3:30 p.m. – 5:00 p.m.
R. Miller, W. Rubinstein, American Society of Clinical Oncology; A. Norden, COTA, Inc.; S. Sweeney, American Association for Cancer Research; W. Dalton, Oncology Research Information Exchange Network
This didactic panel will describe how real-world data, traditionally defined as those data derived from a source other than a traditional clinical trial, can generate real-world evidence and can be used to facilitate research and inform clinical care. Focusing on the field of oncology and the collection, aggregation, and transformation of cancer data predominantly but not exclusively contained within electronic health records, panelists representing four platforms currently delivering solutions – CancerLinQ®, COTA, GENIE, and ORIEN – will provide their perspectives on what it will take to improve cancer data interoperability and make meaningful progress in knowledge- and data-sharing to benefit patients. The panelists will discuss the need for alignment on data standards, challenges around curating and incorporating rich clinical data sourced from unstructured documents and free-text, and the particular data needs of precision oncology.
J. Osheroff, TMIT Consulting, LLC; C. Robbins, Kaiser Permanente; B. Alper, EBSCO Health; D. Little, Epic; E. Lomotan, AHRQ
Over 50 individuals from many stakeholder groups convened in 2018 to develop and begin executing an action plan to improve pain management, opioid use, and opioid use disorder treatment through more widespread and effective patient-centered clinical decision support (PCCDS). The opioid action plan (OAP) includes five consensus, highly desirable opioid-related future vision scenarios for PCCDS-enabled care. The scenarios reference 19 PCCDS interventions that are not widely or well implemented today but could be highly valuable. The OAP includes recommendations for various stakeholder groups to make the future vision and underlying interventions widely implemented. Most importantly, the OAP includes over two dozen examples of specific steps toward the future vision that organizations have taken as a result of contributing to the OAP. During this panel, the OAP lead author will outline the future scenarios and recommendations, and 4 individuals who helped develop the plan will describe OAP-driven actions they have taken and/or implications for their work with healthcare systems, CDS suppliers, electronic health record vendors, and government agencies. Half the session will be interactive discussion between panelists and audience members to surface opportunities to further accelerate PCCDS-enabled pain and opioid management within and across organizations.
Y. Kumah-Crystal, Vanderbilt University; L. Salmi, OpenNotes, Beth Israel Deaconess Medical Center; C. Sarabu, Stanford University; J. Smith, AMIA; J. Warner, Vanderbilt University
Electronic health records (EHRs) have become ubiquitous tools and represent the standard of care for 96% of hospitals and 86% of ambulatory physicians in the United States. With adoption of EHRs came the promise of improved efficiency, higher quality care, and lower costs. Unfortunately, some clinicians are now spending twice as much time on documentation as they spend seeing patients, and the documentation paradigm of problem-oriented medical records is contributing to this imbalance. It is time to consider new innovations. In this panel, we present several ways that we can overturn the current paradigm, with a focus on increased readability, shareability, and information channels other than raw text. Clinical documentation serves many masters and exists within a complex regulatory environment, but we will tackle the implications of change head-on. We anticipate that this panel will be of interest to innovators and disruptors across the care continuum.
T. Carton, Louisiana Public Health Institute; M. Zirkle, PCORI; E. Shenkman, University of Florida; A. Kho, Northwestern; A. Hernandez, Duke Clinical Research Institute
The Patient-Centered Outcomes Research Institute (PCORI) funded the National Patient-Centered Clinical Research Network, PCORnet. PCORnet leverages the power of electronic health data in a national distributed research network (DRN) to conduct patient-centered research. From late 2013 to present, PCORnet has built, refined and expanded its infrastructure to connect various stakeholders in productive research partnerships that leverage the power of large volumes of high-quality health data maintained by partner networks. This panel will focus on the unique attributes and lessons learned in the areas of governance, research capabilities, and sustainability during PCORnet’s evolution toward a cohesive self-governing Network.
This panel will describe PCORnet’s advancement across governance, research infrastructure and overall network sustainability.
J. Adler-Milstein, University of California, San Francisco; Y. Chen, Vanderbilt University Medical Center; M. Hribar, Oregon Health & Science University; J. Popovic, RTI International; J. Overhage, Cerner Corporation
EHR systems generate log data, which can be productively employed to improve the efficiency, quality, and safety of healthcare. There is a growing number of studies that use EHR log data but, as with any new set of data and associated methods, there is little consistency in approach and there are substantial inefficiencies from researchers replicating foundational work. Following a highly successful AMIA panel in 2018 that introduced log data and its potential uses, we convened 3 workgroups to develop measurement and methods guidance for working with log data. Our panel will present results from each workgroup and illustrate how the products could be used to support a range of research topics. The panel will also engage in a discussant-led conversation about the opportunities, limitations, and methodological challenges of common approaches, and where additional efforts may be needed. 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 new methodological guidance for how to use these data in a research setting, and understanding the remaining domains in which community-led work could improve the quality and efficiency of EHR log data research.
Y. Aphinyanaphongs, NYU Langone Health; J. Wilt, Oschner Health System; C. Chivers, Penn Medicine; M. Sendak, Duke Institute for Health Innovation
Significant barriers to the translation of machine learning based predictive models in clinical care continue to exist. The barriers are both technical and social and successful deployments require significant effort. This panel will provide targeted case studies of machine learning based predictive models through model evaluation, deployment, and adoption. The panel will demonstrate end to end examples of machine learning based predictive models and the issues and processes leading to their successful deployment. In some case studies, model evaluation results demonstrate improvements in quality of care and reductions in cost of care.
8:30 a.m. – 10:00 a.m.
U. Topaloglu, M. Kohn, M. Gurcan, B. Wells, A. Dharod, Wake Forest Baptist Medical Center
Biomedical Informatics (BMI) is the pillar of a functioning Learning Health System (LHS) that enables bringing translational clinical research findings into clinical practice and thereby, improving the health of patients and populations. It is a compelling promise; hence, medical centers invest toward establishing sustainable BMI programs in the form of a Center for BMI or an academic division. WFBMI attempts to foster learning from care experiences and develop and/or deliver better and more robust algorithmic models (e.g. Clinical decision support [CDS] systems, etc.) utilizing its integrated data ecosystem. In this panel, starting with a historical background, we will discuss challenges and opportunities of establishing a BMI program to support the operation and necessary changes to the institution.
