• AMIA 2018 Annual Symposium

    Data, Technology, and Innovation for Better Health

AMIA 2018 Annual Symposium Panels

For updated information please visit the online program.

Sunday, November 4

S01: Panel – Innovative Informatics Research and Practice in the Era of Commercial Electronic Health Records: Our Experience Using Epic

A. Wright, D. Bates, Partners HealthCare/Brigham and Women's Hospital/Harvard Medical School; A. Auerbach, University of California San Francisco; A. Weitkamp, Vanderbilt University; E. Kirkendall, Cincinnati Children's Hospital Medical Center/University of Cincinnati

Over the past five years, several leading informatics organizations have transitioned from self-developed electronic health records (EHRs) to commercial systems. For example, Partners HealthCare, Vanderbilt University Medical Center and Eskenazi Health (formerly Wishard Hospital, partners with the Regenstrief Institute) have moved to Epic, Intermountain Healthcare has moved to Cerner, and the VA health system is also negotiating to move to Cerner. Historically, these organizations had considerable control over their information systems, as they directed their development and roadmap. As organizations make the transition from self-developed to commercial systems, new challenges arise, especially around informatics research, practice and innovation.

In this panel, we present the experience of four organizations which are actively innovating in Epic, the most widely used commercial EHR: Partners HealthCare, University of California San Francisco, Cincinnati Children’s Hospital and Vanderbilt University Medical Center.

The presenters will detail their organizations’ experience innovating on the Epic platform through several case studies, introduce the range of technical approaches for developing and integrating innovative tools in Epic, discuss Epic’s approaches to working with innovators and cover issues in organization governance of innovation activities. Though the focus of the panel is the experience with Epic, the lessons learned apply to all EHRs.

S02: Panel – Usability Testing Methods for mHealth Apps Designed for Disadvantaged End-users

S. Stonbraker, Columbia University School of Nursing; E. Heitkemper, Columbia University; H. Cho, M. Beauchemin, R. Schnall, Columbia University School of Nursing

Mobile health (mHealth) applications (apps) have the potential to improve health behaviors and subsequently improve patient outcomes. Usability testing is fundamental to the uptake and efficacy of mHealth apps for improving end-users’ health. Numerous methods to assess the usability of mHealth apps are available in the literature, thus selecting an appropriate method can be challenging. Additionally, usability testing may be particularly difficult among disadvantaged populations particularly those with low socio–economic status (SES) groups who have limited economic resources, or minimal educational backgrounds, as these groups are more likely to have lower health literacy and difficulty using technology. In this didactic panel, four presenters will each discuss a discrete usability evaluation technique and a case study to illustrate the use of this methodology for evaluating mHealth apps with disadvantaged groups. By the end of this panel, participants will be able to: 1) articulate the importance of usability evaluations, especially among low SES groups; 2) differentiate between usability evaluation techniques; 3) describe evaluation techniques for achieving different levels of usability evaluations; and 4) select an appropriate usability evaluation method for mHealth apps in specific contexts with varying end-users.

S03: Panel – National Patient-Centered Clinical Research Network (PCORnet) Distributed Research Network (DRN): Questions Answered from Preparatory-to-Research to Comparative Effectiveness

J. Sturtevant, Harvard Pilgrim Health Care Institute; D. Louzao, Duke Clinical Research Institute; K. Marsolo, Cincinnati Children's Hospital Medical Center; K. McTigue, University of Pittsburgh Medical Center; L. Curtis, Duke Clinical Research Institute

The National Patient-Centered Clinical Research Network (PCORnet) conducts multi-site preparatory to research (PTR) queries, which lead to observational research studies and clinical trials using data that have been standardized across 80 Network Partners (NPs) into a Common Data Model (CDM), which allows a single analytic program to ask the same question of all NPs. Utilizing a distributed research network (DRN) such as PCORnet, requires in-depth knowledge of the CDM, the underlying partner data, and best practices in analytic programming so that a query will execute across 80 contributing NPs with diverse computing infrastructures. To help potential users navigate this complexity, the PCORnet Coordinating Center (CC) operates a “Front Door,” which provides a number of intake services, including requests for data. The Front Door formalizes such requests into research questions, which are then operationalized by the analytic team into PTR or research queries. The CC has supported dozens of PTR questions and 3 PCORnet Demonstration Projects, including the PCORnet Bariatric Study (PBS). This panel will describe how research questions are turned into PTR or complex study queries within PCORnet, and will include perspectives from CC personnel as well as study investigators.

S04: Panel – The Healthcare Services Platform Consortium: Building a Marketplace for Healthcare Applications

P. Haug, Intermountain Healthcare/University of Utah; K. Rubin, University of Utah/U.S. Department of Veterans Affairs; S. Huff, Intermountain Healthcare, University of Utah; S. Hasley, University of Pittsburgh; R. Freeman, Interopion, Inc

The growing use of electronic health records (EHRs) in patient care has created an increasing need for new applications capable of improving the healthcare experience and outcomes for both patients and caregivers. The success of innovative software tools in this marketplace is frustrated by the lack of well-established, unambiguous standards allowing access to the information stored in these EHRs. The Health Services Platform Consortium (HSPC) has as its purpose the identification, evaluation, and refinement of healthcare standards with the express goal of fashioning interface specifications that can be used to create fully interoperable applications enabling plug-and-play transfer between different EHRs.

Our panelists include participants from standards organizations, the U.S. Department of Veterans Affairs, academic medical centers, and business dedicated to providing services to the medical software market. In this panel, they will introduce the HSPC. They will describe its organization and goals, its approach to standards refinement and endorsement, applicable experience with standards–oriented application design and development, and the HSPC Sandbox, an environment for developing interoperable clinical applications. They will introduce and describe HSPC’s overarching goal: to enable a marketplace for interoperable applications and to speed delivery of innovative tools that can extend the capabilities of all standards compliant EHRs.

