This panel format is on a specific aspect of theory, application, or experience pertaining to any aspect of biomedical or health informatics, or provides interdisciplinary viewpoints that cut across traditional themes.
Sunday, November 13
3:30 p.m. – 5:00 p.m.
S01: Patient Generated Data: The Missing Link in Patient Centered Care?
N. Elhadad, L. Mamykina, Columbia University; X. Zhu,C. Chen, IBM Watson Research Center
In this interactive panel, we aim to engage the audience in an in-depth discussion on the topic of patient-generated data, as captured through smart phones and wearable devices. As patient-generated data receive increased attention from the patient, clinical, and research communities alike, there are numerous synergies and avenues for informatics research. We expect a lively discussion around existing use cases and a diverse set of axes (technological, analytics, human-computer interaction, varied stakeholders).
S03: Watson Cognitive Computing for Electronic Medical Records
M. Devarakonda, IBM Watson Health; N. Mehta, Cleveland Clinic; C. Nielson, VHA; K. Ng, IBM Research; P. Raghavan, IBM Watson Health
With the explosive growth of data, the human brain is substantially overloaded with more information than it can make sense of. This is painfully clear in patient care where critical information may be buried in the mountains of data entered in to the Electronic Medical Record (EMR) system while clinicians struggle to make sense of this information at the point of care. Cognitive computing, exemplified by Watson, offers the promise to transform this data to an intelligent assistant that provides precise insights when needed. This didactic panel examines recent advances in the specific capabilities of such cognitive computing and prospective benefits of these capabilities in patient care. The panel will discuss cognitive needs of practicing physicians. Watson capabilities of automatically summarizing a patient record and natural language question-answering on a patient record as examples of intelligent assistance cognitive computing can provide. The panel will also discuss complementary technologies from massively large scale data analysis and knowledge management that will play a role in cognitive computing. The learning objectives include developing an appreciation for working examples of cognitive computing for Electronic Medical Records, and an understanding of how these capabilities could mitigate physician’s cognitive load.
S04: Wait, My Patient is Where? Promises, Challenges, and Impact of Automated Event Notification Systems
B. Dixon, Regenstrief Institute/Indiana University Richard M Fairbanks School of Public Health/Department of Veterans Affairs; J. Shapiro, The Mount Sinai Hospital; J. Ancker, Weill Cornell Medical College; J. Vest, Indiana University Richard M. Fairbanks School of Public Health/Regenstrief Institute; S. Haque, RTI
Electronic event notifications, also called alert or subscription services, provide real-time automatic alerting of patient health care encounters to healthcare providers at other institutions. For example, an ambulatory primary care provider might use the service to sign up for alerts when his/her patients are admitted to regional emergency departments or inpatient departments. In this didactic panel, panelists will discuss use cases for event notification systems, the current state of the evidence about whether these systems affect healthcare quality or cost, sociotechnical considerations in developing these systems, and ongoing evaluation work to move the evidence forward.
S08: Building Software Platforms that Integrate with the EHR: Implementations, Frameworks, and Industry Experiences
M. Tobias, R. Grundmeier, The Children's Hospital of Philadelphia; J. Mandel, Harvard Medical School/Boston Children's Hospital; A. Neinstein, University of California San Francisco; P. DeVault, Epic
Despite billions of dollars invested nationally and decades of enthusiasm for better interoperability between electronic health record (EHR) systems, there have been few examples of successfully integrating customized software applications. In some cases, barriers to disseminating these developed tools arise even when attempting to share with other institutions using the same EHR product. The recent emergence of readily implementable standards such as Fast Healthcare Interoperability Resources (FHIR) and application platforms built upon these standards are rapidly changing the landscape. Additionally, vendor-specific “app-sharing” frameworks are providing new ways for customers to share their developments. These initiatives have taken the form of institutional research vehicles, open standards frameworks, and commercial endeavors. The panelists will describe their experiences developing and implementing platforms that can embed software-like web applications. Example projects will include the use of web-services to implement complex decision support rules, shared-decision making projects that directly engage patients, and workflow tools that facilitate communication across health disciplines.
S09: The Human Microbiome: Informatics Challenges and Opportunities
A. Alekseyenko, Medical University of South Carolina; M. Becich, University of Pittsburgh School of Medicine; T. DeSantis, Second Genome; J. Gilbert, University of Chicago; G. Gerber, Harvard Medical School
Human microbiome, the collection of all microorganisms cohabitating the human body, is making an impact on a broad range of biomedical fields. Actionable associations of the host microbiota with health conditions span many health areas, such as oral and dental health, inflammatory bowel disease, metabolic syndrome, and other systemic conditions, and cancers. The technologies for microbiome science are inherently heavy on informatics. Many leading microbiome scientists would admit inadequate involvement of informatics experts is currently the key obstacle to progress. The informatics-related challenges and opportunities include issues with representation of multi-species data, creation and validation of reference databases for genes and genomes, benchmarking and validation of computational and analytic techniques for microbiome data upstream processing and downstream analysis, and integration of microbiome data with clinical data and other multi-omic data both for research and at bed side. In addition, the White House Precision Medicine Initiative specifically calls the biobanking and sharing of microbiome cohorts an important characteristic that will enable health care providers to tailor treatment and prevention strategies to individual patients. This panel will stimulate the development of specific action plan to address the challenges and capitalize on opportunities for informaticists in translational microbiome science.