S. Al-Showk, E. Lomotan, AHRQ; K. Miller, A. Hettinger, MedStar Health; J. Michel, Children’s Hospital of Philadelphia
Clinical decision support (CDS) is the process of bringing evidence into practice by providing the right information, to the right audience, using the right channel and format, at the right time (known as the “Five Rights” of CDS). Health care systems have traditionally developed CDS in siloes, which is a sign of redundancy and inefficiency of an overall ecosystem where CDS knowledge might be shared. MedStar, through contract with AHRQ, studied potential efficiencies gained by health care systems using shared, interoperable CDS available through CDS Connect, AHRQ’s web-based platform for sharing CDS artifacts. This panel describes how researchers worked with four health care systems to quantify differences in process, time, and resources between CDS developed entirely internally versus CDS developed with shared CDS artifacts. At the end, participants will be able to translate this knowledge into a business case for health care systems. Participants will be able to access CDS Connect and review its collection of shareable, interoperable clinical decision support. Participants will be able to compare and contrast processes for developing CDS in siloes vs. developing CDS using shared CDS artifacts. Participants will be able to identify measures that quantify potential efficiencies, including time and resources, towards developing CDS.
F. Eisenberg, ESAC, Inc.; C. Nanjo, University of Utah; J. Rubini, Mathematica Policy Research; J. Kunisch, Memorial-Hermann Healthcare System; K. Lesh, Battelle Memorial Institute
The panel will discuss how existing efforts to express electronic clinical quality measures (eCQMs) have been mapped to new, evolving standards to enable and encourage continuity with emerging health information technology (IT) efforts without disrupting existing efforts to evaluate clinical care in current healthcare settings. The work is based on efforts to harmonize standards for quality measurement and clinical decision support (CDS) using the Health Level Seven International (HL7) logical representation, Quality Improvement Clinical Knowledge (QUICK)data model, and HL7 Fast Healthcare Interoperability Resources (FHIR) Clinical Reasoning resources as the structural components and HL7 Clinical Quality Language (CQL) as the expression language.
A. Neinstein, University of California, San Francisco, UCSF Center for Digital Health Innovation; J. Adler-Milstein, University of California, San Francisco; W. Morris, Cleveland Clinic Main Campus; E. Sweeney Anthony, Office of the National Coordinator for Health Information Technology; A. Sethi, Citizen Corporation
Application programming interfaces (APIs) underpin the progression from data to knowledge to action in so many walks of life. In healthcare, APIs have recently received national attention as Medicare and Medicaid began requiring doctors and hospitals to provide their patients with digital access to their health data through patient-facing APIs. Now the Office of the National Coordinator for Health IT has released groundbreaking rules for standardized APIs as a cornerstone of national interoperability and patient access. As healthcare’s API era launches, our panel will present four key perspectives on patient-facing APIs—academic research, frontline health systems, app innovators and developers, and national policy—and strategies and priority use cases in the real world today. We will present findings of a just-completed national study of leading health systems and their uses of patient-facing APIs, including best practices, early challenges, and potential policy implications, followed by initial research on patients’ usage of those APIs, a deeper look at one leading health system’s strategy and experience, a leading app developer’s perspective, and ONC’s major policy proposal this year using APIs to advance patient access and interoperability. This panel provides an eagerly anticipated first look at real-world usage and impacts of patient-facing APIs.
J. Schneider, Harvard Catalyst | Clinical and Translational Science Center; K. Holmes, Northwestern University Feinberg School of Medicine; R. Revelo, Oregon Health Sciences Institute, Oregon Health & Science University; D. Keogh, tranSMART
This panel will follow the stories of three informatics tools or methods developed by CTSAs to look back on the successes and failures of ten years supporting clinical and translational science. The stories will highlight successes where they were perhaps not expected, lessons learned about tool development and outreach, how sustainability is supported, or why a great product never got widely adopted. The discussion will then turn to the current environment of open science, and how the changes in the research lifecycle could have impacted the success of tools that were ahead of their time, or made formerly widely used tools obsolete, and how future development needs to take into account open science best practices to maximize data and software reusability and funding requirements for data management. A discussion will elicit ideas and opinions from other stakeholders in the CTSA program, and how we can use the lessons of the past to ensure success in the next five-year cycle.
10:30 a.m. – 12:00 p.m.
L. Schilling, University of Colorado School of Medicine; R. Greenes, Arizona State University, Mayo Clinic; B. Alper, EBSCO; B. Rhodes, Dynamic Content Group, LLC; M. Michaels, Centers for Disease Control and Prevention
This panel brings together informatics experts in key domains necessary to leverage current and future technologies to efficiently translate evidence into action by making evidence and clinical practice guidelines computable and represented in a standard, interoperable format. There is a recognized lag of about 17 years in the translation of evidence to practice. The Internet, new data sources (such as ‘omics profiles and home sensor data), and new methods (such as causal inference and predictive analytics), and sources of evidence (such as ‘real-world evidence’) will lead to ever faster and broader evidence generation. Computable clinical guidelines, even when they exist, have been difficult to integrate into care processes due to a lack of pragmatism and patient-centeredness, among other factors. The panel will address challenges of guideline creation and adoption in the digital age, and the extent to which the availability of data and ways of acquiring and delivering knowledge provide new opportunities for integration and use of best-practice knowledge. The panel will discuss innovation in FHIR to transfer information regarding evidence (EBMonFHIR), clinical practice guidelines (CPGonFHIR), various types of computable knowledge and standards for biomedical knowledge (the Mobilizing Computable Biomedical Knowledge (MCBK) initiative), and redesigned processes for clinical practice guidelines.
M. Laing, Department of Health and Human Services; T. Cullen, Regenstrief Institute, Inc., Indiana University; J. Shivers, J. Amlung, Regenstrief Institute, Inc.; S. Fonseca, PRTNRS (&Partners) Product Design and Development Studio
The HHS IHS HIT Modernization Project is sponsored by the U.S. Department of Health and Human Services Office of the Chief Technology Officer (HHS OCTO) in conjunction with the Indian Health Service (IHS) to provide an analysis and recommendations to HHS concerning the modernization of the IHS Health Information Technology (HIT) system (RPMS). This project is anticipated to be the first step of a multi-faceted project. Human-Centered Design (HCD) is an integral methodology for the evaluation of the current state and is being used to evaluate the current state of HIT across the I/T/U (Indian Health Service/tribally operated/urban programs) and to identify alternatives.
This panel will demonstrate the value of a human-centered research approach early in the software development lifecycle and will provide participants with methods, tools and recommendations for applying HCD in other health informatics projects. They will share the experience of applying HCD methodology to help answer “What does modernization look like from the end-user perspective?” The methods described were applied during IHS hospital and clinic site visits and during analysis sessions. Lessons learned, including appropriate tools to use for gathering, synthesis and analysis, will be shared.
