hong is selected to serve on the interprofessional education program team to facilitate collaboration on the EHR usage at the university level. Hong presents his teaching and research findings at national and international professional conferences.
Abstract:
With the rapid implementation of EHRs and health information technologies (HITs) in hospitals, it’s important that entryto-practice nurses acquire HIT competency to provide patient-centered care, collaborate with interdisciplinary teams, and improve care quality. The most recent Nursing Residency Program (NRP) accreditation standards from the Commission on Collegiate Nursing Education (CCNE) add informatics and technology in the NRP to expand resident's knowledge and skills acquired in their prelicensure programs to analyze and implement best practices in effective use of information technology to safely manage patient care. A recent survey result of 86 senior student nurses conducted to understand the gap between academic settings and real-world practice of using HITs indicated 5 challenges of utilizing HITs in practice: a) technical issues (58%), b) patient privacy concerns (51%), c) human errors (34%), d) less time for patient care (26%), and e) communication between disciplines (15%). Using the CCNE NRP standards and the survey results, an online learning module was created to assist nurse residents in learning applicable HITs in their practice. Six learning objectives include: a) understand health/clinical information systems in hospitals, b) use effective electronic communication for team-based care delivery, c) evaluate information resources for evidence-based practice, d) apply health information technology in care quality improvement and error reduction, e) comply with policies and confidentiality laws when using social media, f) exercise safety, security, and emergency backup plan in HITs. In addition to the lecture, gamified simulation
activities and required self-assessment quizzes are integrated to help nurse residents achieving the identified learning objectives.
Henriette Loeffler-Stastka
Medical University of Vienna, Austria
Title: CASE-BASED BLENDED E-LEARNING SCENARIOS – ADEQUATE FOR COMPETENCE DEVELOPMENT OR MORE?
Time : 12:05-12:35
Biography:
Henriette Loeffler-Stastka has completed her MD at the age of 24 years from Medical University Vienna, is a Psychiatrist and Psychotherapist/Psychoanalysis and Associate Professor of Psychoanalysis and Psychotherapy and the Medical University Vienna. She is the deputy director of the postgraduate unit of the teaching center, developed medical curricula, including a case-based e-Learning program and different postgraduate programs. She is Head of the Advanced University Course for Psychotherapy Research. She has published more than 110 papers in reputed journals and has been serving as a editorial board member of repute.
Abstract:
Learning, competence development and research processes in medicine need several strategies to facilitate new diagnostic and therapeutic ways. The optimal collaboration between creative design thinking and biomedical informatics provides innovation for the individual patient and for a medical school or society. Fast processes in observing and understanding are needed to generate ideas
for the development and testing prototypes: First, declarative knowledge has to be acquired and collected in basic medical sciences, knowledge that is in fact available and can be accessed on the conscious and preconscious level in long-term memory. Second, associative learning describes the formation of neuronal connections between a neutral stimulus and a second. This conditioning is an important form of learning and discovering and founded in neural associations. Third, polythematic-crosslinking thinking is needed as ability to link information (thoughts, symbols, images, scenes) in a meaningful way. These steps are a typical intellectual ability of gifted learners and researchers, creative enough that they succeed to combine previously seemingly unrelated areas to each other and drive innovation. Utilizing the flexibilities of an e-learning platform, a case based blended learning (CBBL) framework consisting of A) case based textbook material, B) online e-CBL with question driven learning scenarios and C) simulated patient (SP) contact seminars was developed and implemented in multiple medical fields. Satisfaction with this kind of learning lead to formation of innovative learning and publication groups that began to develop critical reflection on curricular development, patient-centered clinical reasoning processes and research questions – both in students and teachers.
Kim Powell
The Ohio State University, USA
Title: Atlas-based segmentation of temporal of bone anatomy
Time : 12:35-13:05
Biography:
Kim Powell completed her Ph.D at the Ohio State University in 1992. She is an assistant research professor in the Department of Biomedical Informatics at OSU and the director of Small Animal Imaging for the University. Powell is an imaging scientist who has extensive research experience in microscopy, small animal, clinical imaging and image analysis. She has published more than 50 papers in reputed journals.