Monday, November 14
8:30 a.m. – 10:00 a.m.
S15: Transforming Clinical Documentation: Nursing and Big Data Working Group 10's Recommendations and Action Plan
C. Weaver, Weaver & Company; J. Effken, University of Arizona; K. Cochran, American Nurses Association; R. Freeman, HHS; A. O'Brien, Kaiser Permanente
In acute care, nursing documentation in today’s electronic health records systems (EHR) has become so burdensome that its sheer volume ensures that virtually no member of the care team reads it, not even nurses. EHR documentation was created in the image of the paper chart with a focus on data capture; data views that could assist clinicians in assessment, on-going evaluation, and team communication/data sharing are woefully absent. Thus, instead of promoting nurses’ core functions of monitoring and early-warning surveillance, this data-rich but information-poor (DRIP) state undermines patient safety by not supporting effective sharing of information among nurses or with the care team1. This panel will describe how nurses view current EHR systems and how they restrict nurses’ ability to provide optimal, safe, efficient, and quality patient care. Panel members will present current initiatives addressing some of these pain points and offer specific recommendations for how the best-practices resulting from these initiatives might be rapidly shared with others through a vendor-neutral library.
S16: Protecting Patient Privacy in Cyber Environments
J. Aarts, University at Buffalo/Erasmus University Rotterdam; S. Adams, University of Tilburg; B. Kaplan, Yale University; P. DeMuro, Broad and Cassel, Attorneys at Law; T. Solomonides, NorthShore University HealthSystem
Confidentiality in the medical encounter is crucial to providing adequate patient care. Health data is therefore privileged and protected by legal mechanisms. Health systems use electronic records and large-scale databases. Increasingly consumers use also IT to collect, store and share data about daily life and health behaviors. Sharing data via network-based systems or storing it ‘in the cloud’ produces multiple ‘digital selves,’ health ‘data doubles’ and ‘virtual patients.’ With so many stakeholders involved much data is produced without clear governance structures, blurring the view of what is done with the data. These problems are exacerbated through the networked, distributed nature of health data collection and convergence of protected hospital systems, commercial collection and aggregation of data and consumer health technologies. This brings patient privacy into the realm of cybersecurity. This panel explores how cybersecurity impacts the governance of critical IT infrastructures and mitigation of threats, what sociotechnical challenges are related to protection of large-scale HIT systems, how surveillance and bioethics studies seek to understand threats to personal privacy in the context of networked technologies and finally what changes to laws and regulations would be required.
S17: Extending Commercial Electronic Health Record Systems through Interoperable Applications and Services: Experiences and Lessons Learned from Four Leading-edge Institutional Programs
K. Kawamoto, University of Utah; S. Narus, University of Utah/Intermountain Healthcare; R. Bloomfield, Duke University; A. Erskine, Geisinger Health System; B. Middleton, Apervita, Inc.
Commercial electronic health record (EHR) systems such as Epic and Cerner are increasingly offering powerful interoperability frameworks for extending their capabilities through interfaced applications and software services. These interoperability frameworks include application programming interfaces (APIs) for reading and writing data from the EHR, embedding Web applications seamlessly into the native EHR user interface, and consulting external clinical decision support Web services from within the EHR’s native rules engine. Importantly, several of these APIs utilize standards-based, vendor-agnostic interfaces such as the Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR) data interface. In this panel, leaders from four leading healthcare organizations will describe how their institutions have established operational programs for developing applications and services that extend the capabilities of their EHR systems to improve clinical care. Example applications and services that have been developed by the panelists’ organizations include an integrated rheumatology management application; patient-facing medication compliance applications; a pediatric growth chart application; and a sepsis predictive modeling service. The panelists will also describe some of the key challenges and lessons learned, solicit additional insights from the collective experience of the audience, and provide guidance for colleagues seeking to extend their EHR systems at their respective organizations.
S20: Electronic Case Reporting: 3600 Perspective by Public Health, Informatics, and Healthcare Stakeholders
C. Staes, University of Utah; S. McGarvey, Northrup Grumman; S. He, Intermountain Healthcare; R. Arnold, Houston Department of Health; L. Conn, Centers for Disease Control and Prevention
This panel brings together key perspectives on the challenges and strategies used for public health case reporting, a Meaningful Use (MU) requirements for the electronic health record incentive program recently included in MU Stage 3. As public health and health care enterprises, and electronic health record vendors prepare for electronic case reporting, there are questions about the proposed strategies, current functionality, and readiness of the systems required to move from vision to reality. This panel includes presentations that will allow the audience to understand the need and systems from the public health (receiving) perspective, the informatics methods and tools being leveraged, and early experiences with performing case reporting from a health care enterprise. Questions will be posed for 30 minutes of discussion at the end to garner the perspectives and input from an audience of informatics experts working with the systems and technologies required to realize electronic case reporting.