K. Kawamoto, University of Utah; S. Lane, Sutter Health; C. Caraballo, Audacious Inquiry ; T. O'Malley, Partners HealthCare
While substantial progress has been made in recent years to improve health care through standards-based interoperability, much remains to be done. The U.S. Health IT Advisory Committee (HITAC) was established in the 21st Century Cures Act of 2016 to provide recommendations to the National Coordinator for Health IT to advance the electronic access, exchange, and use of health information. In this session, the co-chairs of the HITAC U.S. Core Data for Interoperability (USCDI) and Interoperability Standards Priorities (ISP) Task Forces will describe the HITAC recommendations for advancing interoperability in the U.S. in these areas. These recommendations encompass the process for expanding the data elements included in the USCDI; the specific data elements with highest priority for near-term inclusion in the USCDI; the priority uses of health IT; gaps in standards or their implementation that hinder these priority uses; and recommended industry and government action to address these gaps. Following an overview of the HITAC recommendations in these areas, the panelists will engage in an interactive dialogue with the audience on what is most needed to advance health care in the U.S. through standards-based interoperability.
1:45 p.m. – 3:15 p.m.
S. Boca, Georgetown University; D. Hanauer, University of Michigan; I. Holmes, UC Berkeley; D. Kontos, University of Pennsylvania; J. Klemm, National Cancer Institute
This panel will introduce the Informatics Technology for Cancer Research (ITCR) program, an NCI initiative to support investigator-initiated informatics technology driven by critical needs in cancer research. Investigators supported by ITCR will share their perspective on the importance and challenges in developing and disseminating informatics technology stemming from an academic setting, using their own projects as case studies, and will discuss how programs like ITCR can benefit these efforts. Panelists will also describe how their tools are benefiting the scientific goals of the targeted cancer research communities and how interactions with end users have shaped the development of their tool. In addition, the panel will discuss ways in which the cancer informatics community of practice that is evolving through ITCR is promoting collaboration and interoperability within and beyond program participants. The NCI ITCR program coordinator will moderate the discussion and will provide a brief overview of NCI’s goals for supporting informatics technology development.
T. Cullen, Regenstrief Institute, Inc., Indiana University; M. Laing, Department of Health and Human Services; C. Crowder, Crown Consulting & Management Services; P. Biondich, Regenstrief Institute, Inc., Indiana University; H. Hays, Hays Health Informatics Consulting, LLC
The Indian Health Service (IHS) has enjoyed a 50-year history of successful, innovative and open source development of health information technology (IT). Due to multiple factors, the IHS is currently engaged in evaluating options for health IT system modernization. The Office of the Chief Technology Officer (OCTO) at the Department of Health and Human Services (HHS) collaborated with IHS and the Office of the National Coordinator (ONC) to design and conduct a multi-faceted, expert-driven and research-based approach to evaluate alternatives for health IT modernization. This panel will describe the drivers for IHS health IT modernization, the design and execution of the HHS/IHS Health IT Modernization Research Project, and methods used to ensure community engagement in modernization efforts. The panel will focus on sharing lessons learned and rubrics developed for HIT modernization applicable to other healthcare organizations.
V. Curcin, King's College London; E. Ford, Brighton and Sussex Medical School; J. Pathak, Cornell University; G. Nenadic, University of Manchester
Social networks are increasingly seen as a promising source of data for mental health research. The resulting applications are manifold, from improving diagnostics and monitoring patient activity, to developing digital interventions and targeted advertising of mental health services. This panel will deliver on some key learning objectives in the area: (1) Understand the ethics and governance frameworks; (2) Understand recommended technology for collecting data from popular social networks (Twitter, Facebook, Reddit); (3) Discuss approaches to complex concept definitions needed for working in the mental health domain; and (4) Learn about current state-of-the-art NLP and machine learning techniques for processing social network data, their capabilities, and what the future may hold for this field.
S. Bhavnani, UTMB; S. Visweswaran, University of Pittsburgh; E. Kummerfeld, University of Minnesota; C. Clark, R. Penton, University of Texas Medical Branch
The rapidly maturing fields of translational science and implementation science have recognized the multidimensional nature of health outcomes and assembled multidisciplinary scientific teams to address them. Furthermore, team science researchers have proposed several methods to improve the ability of such teams to integrate diverse disciplinary knowledge, with the goal of generating novel insights that transcend disciplinary boundaries. However, current informatics tools are mostly targeted to individual researchers working with a few dimensions of health outcomes such as genes or EHR data. This panel poses the question: Could the fields of translational science and implementation science benefit by a new class of informatics tools that are team-centered?
To address this question, this panel brings together researchers from human-computer interaction, machine learning, causal modeling, hospital administration, and nursing to explore two challenge questions: (1) is team-centered informatics old wine in new bottles? (2) does team-centered informatics provide low bang for the buck? The attendees will be asked if they agree or disagree with the conclusions of each panelist, with the goal of generating an animated discussion. The panel discussion should benefit researchers engaged in translational and implementation science, in addition to program managers from funding agencies aimed at supporting team science.
3:30 p.m. – 5:00 p.m.
S. Garcia, Department of Health and Human Services; R. Freimuth, Mayo Clinic; D. Jones, Utah Department of Health; R. Milius, National Marrow Donor Program
Genomic data sharing is critical to the Precision Medicine Initiative (PMI) and the next impetus of genetic research. Launched in 2016, Sync for Genes aims to standardize the sharing of genomic information between laboratories, providers, patients, and researchers. Two phases of this project have culminated, both resulting in the advancement of the Health Level 7 International (HL7®) Fast Healthcare Interoperability Resources (FHIR®) Clinical Genomic profile. During the course of this project, several important gaps were identified in the emerging standard that must be filled to set a firm foundation for the use of genomic data in the provision of tailored medicine and research. This panel will share results and findings from the Sync for Genes project, review the semantic data model work conducted to date, and discuss with the audience priorities for future work. The discussion will explore roadblocks that may be hindering widespread adoption of FHIR for genomics and weigh alternatives that may be in use to support the exchange of genomic data for use at the point of care and for research.