Monday, November 5

S12: Panel – Surveillance Methods for EHR Safety Hazards

N. Muthu, Children's Hospital of Philadelphia/Perelman School of Medicine at the University of Pennsylvania; A. Fong, Medstar Health National Center for Human Factors; D. Ferro, Children's Hospital of Philadelphia; E. Orenstein, Children's Healthcare of Atlanta

Existing methods for electronic health record safety surveillance often miss “hidden” hazards because they rely on the user alone to 1) perceive the risk, 2) report the risk and 3) explicitly identify health information technology (HIT) as a cause. Systems relying on user identification of risks also require significant resources to identify the errors that are potentially caused by HIT. Additionally, many error types are never perceived or reported by users, leading to a substantial underestimate of the overall amount of safety hazards and omission of some types of hazards entirely. In this didactic panel, researchers will present innovative approaches to the identification of these previously “hidden” hazards, including automated classification of HIT related medication errors from safety event reports, use of alert logs to identify alert malfunctions and process gaps, and identifying discrepancies in clinical documentation that portend a high risk of potential harm. These advances are important steps in developing a more comprehensive framework for researchers and informatics practitioners to proactively surveil the EHR for safety hazards.

S13: Panel – Blockchain Fundamentals for the Healthcare Professional: Realities, Use-cases, and Future Implications

R. Boodoo, M. Diaz, University of Illinois at Chicago; J. Taylor, Walter Reed National Military Medical Center; A. Alsadi, University of Illinois at Chicago

The panelists are practicing physicians and informaticists who believe that there is fertile ground for innovation at the intersection of blockchain and healthcare. Blockchain is a distributed ledger technology that is likely to transform the processes of information exchange for numerous industries, including healthcare. In 2008, blockchain, along with its use of cryptographic techniques, was made famous by the novel process of mining digital currencies such as Bitcoin. While many misconceptions about blockchain technologies have stemmed from an association with Bitcoin, the underlying principles can empower many applications beyond the exchange of value via digital currency. Some valuable characteristics include an immutable audit trail, decentralization of systems, and distributed data. The panelists are blockchain enthusiasts and physicians who are actively working on applying this technology to positively influence the delivery of healthcare. The fundamental technology underpinning blockchain applies to many healthcare use-cases and can potentially lower costs, increase efficiency, restructure security and engage more stakeholders. Most importantly, patient incentives can be reimagined in ways never before possible. Blockchain continues to evolve, adding transactional capabilities via the incorporation of variable functionalities, such as smart contracts. We aim to demystify blockchain and articulate its value added to the healthcare industry.

S14: Panel – Preemptive Clinical Decision Support: Delivering Precise Information Clinicians Will Actually Use to Prevent Harm

J. Hoffman, St. Jude Children's Research Hospital; J. Peterson, Vanderbilt University Medical Center; N. Muthu, Children’s Hospital of Philadelphia; H. Dunnenberger, NorthShore University HealthSystem

Clinical decision support (CDS) in actual practice is primarily reactive, interrupting clinicians when prior clinical decisions are rated by rule based knowledgebases as suboptimal; such an approach is typified by drug safety alerts which often have disappointing clinical impact. Ideally, CDS is designed to proactively guide clinicians toward data–driven decisions through highly integrated, seamless, and interactive interventions. As part of this panel, examples of preemptive clinical decision support will be presented along with data about clinician response. The panelists will explore how existing systems to manage panel–based pharmacogenetics, improve safety of prescribing, and detect early clinical deterioration meet the goals of preemptive CDS. The panelists will discuss how these examples could be redesigned or further improved. The importance of human factors and visualization to improve preemptive CDS will be discussed throughout the session.

S15: Panel – Standardization of Clinical Decision Support: An Essential Strategy for Interoperability and Optimization of Patient and Provider Experience

E. Barone, VA-Fayetteville VAMC; D. Montella, L. Wedemeyer, K. Campbell, Veterans Health Administration

The Veteran’s Administration (VA) currently has 9 million enrolled patients Department of Defense (DoD), as of 2017, had 1.3 million active duty soldiers who represent potential VA patients over time. The VA electronic medical record, VistA, and its EHR module known as CPRS, were developed and enhanced largely by clinicians over the last 25 years. In June of 2017, Secretary of Veterans Affairs Dr. David Shulkin announced VA’s intent to adopt a commercial EMR similar to the one being adopted by the Department of Defense – a maneuver intended to facilitate free flow of information between the two organizations which share patients. Regardless of eventual outcome, the announcement of this plan created an urgency to evolve clinical decision support (CDS) in the legacy VA electronic health record with intent to streamline and optimize CDS to improve patient and provider experience, reduce clinician cognitive burden, and reduce implementation burden by standardizing the representation of clinical statements and integrating the terminologies we use to populate those clinical statements. This last goal is a challenge whose success is essential to achieving true interoperability of CDS not only between VA and DOD, but among all health care entities, regardless of EMR platforms.

S23: Panel – The Intersection of Data Science, People, and Organizations in Health Care: An Interactive Discussion of Challenges and Solutions

L. Novak, Vanderbilt University; R. Valdez, University of Virginia; C. Walsh, Vanderbilt University; H. Salmasian, Brigham and Women's Hospital; E. Wynn, Ronin Institute

Data science holds promise for producing innovative clinical insight and transforming health care operations. However, the real world is complex and holds variation in information practices, undiscovered clinician information and visualization needs, and institutional politics. Data scientists encounter these and other challenges in design and implementation of data scientific solutions. Increasingly, data scientists are collaborating with researchers who specialize in social and organizational issues to study and improve the process of translating health-related practices for the purpose of data science interventions. This interdisciplinary, interactive panel will discuss real-world challenges of data science in health care and produce insights based on the panel and audience expertise.

S24: Panel – The Road to Broader Adoption of CDS in a Learning Health System

A. Boxwala, Elimu Informatics, Inc.; B. Middleton, Apervita Inc.; J.M. Overhage, Cerner Corp.; J. Adler-Milstein, University of California San Francisco

This interactive panel will discuss broad adoption of CDS and, in particular, the barriers and the solutions to implementing CDS at scale. The discussion will be framed around the use of CDS as a key component of a learning health system, which requires wider use of CDS than is the current state. The panel will describe the barriers to implementing CDS at scale and the different solutions to those barriers. These solutions include implementation of new CDS capabilities, newer technological approaches to delivering CDS using web services, a framework for creating and managing CDS knowledge bases at scale, and standards that enable sharing of these knowledge bases. The panelists bring a range of perspectives that includes an electronic health record systems developer, a CDS platform developer, a knowledge representation and knowledge translation analyst, and a learning health system scientist.