10:30 a.m. – 12:00 p.m.
S24: The Best of Imaging Informatics Research 2016
W. Hsu, University of California, Los Angeles; C. Kahn, University of Pennsylvania
The field of imaging informatics is rapidly advancing in its ability to address challenges related to clinical big data and harnessing this information for precision medicine. In the past year, the field has experienced growth in a variety of areas including radiomics (the generation of high dimensional features from images), development of new ontologies and standards for capturing information from images and reports, and unsupervised learning from images to predict the course of a disease and treatment response. In addition, we have seen a remarkable growth in novel approaches that go beyond pixel data by integrating imaging with other biomedical data, standardizing imaging workflows, and improving the quality and utility of image-derived information in clinical practice. In this session, we will review key advances in imaging informatics research published this past year.
S26: Global e-Health: The Latin American Perspective
H. Marin, UNIFESP; D. Novillo-Ortiz, PAHO/WHO; A. Barbosa, CETIC.BR; F. Gonzalez Bernald de Quiros, Hospital Italiano
Health systems around the world face considerable challenges in providing health care services. Governments of the Latin America region have also realized the importance of e-Health and the need to create adequate policies and national strategies that could facilitate international and regional cooperation in this area. This panel will discuss the vast possibilities of ICT in health care focusing on research and initiatives developed by countries in Latin America to influence policy makers and establish indicators for fostering ICT use in health care to improve health conditions for the population.
S30: Personalized and Precision Medicine At Scale Genetic Testing and Clinical Intervention in a Large Population
K. Yale, ActiveHealth Management; L. Frey, Medical University of South Carolina; N. Walton, Washington University St Louis; D. Sands, Harvard Medical School Society for Participatory Medicine
Personalized and precision medicine has been an interesting topic of discussion, with few examples of large-scale, replicable, population-wide programs - until now. In 2014 Aetna completed one of the first health-plan sponsored programs using genetic testing in a large population to enhance diagnosis and target treatment for persons at-risk for metabolic syndrome. The original pilot was based on sound science, however the shortcomings of current knowledge on the use of genetics for diagnosis and treatment are apparent. Not as obvious, however, are the proven benefits to population health. Based on published findings from the pilot and lessons learned, we are refining the original study and expanding to other conditions. Goals of this panel include: discuss the different genetic tests useful for metabolic syndrome and other chronic conditions, determine appropriate pre-screening procedures using big data and novel predictive analytics, understand the probabilistic nature of genetic variants and implications for precision medicine, acknowledge provider, payer, and patient skepticism of the promise of genetic testing, and identify other conditions currently appropriate for personalized medicine studies base on existing genetic knowledge and return on investment. This submission is sponsored by the AMIA Genomics Workgroup
1:45 p.m. – 3:15 p.m.
S35: Building a Research Ecosystem upon the Early Success of ResearchKit
K. Mandl, Boston Children's Hospital/Harvard Medical School; R. Bloomfield, Duke University; D. Estrin, Cornell Tech; S. Friend, Sage Bionetworks; P. Pfiffner, Boston Children's Hospital
In March of 2015, Apple introduced ResearchKit, an apps framework and tool chain for conducting clinical studies that begin when subjects discover and interact with a smart phone app. ResearchKit enables investigators to readily inform and consent patients into medical and health studies. Once a subject is enrolled, ResearchKit facilitates collection of personal health data, device data and patient reported outcomes (PRO) from distributed populations. The framework launched with five initial studies that immediately demonstrated unprecedented brisk enrollment of subjects “in the wild.” Because ResearchKit is open source, extensions and adaptations began immediately. This panel explores new use cases and platforms for the technology. Topics for presentation and discussion include the role of ResearchKit in Precision Medicine, the value of connecting HealthKit to the research enterprise, transposition of the framework to the Android platform, connection of ResearchKit to the i2b2 ecosystem, and alternative enrollment pathways, including “prescribing” ResearchKit apps to selected subjects in a trial.
S36: Ensuring Reproducibility in Observational Research: Building and Sharing Knowledge Resources in the OHDSI Network
J. Duke, Regenstrief Institute; N. Shah, Stanford University; G. Hripcsak, Columbia University; P. Ryan, Janssen Research and Development
The Observational Health Data Sciences and Informatics collaborative (OHDSI, pronounced ‘Odyssey’) was formed in 2013 with the goal of creating reliable scientific evidence through large-scale analysis of observational health data from around the world. To ensure transparency and reproducibility in our research, we have created tools to share knowledge artifacts and other resources efficiently amongst OHDSI community members. We have also worked to facilitate best practices in the validation and governance of these resources. In the proposed panel, we will discuss our approach to building shareable knowledge assets, conducting research using these assets, and supporting access to these resources beyond the OHDSI community.