H. Singh, A. Meyer, T. Giardina, Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center, Baylor College of Medicine; D. Upadhyay, Geisinger; D. Sittig, UT Memorial Health Care System
Diagnostic errors pose a significant threat to patient safety. Learning health systems (LHS) provide opportunities to investigate, monitor, and measure diagnostic errors in real time within a healthcare system. A multidisciplinary panel will discuss emerging priorities and research opportunities related to health information technology (HIT)-enabled measurement of diagnostic errors, now considered the next big patient safety challenge. This session will explore how to use multiple sources of data from the health system (electronic triggers), providers, and patients, to identify and analyze missed opportunities in diagnosis. These data can be used to understand how health systems can learn and enhance the safety and accuracy of the diagnostic process. The learning objectives of this panel include: 1) how HIT-enabled health care organizations can become LHS for improving diagnosis; and 2) how informaticists can help implement pragmatic approaches to measure and understand diagnostic errors and engage in development of institutional infrastructure and feedback processes to promote individual and organizational learning.
K. Mandl, Boston Children's Hospital, Harvard Medical School; A. Gupta, Google Cloud; J. Mandel, Microsoft Healthcare, Boston Children's Hospital, Harvard Medical School; M. Obenhaus, Cerner Corporation; D. Rucker, Department of Health and Human Services; I. Vetter, Epic
Draft regulations following from the 21st Century Cures Act mandates the use of FHIR as a universal format for clinical data and the SMART API as a universal app launch standard. This creates a form of interoperability known as substitutability—an app can be added or deleted from the electronic medical record as easily as it can from a smartphone. It also enables the apps to be reusable—an app written once can run anywhere in the healthcare system. This panel brings together leaders representing public API development, the government agency spearheading health IT interoperability, the two largest EHR vendors, as well as relatively new entrants into the space—major cloud services companies. This interactive discussion will engage the audience in exploration of barriers to and opportunities for creating a burgeoning health apps economy, one that accelerates the translation of innovation to care, well established in provider workflows.
L. Clack, University of Georgia; Q. Zeng, Veterans Health Administration, George Washington University School of Medicine; G. Jackson, IBM Watson Health, Vanderbilt University Medical Center; W. Hersh, Oregon Health & Science University; A. Mohanty, University of Utah, Veterans Health Administration
A career in biomedical informatics typically requires a significant time and energy commitment, which must be balanced with other areas of life. An informatics professional can serve in multiple roles, such as domestic partner, dependent caregiver, etc. which may be challenging to their career trajectory and opportunities and may be hard to blend, though all of these roles can contribute to an enriching and fulfilling life. Biomedical informatics is especially challenging, given its interdisciplinary nature and wide breadth and depth. Key topics relevant to work-life balance include identifying values and priorities, creating a career matrix/timetable, family planning (e.g. family leave, dependent care), time management, burnout and managing daily struggles, work arrangements, dual-careers. In this interactive panel, a diverse group of informatics professionals will present different perspectives and discuss relevant issues, resources and recommendations related to balancing personal and family life while advancing their biomedical informatics careers. As these issues affect women and men, all informatics professionals who are interested in gaining insight about and advancing their informatics careers from various perspectives and engaging in a discussion about work-life balance issues are encouraged to attend. This interactive panel supports AMIA’s strategic directive of professional growth and leadership development of our members.
Y. Zhang, E. Sholle, M. Sharko, Y. Zhang, J. Ancker, Weill Cornell Medical College
It is well-established in the public health literature that health is strongly influenced by socioeconomic, cultural, and environmental conditions, community factors, and lifestyle. As healthcare organizations are becoming interested in these social determinants of health (SDH), informatics has multiple opportunities to shape and facilitate the collection, analysis, and application of SDH data to improve health equity. In this panel, we discuss multiple challenges and opportunities around three areas: how to collect SDH data; how to analyze these data; and how to apply these data to improving healthcare operations and health equity. Panelists will also analyze methodological reasons why, when the direct effects of SDH on health are so clear, modeling approaches using large clinical data sets sometimes fail to find effects of SDH. After participating in this session, attendees should be able to define SDH, identify multiple barriers to the efficient and accurate collection and analysis of these data, describe several informatics approaches to overcoming these barriers, and demonstrate how SDH influence the course of a chronic disease (heart failure).
8:30 a.m. – 10:00 a.m.
T. Leung, Maastricht University; M. Ball, IBM Research, Johns Hopkins School of Medicine; C. Brandt, VA Connecticut Healthcare System, Yale University School of Medicine; G. Jackson, IBM Watson Health, Vanderbilt University Medical Center; T. Schleyer, Indiana University School of Medicine, Regenstrief Institute, Inc.
Sponsorship is important for career development, but infrequently discussed and often not well understood. Sponsors are advocates for their protégés, recommending them for promotion, connecting them with senior leaders, promoting their visibility – pulling them up the ladder with them. In this panel, the speakers will introduce the concept of sponsorship, its impact on careers of young, mid- and experienced informaticians, and describe their own experiences from both sides of the sponsorship relationship. The objectives of this panel are to: (1) learn the importance of sponsorship in one’s career; (2) learn how to be a sponsor or to find one; (3) share experiences with others; and (4) develop an action plan for what to do during the AMIA symposium to put in to practice what is learned.
E. Poon, Duke University Health System, Duke University School of Medicine; C. Friedman, Michigan University School of Medicine; P. Payne, Washington University School of Medicine, Washington University McKelvey School of Engineering; M. Pencina, Duke Clinical Research Institute; K. Johnson, Vanderbilt University
The confluence of EHR adoption, advances in data science, and intensified pressures to achieve the quadruple aim has provided fertile ground for the development and spread of Learning Health Systems (LHS), which the Institute of Medicine has defined as systems in which ‘science, informatics, incentives, and culture are aligned for continuous improvement and innovation, with best practices seamlessly embedded in the delivery process and new knowledge captured as an integral by-product of the delivery experience.’ Informaticians will need to play a key role in these ecosystems, but many socio-technical barriers may stand in the pathway to success. This panel will showcase the experience of four academic medical centers that have been early developers of LHS. Panelists will describe strategies employed and investments made to overcome these barriers and sustain the success of LHS.
K. Chaney, T. Zayas-Cabán, Office of the National Coordinator for Health IT; K. Mandl, Boston Children's Hospital; K. Miller, MedStar Health
It is critical the field of health IT innovate far quicker to optimize real-time solutions in areas ripe for momentum. The 21st Century Cures Act requires health IT to have application programming interfaces (APIs) which enable access to all of a patient’s data with no special effort. The Office of the National Coordinator for Health Information Technology (ONC) published a proposed rule to implement these APIs, both for bulk data access and for one at a time patient access. In anticipation of a new generation of health IT, ONC created the Leading Edge Acceleration Projects (LEAP) in Health IT funding opportunity, informing the development, implementation and use, and/or advancement of well designed, interoperable, and scalable health IT. Two projects under LEAP in health IT, will be showcased on this panel. Dr. Mandl will discuss and demonstrate a provider-payor use case employing the emerging SMART/HL-7 bulk-data API, which defines a Fast Healthcare Interoperability Resources (FHIR®) “Flat FHIR” format. Dr. Miller will present on work advancing clinical knowledge at the point of care via risk calculator implemented as a SMART on FHIR app into an active surveillance tool clinicians can use in electronic health records (EHRs) to integrate and improve patient care.