S25: Panel – Women in AMIA – Resources for Emerging Leaders

G. Savova, Boston Childrens Hospital; M. Johns, The Monarch Cetner; N. Lorenzi, Vanderbilt University Medical Center; P. Brennan, National Library of Medicine; R. Crowley Jacobson, UPMC

This panel provides multiple perspectives on empowering emerging leaders through a variety of resources. The panel is organized by Women in AMIA taskforce and follows in the footsteps of similar panels at 2016 and 2017 AMIA annual symposia led by the taskforce. The topics of the panel are (1) The Intersection of Mindfulness and Leadership to explore insights about the evidence-based science that supports the benefits of mindfulness in the workplace, as well as introducing the results of evolving research suggesting that mindfulness helps women avoid stereotypical constraints of gender roles, (2) Strategic Thinking about Your Problems to explore proven strategies and tactics that can be used for short and long term success, (3) Negotiations to cover the difference between distributive negotiation (dividing up a fixed resource, also called “zero-sum”) and integrative (or “principled) negotiations and discuss how winning negotiations lead not only to clear paths for action that can be supported by all parties but also how getting the negotiation stance right improves working relationships, and (4) Team Building to discuss a variety of aspects of team-building including how to improve your skills in creating and sustaining outstanding informatics teams. The panelists are leading female informaticians.

S26: Panel – Data Science in Biomedical Informatics Education: Critical Problems and Innovative Solutions

H. Hochheiser, University of Pittsburgh; J. Mostafa, University of North Carolina; N. Gehlenborg, Harvard Medical School; S. McWeeney, Oregon Health & Science University; V. Florance, National Library of Medicine

The growth of interest in data science presents educators with opportunities in preparing researchers and professionals for future changes in biomedical informatics research and training. The NIH Strategic Plan for Data Science, released as a draft in March 2018, identifies workforce development for biomedical data science as a key goal. In 2017, NLM Training Programs in Biomedical Informatics and Data Science launched an initiative aimed at developing new data science educational materials, curricula, and instructional delivery modalities. The initiative now has reached approximately its halfway mark and recently a meeting was held for 14 training programs in which updates on critical barriers, lessons learned, and new strategies for data science pedagogy and training were collected from all the program leaders. A key goal of the panel is to provide a summative update to a wider audience, based on the cumulative knowledge from all the participating NLM training programs, on core skills and competencies associated with data science, proposed approaches for addressing data science, and the role that data science plays in biomedical informatics education, research, and practice. Panelists will share perspectives on data science education efforts at their institutions and discuss the relationship between data science and biomedical informatics.

S34: Panel – Lessons Learned from Using Informatics to Address the Social Determinants of Health

M. Cantor, NYU School of Medicine; R. Gold, Kaiser Permanente; L. Gottlieb, University of California San Francisco; H. Kharrazi, Johns Hopkins School of Public Health; T. Cullen, Regenstrief Institute

As the clinical impact of the social determinants of health (SDH) is increasingly recognized, more institutions are implementing HER-based approaches to addressing these determinants. Organizations are developing the infrastructure to integrate SDH-related data into clinical practice, population health management, and research. However, the U.S. healthcare system is still in the early days of dealing with social determinants, so there is no standard roadmap for collecting SDH-related data, using the data in billing or diagnosis, or acting on the determinants in the clinic or through referrals to community-based organizations. This panel will discuss the experience of four different institutions in implementing programs to address the SDH and the core role of informatics in their approaches. Lessons learned by these groups can help others implement their own programs successfully.

S35: Panel – Embracing Interdisciplinarity: A Commemoration of the Work of Dr. Samantha Adams

C. Kuziemsky, University of Ottawa; L. Novak, Vanderbilt University Medical Center; C. Petersen, Mayo Clinic; A. Solomonides, NorthShore University HealthSystem Research Institute; J. Aarts, Erasmus School of Health Policy & Management

When Dr. Samantha Adams passed away in December 2017 she left behind a scientific legacy that bridged multiple disciplines including computer science, information science, medicine, social science, and popular culture. Of greater importance, Sam was also a mentor, colleague, and friend to our community. One of Sam’s biggest contributions was providing an interdisciplinary perspective on medical informatics to enhance our understanding of how people, processes, and technology interact in our desire to improve human health. This panel presentation summarizes Sam’s work in four areas: social theory, social constructionism and patient empowerment, consumer and pervasive informatics, and discourse analysis.

S36: Panel – Using HL7 FHIR to Improve Standardization and Interoperability of Common Data Models for Clinical and Translational Research

G. Jiang, Mayo Clinic; J. Duke, Georgia Institute of Technology; D. Meeker, University of Southern California; M. Rocca, Food and Drug Administration; H. Solbrig, Mayo Clinic; S. Murphy, Harvard Medical School

Fast Healthcare Interoperability Resources (FHIR) is becoming a next generation standards framework created by HL7 and leveraging the latest web standards for exchanging electronic health records (EHR) data. The clinical research datasets indexed by the HL7 FHIR will not only enable existing data/datasets integration, but also provide a path towards future data integration as the EHR vendors provide FHIR-compliant interfaces to enhance interoperability. This thematic panel will highlight five projects that are developing technology infrastructure and tools leveraging the HL7 FHIR standard for the harmonization, standardization and applications of a variety of common data models (CDMs) developed in the clinical and translational research communities. Attendees will learn about novel open source FHIR-based informatics platform, and how the FHIR standard is used in data sharing platforms. Attendees can engage with panelists on the FHIR-based CDM harmonization and interoperability challenges and potential solutions in support of clinical and translational research applications.