S37: The OpenMRS Community's Experience: A Decade of Developing and Implementing Medical Record Systems Within Constraint
T. Cullen, P. Biondich, B. Mamlin, Indiana University/Regenstrief Institute; J. Gichoya, Indiana University; H. Fraser, Leeds Institute of Health Sciences
In 2016, the Open Medical Record System (OpenMRS) community celebrates its ten year anniversary. OpenMRS, a collaboratively developed free and open platform for building electronic medical records (EMR), is currently in use in over 1200 facilities, and supports care provision to over 5 million patients. These software components are developed and supported through the OpenMRS rapidly evolving and engaged multinational open source community. This panel is designed to share the experience of the last ten years from the vantage point of key community participants who have helped lead the community over the last decade. The panel will discuss the history of OpenMRS, its current state (including a technical overview and description of the close alignment with standards development activities), implementation challenges faced, and future growth plans. The panel presentations will emphasize lessons learned, including predictive factors for a successful open source health information technology community.
S38: Mental Health/Substance Use Disorder Care, Privacy, and HIT – Can We Make it All Work?
D. Bates, Brigham and Women's Hospital; A. Busch, McLean Hospital/Harvard Medical School; M. Gordon-Nguyen, U.S. Department of Health and Human Services; P. Tang, Palo Alto Medical Foundation
Electronic health records (EHRs) enable easier, more complete sharing of health information across a patient’s providers. Mental health and substance use disorders (MH/SUDs) are common, and often cause or exacerbate other medical conditions, so patients could benefit from integration of MH/SUD information into EHRs and other health information technology. However, patients view some health conditions as more private than others and would like more control over how, or to whom, their health information is shared. For example, patients expect, and some federal and state regulators require, special privacy protections for MH/SUD care, which can pose challenges for EHRs and health information exchange (HIE). Overcoming the challenges of applying additional privacy protections to specific data categories in EHRs and HIE has important implications for patient autonomy, care-seeking, and healthcare quality and safety. It also has important implications for behavioral health providers’ ability to participate in HIE, and for our healthcare system to move towards increasingly efficient, higher quality, and safer healthcare. Recent federal and state efforts are aimed at addressing some of these issues. The goal of this interactive panel is to discuss the ethical, medical, public health, regulatory and technological challenges and controversies of addressing privacy in EHRs and HIE.
S42: Business Analysis Component of Public Health Informatics: The Missing Link
W. Williams, Centers for Disease Control and Prevention; P. O'Carroll, Office of the Assistant Secretary for Health; D. Lyalin, S. Myerburg, Centers for Disease Control and Prevention; B. Brand, Public Health Informatics Institute
Often the focus of public health informatics is narrowed down to application of IT systems and other technology solutions in public health settings. We argue that actually that scope should be broadened, first, to analysis and improvement of business practices utilized by public health agency program; and second, to a much wider application of information and computer science technology to promote and improve the public’s health, well beyond the reach of public health agency programs. This panel will provide a variety of perspectives on use of business analysis techniques for typical public health informatics projects.
3:30 p.m. – 5:00 p.m.
S45: Transforming Patient-generated Data for Wellness and Biomedical Research: From Behavioral Sensing to Decision Support
P. Hsueh, IBM T.J. Watson Research Center; S. Peterson, The University of Texas MD Anderson Cancer Center; F. Martin-Sanchez, Weill Cornell Medicine (Cornell University); K. Kim, UC Davis; C. Demiralp, IBM T.J. Watson Research Center
The rise of consumer health awareness and the recent advent of personal health IT has contributed to another shift in healthcare landscape and the accumulation of patient-generated data. Despite the rise of health consumers, the impact of patient-generated data remains to be validated. Transforming patient data for wellness and biomedical research poses many challenges. Besides the fact that patient data has been generated in large quantities that exceed what humans can consume directly, its quality and reliability vary. In this panel, we would like to review the status quo of existing evidence and present success stories where hybrid knowledge-augmented, data-driven methods are coupled with visualization tools to overcome barriers of interpretation and increase adaptability. We are also looking into decision support tools in other fields to look for inspirations of possible ways to better incorporate patient data for prevention and in extended care settings. This panel aims to share with leading practitioners and researchers at AMIA their efforts in transforming patient-generated data for wellness and biomedical research.
S50: Who Watches the Watchers: Tools and Practices for Monitoring and Measuring CDS Performance and Assessing CDS Effectiveness and Value
J. Andersen, Brigham and Women's Hospital/Partners Healthcare System; A. Wright, Brigham and Women's Hospital/Partners Healthcare System/Harvard Medical School; A. McCoy, Tulane University School of Public Health and Tropical Medicine/Ochsner Health System/The University of Texas School of Biomedical Informatics at Houston; S. Weingarten, Cedars-Sinai Health System; R. Murphy, The University of Texas School of Biomedical Informatics at Houston
Health IT (HIT), and clinical decision support (CDS) systems in particular, can improve the safety and effectiveness of health care when correctly implemented and effectively used. Over time, HIT and CDS have become more complex and pervasive, necessitating better monitoring of their efficacy and functionality. In some case, CDS alerts fire incorrectly or stop firing without warning, adversely affecting care quality and putting patients in danger. The issue now becomes, who shall watch the watchmen? In this panel we present approaches to this issue from four sites’ perspectives, together with recommendations for best practices for monitoring CDS effectiveness. We also discuss how the value of an organization’s CDS can be measured. The panelists will discuss the software and/or management plans they have developed in order to monitor CDS, the implementation of said software and plans, and the lessons learned. All four institutions represented on this panel are currently running self-developed CDS monitoring dashboards: Brigham and Women’s Hospital (Andersen), UT Physicians (McCoy), Cedars-Sinai Health System (Weingarten) and Memorial Hermann (Murphy) and have also formed CDS committees and governance approaches to oversee their large and diverse decision support implementations.