10:30 a.m. – 12:00 p.m.
T. Walunas, A. Kho, Northwestern University; N. Kotecha, Parker Institute for Cancer Immunotherapy; N. Beaubier, Tempus
Immunotherapies represent the next generation of cancer therapeutics and have shown dramatic impact and durable responses on a wide variety of tumor types. However, these therapies only provider durable outcomes for 20-40% of patients and up to 90% of patients are expected to experience a so-called immune related adverse event. Large scale EHR data from multiple institutions, linked to genomic information as well as other environmental data could be used as an underlying architecture to drive the future of precision medicine in cancer immunotherapy research. Participants in this panel discussion will learn more about 1)current lessons learned and challenges faced by those using informatics approaches to understanding cancer immunotherapy outcomes, 2) the availability of data to support precision therapeutics in cancer, 3) the challenges with using EHR data as a source for phenotypic and treatment data linked to genetic data to identify markers for successful treatment as well as risk factors for adverse events.
K. Natarajan, Columbia University; R. Carroll, Vanderbilt University; T. Campion, Weill Cornell Medical College; J. Grand, MITRE Corporation; S. Visweswaran, University of Pittsburgh
The All of Us (AoU) Research Program has embarked on an ambitious effort to integrate data from a diverse number of sources to create a resource that will enable biomedical research in the next decade and beyond. One large data source is EHR data. We will present the experiences of collecting, submitting, and curating data from various AoU enrollment sites to the Data and Research Center (DRC). The topics covered in this panel will include: the goals of the curation pipeline within the DRC; the experience with participant matching at both the DRC and enrollment sites; the data mapping efforts to the Observational Medical Outcomes Partnership (OMOP) vocabulary; and the development of metadata to support data quality.
Q. Zeng, Y. Shao, George Washington University; C. Dao, Yale University; O. Intrator, University of Rochester Medical Center; J. Goulet, West Haven VA Medical Center
Deep neural network models (DNNs) are being explored in the clinical domain, following the recent success in other domains such as image recognition. For clinical adoption, outcome prediction models require explanation; however, due to the highly non-linear inner transformations, DNNs are viewed by many as a black box. In this panel, we will discuss the interpretability and statistical inferences of the prediction using DNN model, in the context of real clinical outcome prediction use cases. The panel will begin by providing a brief overview of the deep neural network and the latest research to interpret DNN results. We will then introduce two explanation methods. We will also present methods to construct confidence intervals and p-values for the impact scores and validate them through simulations and data application. Finally, we will discuss the challenges and future work to improve the understanding and acceptance of DNN models by researchers and clinicians.
N. Shimpi, Marshfield Clinic Research Institute; K. Williams, Regenstrief Institute, Inc.; A. Alekseyenko, Medical University of South Carolina
Advances in the understanding of oral and systemic connections and a paradigm shift in delivering patient-centered care has created a demand for incorporating principles of interprofessional education and healthcare delivery. Historically, dentistry and medicine have evolved as separate domains, and this has influenced education and healthcare delivery. Lack of interprofessional training during professional school critically influences practice patterns of dental and medical providers. Interprofessional health information exchange and integrated medical-dental electronic health records are critically vital to integrated care delivery. The lack has raised challenges as well as opportunities for establishing informatics infrastructure to support interdisciplinary care delivery in the context of oral-systemic health conditions. The purpose of this panel is to provide considerations for the establishment of informatics infrastructure and required training of the integrated care delivery team. Our goal is to achieve a roadmap for alignment of informatics and interprofessional efforts to maximize the broad positive impact on continuity of care.
1:45 p.m. – 3:15 p.m.
F. Meric-Bernstam, UT-MD Anderson Cancer Center; M. Rioth, University of Colorado; A. Halevy, 2bPrecise; P. Stetson, Memorial Sloan Kettering Cancer Center
Next-generation sequencing is increasingly being used for clinical decision-making in oncology. There is a growing number of genomic testing platforms, genomically-selected trials, and rapidly evolving literature in genomic medicine and molecular oncology. Thus, there is a tremendous need for decision support to help health care providers and clinician-investigators. Emerging needs include assistance in accessing, displaying, and interpreting genomic reports, appropriately determining actionability of genomic alterations and identifying approved and investigational therapeutic options. The learning objectives of this panel include determinants of actionability for genomic alterations, needs for decision support in annotations of functional implications of specific alterations and therapeutic implications, and resources and approaches for decision support for genomically-informed therapy.
A. Smith, National Committee for Quality Assurance; L. Anderson, The Joint Commission; R. Samples, ESAC, Inc.; V. Nguyen, Stratametrics, LLC
Challenges to some of the Centers for Medicare & Medicaid Services (CMS) quality incentive programs have cited provider burden and feasibility concerns related to electronic clinical quality measures (eCQMs.) The move to use Health Level Seven International’s (HL7’s) Clinical Quality Language (CQL) for 2019 reporting allowed greater expressivity and more precise, unambiguous logic statements. However, retrieving data expressed in CQL is challenged by mapping requirements to access data entered and shared as part of healthcare delivery. This fact also limits sharing expressions with clinical decision support (CDS) artifacts to enable real-time workflow enhancements to improve performance. Panelists will describe how HL7’s Fast Healthcare Interoperability Resources (FHIR) and emerging HL7 implementation guides can improve measure precision and expressivity and provide insights into expectations for more straightforward implementation. Panelists will use examples of eCQMs developed using these new standards and tested in FHIR Connectathons and in real world pilots.
R. Rudin, RAND Corporation; C. LeRouge, Florida International University; D. Lavallee, University of Washington; M. Reddy, Northwestern University; A. Dalal, Brigham and Women's Hospital
Patient-reported outcomes (PRO) are widely used in research studies, but thus far have limited use in routine clinical care. Despite the potential to make care more patient-centered, improve the accuracy of symptom recognition, facilitate shared decision making, and improve patient outcomes, health systems struggle to implement PROs in a sustainable way that generates value for patients and clinicians. Electronic health records (EHR), patient portals, and mHealth apps have the potential to collect electronic PROs (e-PROs) more seamlessly, but best practices regarding implementation have not been established. In this panel, four presenters will describe emerging lessons from experiences implementing e-PROs in the clinical setting. Topics include methods for PRO visualization; designing and implementing a practice models that collects e-PROs via smartphone app for patients; integrating e-PROs collected from apps into the EHR for clinicians; and governance considerations. Attendees will learn the current state of knowledge for effective e-PRO implementation as well as key challenges and future directions.