S37: Panel – Interoperable Apps and Services to Extend the EHR: Perspectives on Current State and Future Vision from Leading Vendors and Healthcare Systems

K. Kawamoto, University of Utah; K. Shekleton, Cerner Corporation; I. Vetter, Epic Systems Corporation; S. Narus, Intermountain Healthcare

The demand for new features in the electronic health record (EHR) is never-ending, yet the resources available at any single EHR vendor or healthcare organization are finite. In recognizing this challenge, leading EHR vendors are supporting a new paradigm where various stakeholders can develop and share high–value applications and services using interoperability frameworks including FHIR, SMART, and CDS Hooks. In this panel, leaders from major EHR vendors and innovative healthcare systems will describe their experiences in making interoperable apps and services a reality, including through the establishment of app marketplaces; SMART on FHIR apps for enhanced data visualization and clinical decision support; and CDS Hooks services for addressing important clinical priorities such as the opioid epidemic. The panelists will then share their vision of a future in which interoperable apps and services make it easy to provide the best care possible for every patient every time, with apps and services effortlessly shared across organizations and seamlessly integrated with workflows. The panelists will conclude by describing the most pressing challenges to enabling this future vision, and they will engage the audience to discuss how we can best work together to move forward as a community.

S45: Panel – Informatics Challenges, Solutions, and Opportunities for Public Health Electronic Case Reporting

J. Loonsk, The Johns Hopkins University/CGI Federal; C. Staes, University of Utah; J. Hartsell, Utah Department of Health; J. Stamm, Epic

While case reporting is a core public health function, its electronic advancement has been relatively delayed because of a series of technical and policy challenges. Electronic case reporting is enabled by the widespread adoption of electronic health records, but a standards-based architecture, decision support, and other components were also needed to address these challenges before widespread roll-out could occur. With many of these informatics solutions now available, some secondary issues have developed, but new opportunities have also been enabled. This panel will discuss resolved informatics issues, corollary challenges, and new opportunities for public health electronic case reporting from several perspectives in its information supply chain. Architecture and standards, electronic health records, reporting logic and decision support, and public health agency surveillance systems will all be represented to share electronic case reporting challenges and solutions as well as discuss some of the new opportunities that have presented.

S46: Panel – The Linchpin of Interoperability: Challenges and Solutions to Patient Record Matching

R. Rudin, RAND Corporation; S. Grannis, Regenstrief Institute; M. Tripathi, Massachusetts eHealth Collaborative; J. Smith, AMIA; B. Moscovitch, The Pew Charitable Trusts

Failures in patient record matching are persistent and imperil our ability to reap the benefits of federal and state investments in health information technology and interoperability. This panel will discuss a portfolio of studies undertaken in the pursuit of a national strategy for enhancing patient record matching across health care organizations. Panelists will first present possible solutions and evidence supporting them, including those that involve standardizing demographic data elements, and those that involve patients more actively in the matching process, such as through the use of new identifiers or patient-facing software tools. Panelists will then discuss challenges and barriers that providers may face in adopting such solutions, and the patients’ perspectives on the various possible solutions. The session will engage the audience in discussing the various proposed solutions in terms of effect on interoperability and health information exchange, privacy concerns, controversial aspects of specific solutions such as biometrics, and the role of government. The discussion will deeply explore the complex tradeoffs among various potential solutions.

S47: Panel – Automatic Text De-Identification: How and When is it Acceptable?

S. Meystre, Medical University of South Carolina; D. Carrell, Kaiser Permanente Washington Health Research Institute; L. Hirschman, J. Aberdeen, The MITRE Corporation; P. Fearn, V. Petkov, Surveillance Research Program/National Cancer Institute; J. Silverstein, University of Pittsburgh School of Medicine

The adoption of electronic health record (EHR) systems is growing at a fast pace in the U.S., and this growth results in very large quantities of patient clinical data becoming available in electronic format, with tremendous potential, coupled with growing concern for patient confidentiality breaches. Secondary use of clinical data is essential to fulfill the potential for personalized healthcare, improved healthcare management, and effective clinical research. De-identification of patient data has been proposed as a solution to both facilitate secondary use of clinical data, and to protect patient data confidentiality. The majority of clinical data found in the EHR is represented as narrative text clinical notes, and de-identification of clinical text is a tedious and costly manual endeavor. Automated approaches based on Natural Language Processing have been implemented and evaluated, allowing for much faster de-identification than manual approaches, with comparable protection. However, despite these advances, health care providers have been reluctant to use automated de-identification to share clinical data for secondary use. On the one hand, IRBs and privacy officers are wary of the possible risks; on the other hand, researchers are concerned about obfuscation of critical medical information in the de-identification process.

S48: Panel – Family Planning, FHIR, and the Office of Population Affairs: Creating Interoperability Standards for Large-scale Public Health Programs

S. Matney, Intermountain Healthcare; S. Hasley, ACOG; R. Handzel, UPMC; B. Heale, Intermountain Healthcare

The Office of Population Affairs (OPA) manages over 4000 sites across the U.S. that supply Family Planning services to low income individuals. These sites are supported by over 100 different EHR products. In order to improve OPA’s services, and measure that quality of care provided, they need to gather encounter-level data from these sites. A common data model would simplify both the documentation and the collection of this data. Fast Healthcare Interoperability Resources (FHIR) provides a standard data model. Creating a FHIR Profile to constrain the data provided by a FHIR server would provide a uniform, project-specific data stream that could be sent to a registry. This panel will discuss the creation and validation of FHIR Profiles, including the review of the OPA’s data dictionary, creation of appropriate logical models, and the design, creation and validation of FHIR profiles for those data elements. Over 130 FHIR profiles were created. The various workflows, processes and tooling to accomplish these tasks will be reviewed in detail. Challenges that are currently unresolved will also be discussed.

Tuesday, November 6

S56: Panel – Improving Care Delivery in Nursing Homes with Informatics

M. Ozkaynak, B. Reeder, University of Colorado; G. Alexander, University of Missouri; E. Ramly, University of Wisconsin; R. Valdez, University of Virginia

Nursing homes (NHs) are critical components of a national health system. Information technologies (IT) can improve health outcomes and facilitate the connectedness of NHs with the rest of the health system. However, adoption of IT in these settings is low compared to primary and acute care settings. Although the lessons on design and implementation of IT from other settings can be helpful, guidelines and frameworks cannot be directly translated because of the unique organizational and care delivery characteristics of NHs. This panel will begin with a discussion by each panelist of their experience with informatics research related to NHs, describing various aspects (e.g. workflow, usability, resident safety, adoption) of design and implementation of informatics interventions. This will be followed with an interactive session, where the audience and panelists will create a list of challenges of designing and implementing IT in NHs, followed by a discussion of known effective and potentially new strategies to address these challenges.