S53: Interactive Systems in Healthcare
L. Mamykina, Columbia University; M. Reddy, Northwestern University; K. Siek, Indiana University Bloomington; G. Marcu, Drexel University; L. Liu, Indiana University Bloomington
In this didactic panel, we will review an emerging and growing field of investigation within medical informatics, interactive systems in healthcare. A set of panelists with complementary areas of expertise and substantial experience in biomedical informatics and human-computer interaction will provide an overview of the field, provide examples of cutting-edge investigations, discuss grand challenges, and outline training and career opportunities. The information presented during the panel will be informed by the Workshop on Interactive Systems in Healthcare (WISH 2016) co-hosted with the SIGCHI conference on Human-Factors in Computing Systems, CHI 2016. Given the growing interest in this area, we expect the panel to be educational and allow for a lively discussion of relevant topics.
Tuesday, November 15
10:30 a.m. – 12:00 p.m.
S64: Visualization of Patient-reported Outcomes
K. Cato, Columbia University/New York Presbyterian Hospital; A. Arcia, R. Masterson Creber, Columbia University; Y. Senathirajah, SUNY Downstate Medical Center; S. Yoon, Columbia University
Patient-reported outcomes (PROs) are outcomes that patients self-report either through self-administered questionnaires or interviews. Examples of PROs include symptoms, functional status, and health-related quality-of-life. While visualizations of PROs can facilitate communication between client and clinician and inform patients’ efforts to self-manage health, the visualization creation process is often arduous. This panel is intended to provide researchers, clinicians, educators, and policymakers with an introduction to methods for development and testing of visualizations of PROs. To this end, each panelist will discuss issues specific to each phase of visualization development for PROs (design, automation/programming, testing, and selection/deployment) with a particular focus on those generated by the Patient Reported Outcome Measurement System (PROMIS).
S65: New Pathways into Biomedical Informatics: Educational Outreach Programs for High School Student
D. Boone, University of Pittsburgh; J. Finnell, Regenstrief Institute; K. Unertl, Vanderbilt University; I. Sarkar, Brown University
Biomedical informatics education in the United States started with a focus on the graduate and post-graduate level. As the field has evolved, educational opportunities have expanded into new areas, such as certificate and undergraduate programs. More recently, biomedical informatics programs have begun exploring a new frontier for the field: educational outreach to high school students and high school teachers. Questions remain about these outreach efforts. At a foundational level, why is educational outreach to high school students important to biomedical informatics? At a more pragmatic level, what types of structures are necessary to support meaningful involvement of high school students in research? Panelists representing different models of high school outreach will give an overview of program design at their institutions. The panel will then collectively discuss what has gone well in their programs, along with the barriers or challenges they have encountered. The panel will have an interactive format to enable a robust discussion around both foundational and pragmatic questions about the role of educational outreach to the high school level in the future of biomedical informatics.
1:45 p.m. – 3:15 p.m.
S70: Big Data for Healthcare and Life Sciences: Learning Useful Insights from Imperfect Data
J. Hu, IBM T.J. Watson Research Center; N. Shah, Stanford University; B. Malin, Vanderbilt University; P. Ryan, Janssen Research and Development; A. Jain, IBM Watston Health/Explorys
Electronic heath data has been generated and made accessible in unprecedented quantity and richness in the last decade. More recently, there has been much hype about applications of big data for healthcare and life sciences. This hype is due in part to a growing collection of success stories involving advanced data analytics bringing about real impact ranging from addressing unmet information needs for patients, clinicians, administrators and policy makers, to novel scientific research results from carefully constructed observational studies. However, at the same time, there has been much skepticism voiced citing limitations of data derived from insurance claims, electronic health records (EHRs), and other types of observational resources. The goal of this panel is to bring together experts from both academia and industry to share examples and thoughts on what precautions need to be taken in performing analytics on real world data, how careful modeling and investigation can help prevent bad science from being performed on imperfect data, and the types of useful insights that can be derived when the analysis is carried out appropriately. Learning objectives include an appreciation of both the promise and the limitations of observational health data, and practical guidance on how to leverage advanced analytics to harness useful insights from such data.