D. Corrigan, Royal College of Surgeons in Ireland; V. Curcin, King's College London; J. Ethier, Sherbrooke University; A. Flynn, University of Michigan; D. Sottara, Arizona State University
The popularity of health data science and persistent problems with avoidable medical errors combine to put a bright spotlight on the growing need for as much actionable health knowledge as possible to be expressed in standardized, computable forms that can be shared, reused, validated, traced as it evolves across time (provenance) and extended by other researchers. Recognizing this need, several major computable biomedical knowledge organizing initiatives are underway around the globe - Research Objects, MCBK, TRANSFoRm, CodeOcean, Knowledge Grid. This panel shall deliver on four key learning objectives. Attendees will come to: (1) Understand major (a) research, (b) clinical, (c) population health, and (d) educational use cases for actionable Computable Biomedical Knowledge (CBK); (2) Recognize common barriers to the widespread adoption and use of CBK to improve health; including the need to position CBK artifacts in relation to existing standards, such as FHIR, BPMN, DMN, PMML, PFA, RDF/OWL, etc. and be able to discuss strategies for achieving large-scale CBK artifact implementation in concrete scenarios, including the routine use of CBK created outside an organization with possibly different technologies; (3) Understand a series of key design considerations that come into play when creating CBKs. (4) Learn the challenges of managing large numbers of routinely evolving CBK artifacts in ways that span research, clinical, health operations, and educational environments and be aware of emerging software tooling used to address CBK management at scale; (5) With the idea that all knowledge is relational by nature in mind, envision the most critical relationships amongst CBK artifacts and be aware of how these relationships can be used to improve the findability, accessibility, interoperability, and reusability (FAIRness) of CBK artifacts, collections of CBK artifacts, and related CBK-enabled IT services.
3:30 p.m. – 5:00 p.m.
E. Crowgey, Nemours; M. Breitenstein, University of Pennsylvania; T. Walunas, Northwestern; S. Volchenboum, University of Chicago
Genomic testing has led to a paradigm shift in our understanding and treatment of disease, with particularly noteworthy accomplishments in oncology. Recent advances in consumer genomics services such as 23andMe have empowered the patient to sequence their genome independent of their healthcare provider. In parallel, personal wearables [e.g. Garmin, FitBit] have become inexpensive and ubiquitous, providing information for personal health maintenance and activity monitoring. While patients are increasingly empowered with new information about their health, genome, and healthcare risks, there are also growing concerns about how people will access these services, how they will interpret and use the data, and how data will be protected from nefarious uses. As these products and services expand, it is critical that informaticians be amongst the most informed in the healthcare ecosystem. The learning objectives for this discussion are: 1) understand risks and benefits to patients and providers for the different approaches for generating personalized genomics data, 2) learn about strategies for engaging patients and consumers in research and return of results in a patient-driven genomic era, and 3) learn data governance, privacy, and data standards required for these large-scale projects.
B. Kim, AHRQ; T. Payne, University of Washington Medicine; G. Melton, University of Minnesota, Fairview Health Systems; H. Singh, Baylor College of Medicine, Michael E. DeBakey VA Medical Center; C. Dymek, AHRQ
Electronic Health Record (EHR) use has been increasingly associated with clinician burnout. Up to 49% of a clinician’s time is spent working on EHR and other administrative activities. Given competing pressures of being clinically productive and trying to maximize face-to-face time with patients using imperfect and sub-optimized EHR systems, there is a tremendous need to fundamentally transform these systems and improve clinicians’ experiences. This panel will feature findings and implications from important health information technology (IT) research projects aimed to improve clinicians’ EHR experiences. Specifically, presentations and discussions will revolve around evidence-based solutions to improve three of the highest contributors to EHR burden: (1) documentation, (2) chart review, and (3) inbox tasks. The panel will conclude by moderating an in-depth discussion with the panel members and audience about health IT research priorities to further build the evidence-base for alleviating EHR-related clinician burden. The learning objectives for this panel include: increasing awareness of new health IT research aimed to reduce clinician EHR burden; understanding the potential for leveraging these health IT tools and applications into various healthcare settings; and understanding future health IT research directions aimed to improve the clinician EHR experience.
B. Rhodes, Dynamic Content Group; K. Kawamoto, University of Utah; N. Kashyap, Yale New Haven Health; C. Erdmann, Cerner
Commercial electronic health record (EHR) systems are beginning to provide support for standards-based clinical decision support (CDS) frameworks in their general-release software. Initial support is focusing on the HL7 Clinical Decision Support (CDS) Hooks standard, and some EHR vendors are beginning to support the HL7 Clinical Quality Language (CQL) standard as well. EHR vendors are also supporting the US Core Fast Healthcare Interoperability Resources (FHIR) profiles for data access. This panel will describe lessons learned from a Centers for Disease Control and Prevention (CDC) and Office of the National Coordinator for Health IT (ONC)-sponsored effort to pilot the use of opioid CDS knowledge resources using these EHR-supported CDS interoperability frameworks. Important lessons learned include the need for EHR support for additional CDS Hooks trigger events, in particular for medication ordering; the still early nature of direct CQL support in EHR vendor systems; the need to combine standards-based CDS with supplemental logic implemented using EHR platform-specific CDS rules engines; and the need for expansion and EHR vendor support for US Core FHIR profiles and interfaces, including for placing orders and for retrieving information on dispensed medications.
B. Kaplan, Yale University; E. Davidson, University of Hawaii at Manoa; G. Demiris, University of Pennsylvania; R. Schreiber, Geisinger Holy Spirit; A. Waldman, New York Law School
Privacy is protected both ethically and legally as a foundation for creating enough trust during clinical encounters for health care decisions to be based on honest discussion and accurate information exchange. Research, public health surveillance, advances in artificial intelligence (e.g., machine learning and predictive algorithms for health care), and the promise of personalized medicine all depend on accurate, complete data. However, health-related data generated and used outside of clinical settings is not protected through privacy regulation. Data aggregators and companies combine data from multiple sources for health and other purposes, while individual behavioral and social practices are being incorporated into medical records. Meanwhile, boundaries and distinctions are breaking down between different categories of protected health data, and between protected data and data collected via commercial apps and services. Privacy of health-related data requires rethinking in this rapidly changing landscape. Panelists will consider challenges in health data privacy, data governance, and privacy policies and practices, including privacy vs the value of data sharing, the adequacy of current legal and regulatory regimes, and how technological developments affect health data privacy.