S57: Panel – Improving Expressivity in EHR Data Retrieves – Focus on Improving eCQMs

F. Eisenberg, ESAC International; A. Smith, National Committee for Quality Assurance; J. Lehner, PCPI; L. Anderson, The Joint Commission; K. Lesh, Battelle

The Centers for Medicare & Medicaid Services (CMS) quality reporting and value-based purchasing programs have included electronic clinical quality measures (eCQMs) to evaluate clinical performance for the past several years. Providers have experienced challenges interpreting and implementing some of the eCQMs. For the calendar year 2019 reporting/performance period, CMS will provide revised measures using Health Level Seven International’s (HL7’s) Clinical Quality Language (CQL), which allows greater expressivity and more precise, unambiguous logic statements. The CQL expressions can be shared with other measures and with clinical decision support (CDS) artifacts to enable real-time workflow enhancements that improve performance of process and outcomes. Panelists will describe improvements in measure precision and expressivity enabled by CQL and provide insights into expectations for more straightforward implementation based on these updates. Panelists will further discuss how some eCQMs have also been tested using Fast Healthcare Interoperability Resources (FHIR) in HL7 FHIR Connectathons and in real world pilots.

S58: Panel – Building Data Capacity for Patient–Centered Research

S. Smith, S. Lumsden, Health and Human Services; P. Dullabh, NORC at the University of Chicago; D. Meeker, University of Southern California Keck School of Medicine; B. Johnson, AcademyHealth

Patient-centered outcomes research (PCOR) and comparative effectiveness research (CER) are central to producing insights that improve treatment options and inform health care decision making by patients, providers and caregivers. This vision requires a data infrastructure that facilitates data access for research, and generates clinically relevant findings that can be disseminated and translated into practice. The Assistant Secretary for Planning and Evaluation (ASPE) was given the responsibility of coordinating across federal agencies to “build data capacity” for PCOR/CER through investments in data infrastructure via the Office of the Secretary Patient–Centered Outcomes Research Trust Fund (OS-PCORTF) This panel will present on the informatics-related achievements to date and continued opportunities for advancing PCOR data infrastructure. The panelists will describe the strategic framework that was instrumental in building the current portfolio of OS-PCORTF projects, the projects and their tangible contributions to data infrastructure and current policy initiatives like 21st Century Cures, as well as strategic priorities for the remaining years of the OS-PCORTF, and reflections from an on-the-round PCOR researcher.

S59: Panel – Improving Longitudinal Care Coordination with Person-centered Electronic (e-)Care Plans: Lessons from Efforts in Chronic Kidney Disease

J. Norton, National Institute of Diabetes and Digestive and Kidney Diseases; D. Carlson, Clinical Cloud Solutions LLC, ; E. Gallego, EMI Advisors; K. Ali, A. Narva, National Institute of Diabetes and Digestive and Kidney Diseases; T. Cullen, Regenstrief Institute

By supporting longitudinal coordination, communication, and collaboration across healthcare teams in diverse settings, e-care plans have potential to improve care safety and quality and reduce adverse events, diagnostic delays, and healthcare costs. Chronic kidney disease (CKD) is an ideal population in which to deploy and test an e-care plan, given the complexity of care, numerous transitions, need for interdisciplinary collaboration, and many comorbidities associated with the disease. The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) is collaborating with experts from the nephrology and informatics communities via its CKD e-Care Plan Working Group to develop and implement a person-centered e-care plan incorporating data elements essential for CKD care. The Working Group conducted interviews with individuals living with CKD to understand their values, and developed patient and clinician personas to guide identification of data elements and corresponding data standards (i.e., LOINC, SNOMED-CT, ICD-10, CPT, RxNorm). These personas and data elements have been used during HL7 FHIR and C–CDA Care Plan standards related testing events, yielding C-CDA output and FHIR resources. Several challenges and lessons learned have been identified during development and testing that may be relevant to the creation of standardized and person-centered e-care plans for other chronic diseases.

S66: Panel – The Future Ain't What it Used to Be: Negotiating the Changing Digital Health Ecosystem

D. Kaufman, Arizona State University; C. Smith, University of Wisconsin-Madison; A. Keselman, National Library of Medicine; A. Murcko, Arizona State University

The questions of how to judge the quality of online health information and how to support the public in the difficult task of information evaluation has been a central issue in informatics for many years. However, with the emergence of new Internet technologies, including the spread of multimedia channels and the ubiquity of social media, the need for new approaches to critical evaluation has taken on a new dimension and urgency. Fake news, which names a troubling phenomenon of deceptive communication, has become ubiquitous across media platforms. This raises important concerns for public health and consumer health informatics. The panelists are collaborators in a research program focusing on the assessment of quality of online health information from “unconventional sources,” including those that make claims unsupported by the medical establishment. At the conclusion of this panel, audience members will be able to: (1) Characterize the dimensions of the problem confronting health consumers and patients; (2) Describe evolving criteria to render judgments about the soundness of information resources; (3) Discuss the role that education can play in empowering lay health information seekers; (4) recognize the challenges that physicians are confronted with in helping patients negotiate the morass of online health information.

S67: Panel – EHR Log Data: An Untapped Health Data Goldmine for Clinical Informatics Research?

C. Dymek, Agency for Healthcare Research and Quality; M. Tai–Seale, University of California San Diego; V. Patel, New York Academy of Medicine; J. Adelman, Columbia University Medical Center, New York-Presbyterian Hospital; J. Adler-Milstein, University of California San Francisco

EHR systems generate user log data, which can be productively employed to improve the efficiency, quality, and safety of healthcare. While usage of these potentially illuminating data remains limited, given the promising potential, methodological approaches for rigorous use of these data need to be discussed and developed. This session addresses these issues by presenting three different use cases that reveal the research value of EHR log data in a range of topical areas, including provider burden, workflow effectiveness, and health IT safety. The panel will also engage in a discussant–led conversation about the opportunities, limitations, and methodological challenges regarding use of these data. The learning objectives for this panel include: increasing awareness of the value of EHR log data for answering various clinical informatics and clinical research questions, understanding methodologies for how best to use these data in a research setting, and understanding the potential for leveraging these data in novel informatics applications such as measuring the impact of interoperability.