S71: From the Trenches – Issues Facing Clinical Informatics Administrative Clinicians in the Primary Care Setting
C. Boehm, Park Nicollet Health System; J. Tiongco, Penissula Primary Care; D. Dorr, OHSU; D. Pandita, Park Nicollet Health System
As primary care providers struggle to survive in an ever changing health care arena, 2016 is rapidly becoming the year of change. Primary Care Clinicians are increasingly taking on additional roles of Clinical Informatics Specialists and taking on challenging administrative roles along with Clinical Roles. This panel will illustrate some of the struggles faced by primary care providers in informatics related administrative roles such as CMIO and EMR champions, share examples of some of the major challenges and explore possible solutions that others can apply to their own health care setting. This will include strategies to manage change, meeting complex mandates, managing clinician expectations with changing work flows and influencing system design. As more and more Primary Care Clinicians are getting board certified in Clinical Informatics and taking on administrative roles in their health care institutions they feel that they are “in the trenches” without any support from rest of the administration due to the unique nature of their job. They also have to deal clinician opposition to change and inertia when it comes to embracing technology. This creates some unique challenges and clinicians who have done this for a while have valuable insight and possible solutions for these challenges.
S76: Report from the ED-WG Frontline: Biomedical and Health Informatics Baccalaureate (BHIB) Course
S. Khairat, University of North Carolina – Chapel Hill; G. Doyle, British Columbia Institute of Technology; I. Sarkar, Brown University; J. Williamson, AMIA
Biomedical Informatics training at the graduate and post-graduate level has expanded to include almost all clinical specialties. Whether it is a Masters, PhD, or the fellowships in the Sub-Specialty in Clinical Informatics, graduate students or returning professionals have various informatics training options. However, informatics education at the undergraduate level is not as well established. Despite initiatives such as TIGER and the Health Information Technology Scholars Program, as well as a limited number of undergraduate degree programs (e.g., at Arizona State University), there is still a need to provide interprofessional informatics education at the baccalaureate level. The Education Working Group (WG) has endorsed and supported undergraduate education as its priority since AMIA 2014 panel. Education WG members have been leading efforts to develop a Biomedical and Health Informatics course that addresses fundamental concepts and core competency skills. Panelists aim to further report on the expectations of the AMIA community with regards to using an undergraduate, interprofessional informatics course. The goal is to present students with an overview of the role of biomedical informatics in clinical, research, and operation practices. This panel will announce the completion of AMIA’s first undergraduate course geared towards various health and IT specialties including medicine, nursing, and allied professions
3:30 p.m. – 5:00 p.m.
S81: Calibration of Predictive Models for Clinical Decision Making: Personalizing Prevention, Treatment, and Disease Progression
L. Ohno-Machado, University of California San Diego; G. Hripcsak, Columbia University; M. Matheny, Vanderbilt University; Y. Wu, Duke University; X. Jiang, University of California San Diego
A large number of clinical decision support applications rely on predictive models for binary outcomes. These models usually estimate the probability for an outcome of interest (e.g., probability of breast cancer). Given that clinical decisions rely on these probability estimates, it is critical that they be well calibrated (i.e., the estimates are sufficiently close to the true underlying probability that the event will occur). Surprisingly, the evaluation of predictive models rarely includes an assessment of calibration [1–3], while essentially all include an assessment of discrimination (e.g., AUC: areas under the ROC curve). In this panel, we will describe the perils of using non-calibrated models for making clinical and administrative decisions and describe various methods to recalibrate existing models or to include calibration in model development algorithms. We will conclude with a discussion of future directions for research on model calibration.
S82: Women in Informatics Leadership Forum
R. Jacobson, University of Pittsburgh; S. Bakken, Columbia University; W. Chapman, University of Utah; V. Florance, National Library of Medicine; J. Tenenbaum, Duke University
Biomedical Informatics is a diverse field encompassing many different areas of study and providing many different potential career paths. Our field has benefited greatly by the intersection of different disciplines which enhance the diversity of Biomedical Informatics as well as the demographics of its contributors. While women have always represented a significant fraction of the discipline, there have been far fewer women in leadership positions in biomedical informatics, particularly in traditional academic roles and professional organizations. Anecdotally, we observe that there are an increasing number of women entering training programs and early academic positions in biomedical informatics. What can we do now to increase the potential for a larger cohort of women leaders in Biomedical Informatics within the next decade? This AMIA Panel will consist of two parts intended to form a single cohesive experience for participants, although participants may choose to participate in either one or both of the related events.
S87: Clinical Informatics Fellowship - Future Workforce Development
J. Maddela, U. Iqbal, University of Arizona College of Medicine, Phoenix; Y. DiAngi, Stanford University; E. Green, Regenstrief Institute; D. Ferro, The Children’s Hospital of Philadelphia
Clinical Informatics (CI) is the newest addition to the American Board of Medical Specialties (ABMS), with its first board certification issued in 2013 to 456 physicians1. The American Medical Informatics Association (AMIA) was first to recognize the importance of formal physician training to meet the growing need for a CI workforce. AMIA has designed both the Core Content and Program Requirements for the new subspecialty training programs in 20082. Today, 12 Accredited Council for Graduate Medical Education (ACGME) fellowship programs exist nationally with more programs expected to join. Little is known about the structure of the various training programs. In this didactic panel we will: 1) review examples of four ACGME accredited CI fellowship programs, 2) seek opinions from healthcare organizations CMOs, CIOs, CMIOs, and other key stakeholders in physicians’ CI education concerning the skillset needed for graduating fellows, and 3) discuss the challenges of the programs as they evolve to educate and train the next generation of CI physician workforce.