S. Guntuku, University of Pennsylvania, Penn Medicine; L. Southwick, Penn Medicine; I. Vahia, McLean Hospital, Harvard Medical School; I. Barnett, University of Pennsylvania; S. Chancellor, Northwestern University
Digital data available from social media and passive sensing offer a new window into individual health, with the potential to uncover a nuanced understanding of an individual’s lived experiences, both expressed and unexpressed. Harvesting health insights from digital data is significant yet understudied with implications to advance efforts that capitalize on objective digital data and evidence to improve health forecasting and interventions. In this panel, we will explore how digital data can be used to characterize, forecast mental health conditions, and potentially use it in clinical settings. In the first talk, we discuss how digital data may impact the process of care, specifically the doctor-patient relationship. The second talk explores the potential of real-time detection of behavioral anomalies using passive sensing to forecast schizophrenia. In the third talk, we show how social media can be used to better understand the naturalistic manifestations and consequences of loneliness based on a cohort of Twitter users. The fourth talk examines emergent ethical and methods challenges of utilizing novel digital data sources in mental health research and care. Together, this panel demonstrates the potential for social media and wearable sensors to uncover insight into psychological health and well-being. This panel is designed to explain a new generation of innovative and targeted digital decision support tools, and launch large-scale trials exploring the acceptability, efficacy, and ethical as well as clinical implications of incorporating these tools into clinical care.
T. Schleyer, Indiana University School of Medicine, Regenstrief Institute, Inc.; F. Naeymi-Rad, B. Rapchak, Intelligent Medical Objects, Inc.
Biomedical informatics is at a crossroads with respect to the translation of innovation. The center of innovation and implementation in health IT has shifted from its historic origin in academia to industry. Applied researchers must become familiar with a new context, goals, processes and vocabulary in order to be able to translate their research ideas into practical products and services. Most researchers and business entrepreneurs share a common goal: to make a meaningful difference in people’s lives. But, to do so in the commercial space, academicians must keep several principles in mind, such as: addressing a real problem in a meaningful way; focusing on the customer; considering business development strategies up front; leveraging the core skills and competencies of academia and businesses in a complementary way; planning for scaling the innovation; respecting and protecting intellectual property; considering multiple pathways to commercialization; being realistic about valuation; leveraging a variety of sources for capital; and understanding their institution’s technology management and transfer processes. This panel, composed of successful academic and business entrepreneurs, will discuss a framework that can help attendees in going from research idea to initial public offering through a structured, easy-to-understand process.
C. Hsiao, AHRQ; S. Garcia, Office of the National Coordinator for Health Information Technology; D. Meeker, University of Southern California; J. Blumenthal, MedStar Health National Center for Human Factors in Healthcare; K. Marsolo, Duke University
The electronic collection of patient-reported outcomes (PROs) supports patient-centered outcomes research by offering a complementary perspective to clinician assessments, and may provide greater insights into health status, symptom burden, adherence, and quality of life. However, PRO data are not routinely available electronically for clinical care or research due to various challenges including the lack of standards and easy-to-use data collection applications. This session will present findings from a project jointly led by the Agency for Healthcare Research and Quality and the Office of the National Coordinator for Health Information Technology that tested the use of the Health Level 7 International (HL7®) Fast Healthcare Interoperability Resources (FHIR®) technical specifications to collect and integrate standardized PRO physical function data in ambulatory care settings. The panel will engage in a discussion regarding the advantages and disadvantages of the different implementation approaches and architectures that were used in the project’s pilot tests as well as opportunities for the scale and spread of these implementation models to other use cases. The learning objectives include: (1) understanding different technical approaches for PRO implementation using FHIR® in settings with different characteristics and resources, and (2) learning about the required capabilities and resources for standardized PRO implementation.
8:30 a.m. – 10:00 a.m.
G. Demiris, University of Pennsylvania; A. Turner, University of Washington; S. Iribarren, University of Washington; K. Sward, University of Utah
Citizen science refers to the collection and analysis of data by various population groups in real-world settings as part of a collaboration with scientists. Citizen science aims to respect and facilitate contributions from community members to scientific inquiry, build capacity, facilitate community-driven questions and promote dissemination. The advancement of wearable and passive monitoring technologies and direct-to-consumer genetic test kits and the continued diffusion of social media have led to an increase in citizen science projects in health care. Informatics tools introduce opportunities for large cohort research studies to understand disease and drug response and the role of environmental parameters. Vulnerable populations, namely communities with limited access to resources or at a higher risk to experience discrimination, are less likely to engage in citizen science. Informatics tools can address some of the challenges that vulnerable populations experience in engaging in citizen science. This panel will provide examples from ongoing research studies demonstrating the potential for informatics to target access barriers. We will discuss innovative approaches to participant recruitment, data collection, participant engagement and communication platforms. Case studies include participants from diverse groups (across the life-span, in urban and rural settings, within the US and abroad).
M. Kumar, University of North Carolina at Chapel Hill; X. Santas, J. Kariuki, A. Waugaman, U.S. Agency for International Development, Health Data Collaborative; P. Mechael, HealthEnabled
This didactic panel aims to share global and country-specific experiences in implementing national digital health and health information system (HIS) assessment tools to facilitate the transition from data to knowledge to action-namely the Global Digital Health Index, HIS Stages for Continuous Improvement, and the HIS Interoperability Maturity Model. The panel presentations and discussion are targeted toward policy makers, health managers and health informaticians engaged in design and use of digital health technologies to inform data-based decision making at the policy, program and service delivery levels. Discussion is expected to focus on emerging trends and findings from the assessments and what the global health and HIS community can do collectively to address them.
S. Johnson, New York University Langone Health; S. Khairat, University of North Carolina; S. Boren, University of Missouri; V. Mohan, Oregon Health & Science University; E. Manos, University of Kansas; W. Hersh, Oregon Health & Science University
The mission of the AMIA Academic Forum is to foster the development, growth and improvement of academic units that educate students and conduct research in the discipline of biomedical and health informatics. The Forum provides a vehicle for academic units to interact, consider, and act jointly on issues important to them. This didactic panel describes the new organizational structure of the AMIA Academic Forum based on “educational communities”. The learning objectives for the panel include: 1. Understand how educational communities enable Forum members to engage with their peers to share knowledge, collect data, and promote standards across their institutions. 2. Understand the goals and activities of five different communities for informatics education: baccalaureate, professional master's, postdoctoral clinical fellows, nursing, and biomedical informatics academic leaders. 3. Understand the benefits for creating and participating in new educational communities within the Forum.