Abstract:
Surgical approaches, such as mastectomy and cochlear implantation, are the primary treatment for a wide range of hearing and balance disorders. We have developed a surgical simulator that includes volume visualization, haptic modeling and psychomterics to help train clinicians complex temporal bone surgical techniques. X-ray computed tomography (CT) images are used in the surgical simulator and manual segmentation of landmark regions for the simulator is laborious and requires an expert reviewer. Therefore, we have implemented an atlas-based approach to automatically segment 15 critical structures in X-ray CT images of 43 cadaver specimens (22 left, 21 right). First, a rigid-body registration is performed using the whole temporal bone. Then a second rigid-body registration is performed using a smaller region-of-interest (ROI) that includes the otic capsule, ossicles, facial nerve and chorda
tympani. The structure of the bone within this ROI is highly conserved between subjects and temporal bone structures can be directly identified using the reference atlas. To automatically segment surface structures of the temporal bone, such as the signed, tegmen, internal and external auditory canal, we perform a multi-resolution B-spline deformable registration using a Gaussiansmoothed whole bone image. Visual inspection of our atlas-based segmentation approach indicates that it is highly consistent with manual segmentation performed by expert reviewers and can be performed in a matter of minutes as opposed to hours for manual segmentation. Accurate automated segmentation of temporal bone anatomy allows us to further develop the training simulator for use in pre-surgical planning using clinically obtained CT images of patients.
George Tolomiczenko
University of Southern California, USA
Title: Teach one, do one, see one: App design for medical students
Time : 14:05-14:35
Biography:
George Tolomiczenko is an experienced clinician, researcher, teacher and administrator helps him in his Administrative Director role to guide and run the Health, Technology and Engineering program at USC (HTE@USC). After an interdisciplinary which undergraduate degree at Caltech, he trained in Clinical Psychology at Boston University, Public Health at Harvard University and Business Administration at the University of Toronto. He is now focused on developing USC’s interdisciplinary collaborative strengths applied to medical device and process innovation. He teaches courses designed to form and train teams linking engineering and medicine to create innovative technology and start-up companies.
Abstract:
The ubiquity and diversity of apps in everyday life have crossed thresholds that make them expected conveniences in almost all environments including healthcare. Sensors that gather health and behavioral information relevant to the overall aim of personalizing treatment are pervasive and cheap. Much data can now be gathered, but the devil still stubbornly resides in the details of developing algorithms that convert data into actionable information worthy of informing decisions related to health maintenance or improvement, illness prevention and treatment. At the same time, the aspiring app developer must ensure the targeted consumer or patient need is linked to a market compelling enough to attract investment of resources sufficient to create and develop a sustainable business around the app. The lure of digital health can outweigh lack of experience with coding, information technology development and technology assessment resulting in students dropping out of medical school before even starting with a residency program. This presentation describes an approach to immersing students in an experientially-focused elective that teaches a focused approach to need articulation, guides medical students to do customer discovery and development interviews and requires the students to create a wireframe mockup of their app that allows customers and partners to see a prototype. Students who complete the electivehave a more realistic sense of how this sphere of health informatics operates and are better equipped to make informed choices that will determine the future trajectory of their careers.
Subrata Acharya
Towson University, USA
Title: Towards the design of a trusted storage platform for effective big data management in healthcare systems
Time : 14:35-15:05
Biography:
Subrata Acharya received her Ph.D. in Computer Science from the University of Pittsburgh, 2008 & M.S. in Computer Engineering from Texas A&M University,College Station, 2004. She has published over 50 peer-reviewed book chapters, peer-reviewed papers at international conferences and in journals in the area of computer and information security. Acharya has obtained significant extramural funding to support her scholarship efforts, including $450K as PI and $230K as co- PI. of particular note is Acharya’s US patent 7966655 B2, awarded in 2011 with Wang Ge and Greenberg for Method and apparatus for optimizing a firewall. Acharya has also developed new courses in the area of health care informatics. She has mentored various students who have appeared as co-authors on her papers, and has supervised numerous undergraduate research projects, masters’ graduate projects, and doctoral dissertation studies.
Abstract:
Apache Hadoop has the potential to offer powerful and cost effective solutions to big data analytics in health care systems; however, sensitive data stored within an HDFS infrastructure have equal potential to be an attractive target for exfiltration, corruption, unauthorized access, and modification. Pairing Apache Hadoop distribute file storage with hardware based Trusted Computing mechanisms based on TCG standards has the potential to alleviate risk of data compromise and maintain information compliance of federal and/or state governmental standards. With the growing use of Hadoop to tackle big data analytics involving sensitive health care data, an HDFS cluster could be a target for data exfiltration, corruption or modification. By implementing open, standards based Trusted Computing Technology at the infrastructure and application levels; a novel and robust security posture and protection is presented to address the issue. A discussion of the motivation for research on this topic, a threat model and evaluation of a targeted Advanced Persistent Threat against HDFS is presented and a set of common security concerns within HDFS is addressed through infrastructure and software involving integrity validation and data-at-rest encryption. To accomplish these goals, technology from the Trusted Computing Group, such as the pervasively available Trusted Platform Module is used. In addition, a discussion of design
considerations in building an encryption framework for Hadoop in a trustworthy manner is presented along with a description of performance and security results of experiments, creating an encryption scheme for Hadoop utilizing hardware key protections and AES-NI for encryption acceleration (based on data obtained from a real world large scale (> 400 beds) healthcare system). This work includes an evaluation of the recently implemented crypto framework for Hadoop and independent test of the performance claims of AES-NI is regarding mitigating encryption performance overhead.