Wednesday, November 16
8:30 a.m. – 10:00 a.m.
S90: Population Health Informatics: Connecting Consumers to Care
H. Scherpbier, Jefferson College of Population Health; V. Emanuele, Wellcentive; J. Sarasohn-Kahn, THINK-Health; M. Scrimshire, Medyear, CMS
This Interactive Panel Discussion explores the challenges to use Informatics tools and methods to connect patients and consumers to their care givers and make them active participants in their care. In this panel we bring together perspectives from providers, consumers, data science, and technology firms, to engage with the audience participants to discuss technology approaches for connecting consumers and patients to their care: How far we have come? What are the challenges ahead? What are the obstacles that we need to overcome? What will it take to bring the consumer role in health care to the same level as the consumer in retail, entertainment, and other industry sectors? And most importantly, what will that do to the quality and efficiency of our care?
S93: Informatics to Transform Med Wreck to Medication Reconciliation
M. Weiner, Temple University School of Medicine; C. Weir, University of Utah; T. Adam, University of Minnesota; E. Chou, Drexel University College of Medicine
Appropriate conduct of Medication Reconciliation is a fundamental component of meaningful use, and an important contributor to patient safety. In the ambulatory setting, medication reconciliation aims to ensure that the patient and provider are in synch with the medication regimen. In the inpatient setting, the process mandates that medications from a prior treatment location are appropriately addressed in the transition from one level or location of care to another. At discharge, medication reconciliation requires thoughtful decisions about continuing inpatient medications into the ambulatory setting, with appropriate recognition of prior outpatient medications that should be stopped or changed. This panel will discuss some of the informatics, workflow, communication, and human challenges that interfere with optimal completion of medication reconciliation and several institutional initiatives designed to address these challenges.
S94: Secure Record Linkage for Precision Medicine and Patient Centered Outcomes Research
D. Meeker, University of Southern California Keck School of Medicine/RAND Corporation; A. Kho, Northwestern University; T. Ong, University of Colorado, Denver; X. Jiang, University of California San Diego; J. Doctor, University of Southern California
The ability to link records across multiple sources enables richer data for a wide variety of applications in patient-centered outcomes research and precision medicine. The standard implementation of a linkage system uses personally identifying information from different sources to perform a probabilistic or deterministic identity match, with manual inspection to resolve uncertainty. While there are established commercial record matching systems for treatment purposes and Health Information Exchange, regulations, risks and risk tolerance differ in the context of research. Furthermore, if existing systems can be improved upon from the standpoint of both performance and privacy protection, disclosures required for manual adjudication can be limited. Each of the panelists will describe their experiences with implementing a privacy-preserving record linkage (PPRL) approach, and then discuss the pragmatic implications, and share results of a comparative evaluation. The methods discussed will include using seeded one-way hashing, Garbled Circuits, multi-party GMW (Goldreich-Micali-Wigderson), Intel® Software Guard Extensions, weighted and order-preserving encryption methods, and approaches using count queries across parties to infer a total ordered data from partial orders. We will describe how different data elements, weighting strategies, and order-preserving hash functions can impact match performance. Panelists will directly address issues associated with the HIPAA Privacy Rule and implementation of different methods in the presence or absence of a trusted broker. The panel will present considerations for privacy, security, trust, accuracy, and computing resources that merit attention, giving attendees an overview of the different feasible options for privacy-preserving record linkage.
S98: How Informatics are Being Applied in Industry: Challenge and Opportunity
E. Chapman, Department of Veterans Affairs; Z. Cai, Celgene; J. Caballero, Amino; M. Palchuk, TriNetX; S. Ebadollahi, IBM Watson Health
Due to its multi-disciplinary nature, informatics is a field with applications in a variety of settings, among them the secondary use of health care data in industry represents exciting opportunities and challenging career paths for informaticians. These individuals are required to possess innovative thinking and mastery of the cutting-edge technologies, to work on complex, real-world problems in areas such as big data mining, novel treatment development, precision therapies, clinical decision support, mHealth applications and device data analytics, with the ultimate goal of each patient having access to the highest quality and personalized care options. On this panel, representatives from AMIA’s corporate members will share various use cases and discuss how the informatics innovations are being applied in industry to address challenges and make break-throughs. The panel will also share their own career path and how they arrived at their current positions, discuss the important characteristics they are looking for in career seekers and the challenging opportunities residing within the industry. Following individual presentation, the panelists will take questions from the audience on a range of topics related to how informatics is being applied and the role of informaticians in industry.