S68: Panel – Implementation Science for Global Health Informatics Projects

Y. Quintana, Harvard University; P. Biondich, Indiana University School of Medicine; H. Fraser, Brown University; A. Kanter, Columbia University; J. Holmes, University of Pennsylvania

The implementation of informatics programs in low- and middle- income countries remains difficult for many reasons and sustainability is even more challenging. Many emerging economies are beginning to appreciate the role of health information technology, not only within health program oversight activities but as an effector of improved quality and efficiency. As such, eHealth leadership within these countries are actively attempting to rationalize national HIT architectural approaches as a precursor towards large-scale implementation. This panel will describe approaches and lessons learned in the implementation of global health informatics programs and create an opportunity for an interactive discussion with the audience on the on-going challenges in the sustainability of global informatics programs.

S77: Panel – Data Standardization in Cancer: Challenges and Opportunities

R. Belenkaya, Memorial Sloan Kettering Cancer Center; W. Campbell, University of Nebraska Medical Center; R. Chen, Columbia University; M. Gurley, Northwestern University Clinical and Translational Sciences Institute; S. Huff, Intermountain Healthcare; G. Randhawa, National Cancer Institute; C. Reich, Real World Insights, IQVIA; M. Rocca, Food and Drug Administration

This panel will introduce and comparatively review efforts in data standardization in the Oncology domain by the Food and Drug Administration (FDA), HL7 Clinical Information Modeling Initiative (CIMI), Nebraska Lexicon (SNOMED CT), and Observational Health Data Sciences and Informatics (OHDSI). This panel will discuss specific challenges in standardizing oncology data, review the standards in development and their value proposition for addressing these challenges in clinical decision making, data sharing, and research. It will also address the need to unify these efforts and maximize returns on resources invested through innovative interoperability solutions. Panelists will share perspectives that have informed the original conceptualization of their solutions, and will address the topics of adoption, utility, and evolution of these standards.

Panelists will interact with audience and among each other to discuss opportunities that exist in current high profile national initiatives (such as the AACR Project Genomics Evidence Neoplasia Information Exchange (GENIE), National Cancer Institute (NCI) Community Oncology Research Program (NCORP) and Cancer Care Delivery Research (CCDR) for data standardization and large-scale multi-disciplinary research collaborations.

S78: Panel – Sensor Based Assessment for Learning and Care Quality: The New Frontier

A. Hughes, University of Illinois at Chicago; S. Zajac, Houston Methodist Hospital; M. Rosen, Johns Hopkins Medicine; B. Najafi, Baylor College of Medicine; R. Ahmed, Summa Health System

As informatics solution(s) advance knowledge, patient safety, and quality improvement efforts, the need for practical and real-time assessment of safety and performance is critical. Current assessment techniques leveraged in healthcare organizations can be time consuming and impractical, delaying feedback which is important for continued learning and contributing to diminished response(s) to safety hazards. One area of informatics which has received attention is methods for unobtrusively gathering information related to care quality and performance through use of sensors. While various sensor based technologies (e.g., RFID badges) can be leveraged to assess important behaviors (such as teamwork or timely administration of an antibiotic), implementation of sensor based technologies (SBT) requires input from a variety of disciplines in order to be successful. The purpose of this interactive panel discussion is to foster learning of the multidisciplinary nature of SBTs in their application to critical health care settings. The panel features subject matter experts with a variety of professional backgrounds and experiences, including bioengineering, emergency medicine, as well as human factors. Through this discussion, we aim to further the field of assessment to enable practitioners to collect accurate, reliable, and meaningful findings from SBTs.

S79: Panel – Development and Implementation of HIT to Support NCDs in Low and Middle Income Countries

J. Dick, Indiana University School of Medicine/Regenstrief Institute; T. Cullen, Regenstrief Institute; P. Park, Harvard University School of Medicine, Partners in Health

A disproportionate amount of morbidity and mortality due to non-communicable diseases (NCDs) is suffered by individuals living in Low and Middle Income Countries (LMIC). As countries grapple with the development of health care systems to prevent, diagnosis and treat NCDs, many countries are also investing in HIT designed to support providers at the point of care as well as the creation of comprehensive data warehouses to evaluate the quantity and quality of care being delivered. This panel will provide a discussion on the current development and deployment of HIT solutions to support the provision of care for NCDs in large-scale care systems across LMICs with a focus on the role of data re-use in each care setting.

S80: Panel – Collaborative Science Within Academic Medical Centers: Opportunities and Challenges for Informatics

J. Starren, Northwestern University; W. Hersh, Oregon Health & Science University; C. Longhurst, University of California San Diego; P. Payne, Washington University in St. Louis

One of the challenges for Collaborative Science occurs within the walls of academic medical centers. Competing intra–institutional interests and financial disincentives may create barriers to collaboration. This problem is especially acute for Chairs, Division Chiefs, Institute Directors, Chief Health Informatics Officers (CHIOs), Chief Research Informatics Officers (CRIOs), Chief Information Officers (CIOs) and other leaders of academic and operational informatics units. This panel addresses both the opportunities and challenges for intra-institutional collaboration in biomedical informatics within academic medical centers. Each panelist will speak to the successes and failures of collaborating within our own institutions, with the goal of collectively learning from all of these experiences. Each panelist will spend about 10 minutes describing collaborative efforts in informatics at their own institutions, highlighting one success and one failure each. This will be followed by open discussion that will help all attendees who work in academic settings find ways to improve collaboration, especially that which benefits their programs.

S88: Panel – Methods and Tools to Enhance Rigor and Reproducibility of Biomedical Research

H. Kilicoglu, National Library of Medicine; A. Névéol, Universite Paris Saclay; T. Clark, Massachusetts General Hospital; H. Xu, The University of Texas Health Science Center at Houston; N. Smalheiser, University of Illinois at Chicago

Rigor and reproducibility of biomedical research has been the topic of much debate in recent years. Cases of replication failures, increasing number of retractions, and pervasiveness of questionable research practices lead to a lack of confidence in published findings, indicating that a portion of the biomedical research investment is wasted. Some ongoing efforts aim to address the issues in research conduct and dissemination by focusing mainly on standardization. Guidelines and principles pertaining to data, code, and publications have been proposed. The goal of this didactic panel is to engage the medical informatics community in a discussion about strategies to complement such efforts using informatics methods, tools, and resources. In the panel, first, we will provide a brief overview of standardization initiatives. Next, the panelists will present their informatics–based approaches toward improving rigor and reproducibility of biomedical research, focusing on such areas as information retrieval, natural language processing/text mining, and semantic modeling. Finally, with audience participation, we will discuss challenges facing informatics research aiming to address these problems and seek to identify some potentially fruitful research directions.