Jeroen S de Bruin
Medical University of Vienna, Austria
Title: Patient and physician perspectives on nutritional monitoring using a smartphone application for cancer outpatients
Time : 15:25-15:55
Biography:
Jeroen S. de Bruin is graduated from Leiden University, The Netherlands, as a PhD in biomedical informatics, with specialties including medical data mining and ontologies. Later on he started his post-graduation at Leiden University Medical Center with the subject workflow optimization in proteomics. In 2011, he started working at the Medical University of Vienna, where he has continued his research into clinical decision support systems and mobile health, thereby focusing on
infection control in the intensive care setting.
Abstract:
In recent years, the number of mobile health (mHealth) applications available have increased dramatically. These applications register data related to a person’s mental and/or physical state, e.g. for disease self-management, cessation of unhealthy habits, and promotion of healthy behaviour. Given the increased use of such applications, a potentially huge amount of personal health information (PHI) is generated. Besides self-management of health, PHI from mHealth applications could also help to improve the quality of healthcare delivery. In this study, we evaluate a system that integrates PHI from a nutritional monitoring mHealth application for cancer outpatients with data gathered in clinical routine, for the use in a clinical decision support system (CDSS) for nutritional triage. In a clinical pilot study, we recruited 25 oncology outpatients to use a forementioned mHealth application. Data recorded from this application were forwarded to a data repository of the Medical University of Vienna, where they were processed by the CDSS. The results could be accessed directly from the Vienna General Hospital information system. Afterwards, a qualitative questionnaire was taken among patients and medical experts involved with the system. Among patients (N=25), 91% found the application useful as a remote tool for detecting cancer-related malnutrition, and about 75% indicated it should be institutionalized. Among clinicians (N=5), the CDSS was perceived a useful, and enabled them to initiate nutritional interventions sooner. However, its usefulness was limited still as regular nutrition monitoring is not a compulsive part of the overall care workflow.
Steven H Shaha
Center for Public Policy & Administration, USA
Title: EMR-enabled improved clinical, cost and satisfaction outcomes for bedside caregivers
Time : 15:55-16:25
Biography:
In 2016, Steve Shaha was introduced internationally as “premier healthcare outcomes researcher globally in breadth and depth.” With 35+ years of studies, teaching, speaking and advisory work, Steve has addressed the needs of a long list of recognized organizations on four continents, including 11 foreign governments. He has 250+ conference presentations, 125+ peer-reviewed publications, three invited chapters internationally in 2015, and four books. Steve is a full Professor with 4 graduate degrees and has taught or lectured at 30+ universities in 6 countries, among them Harvard, Cambridge (UK), the King’s College, UCLA, Columbia and Cornell.
Abstract:
Background: The acquisition and use of electronic medical records (EMRs) has grown remarkably. Of yet statistically valid verification of beneficial clinical, cost and efficiency outcomes is rare or absent. The vacuum seems largest for bedside caregivers, mainly the hourto- hour vigilant Nurses with critical roles in patient outcomes. Healthcare needs EMRs that favourably impact bedside clinicians.
Methodology/Process: Five healthcare organizations rigorously undertook six projects to improve Nursing impacts through EMR programmable and adaptable solutions. Nurse, physican and IT-professional teams defined crucial Nurse-related documentation impratives that contributed to internal-EMR computations for vigilance, alerting and protection. The IT-proffesional then programmed/adapted the EMR to better value and incorporate Nursing decimation, better alert beside Nurses for key activates, and enable improved physician actions on Nurse-provided documentation.
Findings/Impacts:
Safety and Medication Errors
Pharmacy data and Nurse surveys corroborated (all p<0.01):
• 83.2% of Nurses rated 4 or 5 (1-5 scale) – 38.5% higher versus vbaseline - for improved safety, clinical management, documentation,
communication and “the 5 rigthts”
• 71.7% reduced medication administration errors.
• 83.6% decreased Falls with injuries
• 69.2% fewer acquired level 3&4 pressure ulcers
Nurse Efficiency & Efficacy
• 44-minute decrease in documentation time away from patients
• 21-minute decrease in overall documentation time
• 29.3% increased patient-direct time, shifting from 35% patient-direct to 52%
Conclusion: Programming and adapting the EMR for sensitivity for Nursing needs substantively and significantly improved Nursing efficiency and efficacy. Such adaptability through in-house teams maximizes role importance and recognized criticality. When Nurses
win, everyone wins.