S99: PEDSnet: From Building a High-quality CDRN to Conducting Science
C. Bailey, CHOP; M. Kahn, S. Deakyne, CHCO; R. Khare, CHOP; K. Deans, Nationwide Children's Hospital
Collaborations across multiple institutions are essential to achieve sufficient cohort sizes in clinical research. Within the national PCORnet research network, clinical data research networks (CDRNs) aggregate electronic health record (EHR) data from multiple sites and provide a unified platform to conduct large-scale studies, while dramatically increasing the capabilities to conduct a wide range of scientific studies. There are two critical facets of building a functional CDRN data resource: (i) to address the characteristic data provenance and quality issues that arise from the data’s orientation toward clinical operations rather than research, from semantic heterogeneity across systems, and from clinical data peculiarities, and (ii) to assess and improve the usability of network data to conduct scientifically valid studies. This panel focuses on a pediatric-specific CDRN, PEDSnet, to discuss its recent evolution from initial data collection and generic data quality assessments, to tailoring data characterization to scientific utility, developing a data-compatible approach to constructing computable phenotypes and conducting pilot observational studies.
10:30 a.m. – 12:00 p.m.
S101: Engaging Older Adults in the Design, Implementation, and Evaluation of Health IT
A. Turner, University of Washington/Northwest Center for Public Health Practice; G. Demiris, University of Washington; A. Lazar, Northwestern University ; B. Reeder, University of Colorado
Over the past decade there has been a growing interest in encouraging individuals to take a more active role in their health and health information management. This movement towards patient empowerment is closely tied to patient centered technologies including patient portals, decision making tools, home based telehealth, sensor technologies and patient access initiatives such as Open Notes. Although older adults make up 18% of the US population and are the largest users of healthcare, their uptake of information technology has traditionally lagged behind that of other age groups. Unfortunately, the needs of the largest group of health care consumers are often not considered because of difficulties involving this age group in research and the misperception that older adults are technology adverse. This panel demonstrates the potential of information technology to improve the health and independence of older adults, and showcases the application of participatory design methods to empower older adults in their health care management and decision making. Panel members will discuss their experience in conducting research with and for older adults in various settings including independent living, retirement communities and assisted living facilities.
S104: The Medication-use Process: Current Challengs and Potential Solutions
W. Liu, Vanderbilt University Medical Center; L. Idemoto, Virginia Mason Medical Center; J. Poikonen, Avhana Health; S. Alameddine, Nova Southeastern University College of Pharmacy; S. Nelson, Vanderbilt University Medical Center
The medication-use process of prescribing, perfecting, preparing, dispensing, administering, monitoring, and educating provides many opportunities for health information technology to improve patient safety and process efficiency. However, this complex process is often misunderstood and can have serious downstream consequences in the accuracy and communication of medications, resulting in patient safety concerns or harm. It is imperative that informaticists understand the medication-use process to minimize potential adverse events and unintended consequences. The panel will describe some nuances that affect the medication-use process, such as socio-technical aspects, medication order timing, electronic controlled substance prescribing considerations, and interoperable medication lists.
S109: Standardizing Research Common Data Elements: Initiatives, Exchange Formats and Current Use by Patient-level Trial Results Databases
V. Huser, M. Sincan, NIH ; D. Shmueli-Blumberg, EMMES Corporation; R. Hess, University of Utah
Sharing of de-identified patient level data from clinical trials is increasingly becoming a norm and this trend is increasing the importance of research Common Data Elements. CDEs have the potential to increase the value of trial data by making it easier to integrate data across multiple trials. CDEs can reduce study start-up costs, improve the quality of collected data, and facilitate cross study comparisons, data aggregation and meta-analyses. This panel will (1) describe current CDE initiatives, highlight several trial data sharing platforms (including their use of CDEs), and describe CDE informatics standards; (2) describe a case study a site using REDCap Electronic Data Capture system and their CDE experience; (3) describe a case study of a trial data sharing platform of the National Institute on Drug Abuse; and (4) showcase a multi-site observational study that first selected suitable CDEs and implemented collection of trial data directly within an Epic EHR.
S110: A Forum on Qualitative Research in Biomedical Informatics: Controversies, Challenges, and Opportunities
L. Novak, Vanderbilt University; R. Valdez, University of Virginia; T. Veinot, University of Michigan; J. Talmon, HI-way; N. Lorenzi, Vanderbilt University
Qualitative research has a long history in biomedical informatics. It has often been used in investigations regarding the social influences on emerging technologies, the impact of technology on individuals and groups, and in improving design and implementation of informatics tools. However, researchers using qualitative approaches may face controversies and challenges in obtaining funding, interfacing with researchers using different methods, and in communicating their work. This panel will engage the audience in addressing critical issues facing qualitative researchers and discussing emerging opportunities that impact the direction of the field. Panelists will briefly discuss their experiences in the following areas: Dr. Lorenzi will discuss organizational qualitative research; Dr. Valdez will discuss home- and community-based qualitative research; Dr. Veinot will discuss funding strategies and Dr. Talmon will discuss issues and opportunities for publication of research findings. The subsequent interactive discussion with the audience will build on the panelist comments, creating a forum to explore a variety of challenges and opportunities.