S89: Panel – Navigating the Exposome: Real-world Experiences Connecting Environment, Community, and Behavior to Health

S. Grannis, Indiana University/Regenstrief Institute/Indiana University School of Medicine; J. Vest, IU Fairbanks School of Public Health; C. Patel, Harvard Medical School; E. Holbrook, Colorado State University; B. Dixon, Indiana University/Regenstrief Institute

Much of disease risk – upwards of 50 percent – is attributed to an individual’s environment, from birth to death. Advances in technology and the proliferation of personal devices (smart phones and wearables) is enabling researchers and clinicians to address the Exposome, social determinants of health and other novel data sources. This panel will explore the experiences of those working to develop infrastructure and methods to leverage environmental, sensor, community, social, and behavioral data to inform health and health outcomes. Specifically, the panel will discuss items related to the incorporation of social determinants of health, the efforts to align with a standard common data model, implement new infrastructures, and new methods of Exposome analysis.

S90: Panel – Natural Language Processing at Scale – Perspectives from Five Healthcare Organizations

R. Crowley Jacobson, UPMC Enterprises; W. Chapman, University of Utah; H. Liu, Mayo Clinic; O. Patterson, VA Salt Lake City Healthcare System; D. Zisook, Southern California Permanente Medical Group

Healthcare organizations are increasingly adopting Natural Language Processing (NLP) technologies to support healthcare quality assessment, risk stratification, clinical document improvement, and other tasks. Many providers and payors are using a combination of homegrown, open-source and commercial software. This panel introduces efforts of five different healthcare organizations that have dedicated natural language processing programs and teams. These teams are engaged in developing, deploying, evaluating and/or commercializing NLP technologies. Panelists will share their experience in forming and sustaining an NLP group within a healthcare organization. They will outline the diverse business models supporting their efforts, as well as the range of informatics infrastructures needed to enable their work. Panelists will present a wide range of use cases for NLP within their healthcare organizations, and describe their key stakeholders. A discussion of challenges and future directions will offer a roadmap for other institutions that are considering NLP as a strategic technology investment.

S91: Panel – Implementing Electronic Health Records in Psychiatric Settings: Lessons From the Field

A. Busch, McLean Hospital/Harvard Medical School; J. Luo, University of California Riverside; T. Peters, Vanderbilt University Medical Center/Vanderbilt Psychiatric Hospital

Psychiatric disorders are prevalent and associated with high rates of medical comorbidity. New models of health care aiming to improve health outcomes require robust care coordination among all providers. However, electronic health record (EHR) adoption in psychiatric settings lags behind the rest of health care. Additionally, integrated health care systems face federal and state privacy laws that add additional challenges to EHR adoption in these programs. This interactive panel aims to foster greater education and discussion of unique challenges impeding EHR adoption in psychiatric settings. The panelists, psychiatrists in clinical informatics leadership, will discuss their experiences and challenges implementing EHRs in diverse psychiatric settings, and describe important areas of development in informatics needed to care for patients treated in psychiatric settings and integrated care models.

Wednesday, November 7

S99: Panel – Informatics Needs and Solutions to Support Safe Opioid Prescribing and Effective Pain Care

C. Harle, Indiana University, Regenstrief Institute; L. Militello, Applied Decision Science; S. Anders, Vanderbilt University; R. Hurley, Wake Forest University

The U.S. is in the midst of a public health crisis of opioid misuse, abuse, substance use disorder, and overdoses. Biomedical informatics tools have a significant role to play in combatting this problem. Indeed, in recent years, there has been a proliferation of policies, data, and technology aimed at stemming the inappropriate prescribing and use of opioids and improving delivery of care for substance use disorder, while ensuring effective pain care. However, challenges remain in integrating Prescription Drug Monitoring Program databases with electronic health records, connecting disparate data needed to support clinical decisions, and translating new policies and guidelines into clinical workflows and information systems. In this panel, four researchers from different disciplines will describe several novel efforts to use data and informatics tools to support safer and more effective pain care and opioid prescribing. The panelists will describe efforts that span multiple institutions, and primary and specialty care. The panelists will each discuss successes, challenges, and lessons learned. These presentations will be used to stimulate an extensive audience discussion about the informatics challenges and solutions being explored, implemented, and studied around the country to improve pain care and respond to the opioid crisis.

S100: Panel – Applications for Improving Patient Care that Utilize Customized Common Data Models and Data Visualization Techniques

E. Kolb, E. Crowgey, D. Eckrich, Nemours Alfred I. duPont Hospital for Children; S. Volchenboum, University of Chicago

Implementing a structured and well-defined data dictionary for a phenotype-specific common data model can improve patient care, increase efficiencies at the bedside, and drive precision medicine projects as a utility for integrating clinical phenotype data with research data. This panel will discuss successful national projects that have leveraged common data models at the bedside via data visualization and analysis, and cohort identification. The panelists will utilize their expert knowledge and team science projects and collaborations, such as the Children’s Oncology Group and the International Neuroblastoma Risk Group, and focus on the following learning objectives: 1) establishing a common data model and reference data management for disease-specific data collection and cohort discovery, 2) developing effective data visualization tools for improving clinical utility of data from electronic healthcare records, clinical trials, and other sources, and 3) defining appropriate team members and steps required to build a healthcare environment capable of incorporating standard–based structured data collection into the clinical workflow.

S101: Panel – Challenges and Opportunities in Studying Health IT Implementation: Understanding the Trees and the Forest

K. Unertl, N. Lorenzi, L. Novak, Vanderbilt University Medical Center; P. Sengstack, Vanderbilt University

Transitions between health information technology (IT) systems require attention not just to the technical aspects of implementation, but also to the people and process components. New technology can profoundly change workflow, roles, and organizational structures. Evidence–based strategies are urgently needed to identify best practices for implementation and for understanding the organizational transformation that goes hand-in-hand with large-scale technology implementation. These strategies are especially needed as informatics moves into an era of transitions between technology systems rather than from paper-based processes to technology. This panel seeks to engage the audience in expanding the discussion around best practices for de-implementation, re-implementation, and evaluation of health IT. The panelists represent a diverse set of perspectives on medical informatics research, including sociology, anthropology, engineering, and nursing. Beginning in November 2017, the panelists engaged in a longitudinal mixed-methods study of a large-scale transition from an in-house developed technology infrastructure to a vendor-based infrastructure. The lessons learned from this longitudinal case study can serve as a basis for future implementation and evaluation research efforts. The subsequent interactive discussion with the audience will build on panelist comments, providing cross–organizational insights on the state of the art of health IT implementation research.

S102: Panel – Making Electronic Health Records Safer: Practical Strategies for Evaluation and Improvement

A. McCoy, Tulane University School of Public Health and Tropical Medicine; D. Sittig, The University of Texas School of Biomedical Informatics at Houston; A. Wright, Brigham and Women's Hospital/Partners HealthCare/Harvard Medical School; F. Magrabi, Macquarie University

Despite extensive funding and research efforts to improve patient safety and reduce healthcare costs through use of electronic health records (EHRs), concerns about EHR safety remain. This didactic panel of experts in EHR safety will describe their efforts to evaluate and reduce HER-related safety risks, including: 1) assessing features, functions, and usage scenarios with the potential to cause harm, 2) measurement using existing electronic data to further identify and quantify safety hazards, 3) ongoing monitoring through new technologies to prevent occurrence of additional errors, and 4) investigation and mitigation of events to identify contributing factors and prevent recurrence of errors. Each panelist will provide examples that can be implemented in real world, clinical informatics practice to improve EHR safety.

S110: Panel – Technology, Informed Consent, and Genomic Research: Powering Scale and Adaptability

F. Abujarad, Yale School of Medicine ; M. Doerr, Sage Bionetworks; M. Meyer, Geisinger Health System; S. Kraft, University of Washington School of Medicine, Seattle Children’s Research Institute; T. Bright, Washington University in St. Louis

There is an unprecedented proliferation of large-scale research initiatives that store human biological material as well as vast troves of genomic and phenotypic data for current and future use. Simultaneously, there has been growing recognition of the need for diversity within genomic data sets. With these shifts in scale and focus, we need to consider novel approaches to informed consent that are both scalable and adaptable. This panel will discuss the use of technology to adapt and enhance the informed consent process. At the conclusion of this panel, audience members will be able to: (1) compare and contrast various technology-enabled models of informed consent to facilitate genomic research, (2) discuss the potential impact of these new approaches to the informed consent process on the conduct of research, and (3) describe operational and technical challenges of implementation and evaluation of these enhanced informed consent processes.

S111: Panel – From Evidence to Action: Enabling Opioid Pain Management Guidelines Through Patient-centered Clinical Decision Support

B. Blumenfeld, RTI International; B. Middleton, Apervita, Inc./Harvard TH Chan School of Public Health; S. Sebastian, MITRE Corporation; S. Carney, OCHIN; M. Michaels, Centers for Disease Control and Prevention

Applying the knowledge contained in evidence-based research findings and guidelines in clinical practice is key to improving care quality. A promising way to ensure that research findings inform clinical care is through clinical decision support (CDS). Research in CDS, defined as processes for enhancing health–related decisions and actions using evidence-based information, has shown that the use of CDS can reduce costs, improve quality, and reduce medical errors in clinical settings. However, the process of turning clinical guidelines into point of care clinical decision support is complex, involving a multi–step process to convert guidelines from their narrative form into semi-structured and semi-formal forms that are machine-interpretable. This panel will examine the barriers, facilitators, methods used, and lessons learned at each step in the CDS lifecycle, when transforming clinical guidelines for opioid use in pain management into patient-centered, point of care clinical decision support.

S112: Panel – Computable Longitudinal Patient Trajectories

J. Warner, Vanderbilt University; G. Savova, Harvard Medical School; N. Elhadad, Columbia University; L. Bastarache, Vanderbilt University; D. Gotz, UNC Chapel Hill

Computational analysis of chronic or subacute medical phenotypes requires nuanced representations capable of capturing key details of medical-care journeys lasting for months or years. Flattened definition of phenotypes with a small number of diagnostic codes, medications, or procedures lacks the temporal ordering necessary to understand the treatment, response, and progression of complex conditions that evolve over years or decades, such as cancer or diabetes. Constructing computable representations of multiple health episodes involving direct and indirect clinical encounters presents challenges in extraction, representation, and interpretation. Individual observations must be assembled from source records; ordered and abstracted to identify key concepts; represented in a form suitable for meaningful comparison and analysis; and presented to clinical researchers. Current research projects using differing techniques to address the challenges of developing computational longitudinal phenotypes will be presented, with the goal of identifying, comparing, and contrasting alternative approaches and understanding how they might best be applied to new problems.

S113: Panel – Electronic Referrals for Connected and Integrated Service Delivery for Vulnerable Older Adult Populations

C. Pickering, University of Texas Health Science Center San Antonio School of Nursing; J. Vest, Indiana University; B. Tanzman, Department of Vermont Health Access; F. Abujarad, Yale University; H. Kharrazi, Johns Hopkins University

Social and economic factors play an important role in health. While many communities have resources available to support social determinants of health, connecting those in need to available services can be challenging because of the fragmented nature of multi-sector service delivery. In this panel, four researchers from multi-disciplinary teams working with different high-risk populations will describe novel approaches that address the role of technology and informatics in the alignment and integration of services across delivery systems for medical care, public health, and social and community services. At the conclusion of this panel, audience members will be able to: (1) compare and contrast different models of “electronic referrals” for connecting and integrating service delivery across sectors; (2) discuss the potential impact of electronic referrals on service delivery and health outcomes; and, (3) describe operational and technical challenge of implementation and evaluation of integrated service delivery models.


AMIA 2018 Annual Symposium Sponsors

AMIA First Look

AMIA Working Group Supporters (NIWG)


Student Paper Prize