A. Samarth, Clinovations GovHealth; K. Chaney, Office of the National Coordinator for Health IT; J. Mandel, Microsoft Healthcare, Harvard Medical School; S. SooHoo, Cedars-Sinai Health System
In 2015, the Office of the National Coordinator for Health Information Technology (ONC) and the National Institutes of Health (NIH), in partnership with Harvard Medical School, announced the launch of Sync for Science (S4S) a national collaboration to accelerate adoption of application programming interfaces (APIs) that provide individuals a simplified, scalable, and secure way to share their electronic health record (EHR) data. S4S provides software tools, implementation support, documentation, and a convening platform to foster consistent, standards-based patient API access across diverse EHR systems. Initially tested to support EHR data sharing with the All of Us Research Program under the Precision Medicine Initiative®, S4S is a significant step towards developing a health system that provides patients with greater control over their EHR data, including aggregation of data across multiple provider sites, and sharing it securely and efficiently with other services, providers and research programs via mobile applications (apps). This effort lends ONC, NIH, participating developers, researchers, providers, and participants insight regarding barriers to implementing and using S4S, which can inform other similar endeavors. While S4S is testing sharing of EHR data for precision medicine research, the standards powering it can be leveraged for other uses. This panel will discuss key lessons learned from the S4S pilot thus far, and opportunities for leveraging S4S and standards-based APIs to support health care and research.
10:30 a.m. – 12:00 p.m.
A. Boxwala, Elimu Informatics; J. Desroche, MEDITECH; B. Middleton, Apervita, Inc.; J. Richardson, RTI International; J. Scherer, Motive Medical Intelligence
A confluence of factors is changing patient-facing clinical decision support (CDS) from a possibility to a reality. Recent product advances, the growing adoption of wearable consumer technologies, and policy updates are but some of the many factors to open CDS beyond hospital and clinic EHRs to consumers wherever they go. How is patient-facing CDS to best support patients and how will any resulting data from patients inform clinicians’ decision-making or support learning health systems? This panel is comprised of both pioneers in the field as well as current implementers and will consider critical issues as to how the marketplace is addressing CDS for patients in light of rapidly changing interoperability standards, rapid shifts in the healthcare and consumer markets, and game-changing policies such as the 21st Century Cures Act and the Office of the National Coordinator’s Notice of Proposed Rulemaking. Through field experiences and clinical scenarios, the audience will come to understand the challenges and opportunities for patient-facing CDS and learn new approaches to developing and implementing such as the AHRQ-funded Patient-Centered CDS Learning Network’s efforts to pilot a patient-facing CDS demonstration application and candidate reference architecture.
J. Tenenbaum, Duke; R. Kavuluru, University of Kentucky; T. McCoy, Massachusetts General Hospital; O. Uzuner, George Mason University; S. Velupillai, King's College
It has been shown that application of natural language processing (NLP) and text mining of clinical notes in electronic health records (EHRs) can help facilitate deeper and more accurate phenotyping in many disease areas. These methods are particularly useful in the context of mental illness, where disorders are highly heterogeneous and lacking diagnostic lab tests, medications are non-specific, and diagnosis codes are used inconsistently by different care providers, if not left out altogether. However, NLP approaches in mental health EHRs also pose a number of challenges, some common across all of medicine, some unique to mental illness.
The assembled panelists represent a geographically and scientifically diverse set of perspectives in the field of clinical NLP for mental health. They will discuss how their respective groups have addressed both scientific and infrastructural challenges and how they have helped to advance the field. They will also describe lessons learned as well as tools and artifacts they have developed that will be of use to other groups who are interested in similar questions.
V. Tiase, New York-Presbyterian Hospital, University of Utah, College of Nursing; R. Austin, C. Martin, University of Minnesota; R. Masterson Creber, Cornell University; S. Kitsiou, University of Illinois at Chicago
With the proliferation of mobile health (mHealth) technologies and applications (apps), patients are generating more data than ever before. These electronic patient-generated health data (PGHD) have the potential to yield additional insights, decrease nursing documentation burden, and improve patient care. At the same time, PGHD are also raising questions and posing significant challenges for researchers and clinicians. This interactive panel of nursing and consumer health informatics leaders will discuss key findings and challenges from several studies aimed at engaging patients with mHealth applications with the intent of collecting PGHD for research and clinical use, suggest a road map for additional research, and discuss future directions for nursing and informatics involvement in the design and implementation of mHealth apps.
E. Harper, A. Garcia, The University of Kansas School of Nursing; J. Welton, University of Colorado; L. Moon, University of Minnesota
Examining variation in clinical practice is an important step to measuring efficiency and effectiveness in care delivery. The study of practice variation by an individual nurse and patient characteristics can indicate potential opportunities to reduce costs and improve the value of health care delivery. This didactic panel will describe the overall approach taken to examine the clinical practice of nurses linked to patients at a single pediatric hospital. This exemplary of emerging data science uses data from multiple electronic data sources, including clinical, administrative, human resources, staffing, and financial systems to identify outcomes with a new level of precision. Examples include nurse-sensitive outcomes throughout each shift, length of stay, and traditional discharge measures. Administrative outcomes include calculations of the direct cost of nursing care by nursing unit and diagnosis. Attendees will become familiar with the Nursing Value Data Model and methodology for working with big data. They will examine variation in nursing clinical practice and discuss their findings, including how it can be better supported for clinical and translational research applications.
T. Kuo, University of California San Diego; A. Das, IBM Research; K. Augustine, Deloitte; P. Zhang, Vanderbilt University Medical Center; L. Ohno-Machado, University of California San Diego, VA San Diego Healthcare System
In this 90-minute didactic panel, we will discuss the latest biomedical research and healthcare blockchain applications and implementations. Blockchain is a novel and emerging technology. However, its use in biomedical informatics is still emerging. The target audience includes medical informatics researchers, clinicians, health IT professionals, and leaders of healthcare organizations. The moderator and panelists, who are experts from academia and industry, will provide insights and representative uses cases. We will discuss the utilization of blockchain to advance healthcare informatics, the use of blockchain in clinical research, the combination of blockchain and HL7’s Fast Healthcare Interoperability Resources (FHIR) to share clinical data, as well as the integration of blockchain with privacy-preserving predictive modeling. Our panel provides a unique and timely opportunity to see experts discussing myths, advantages, and potential concerns of adopting blockchain technology in healthcare and biomedical research.
D. Lyalin, CDC; J. Daniel, HHS; D. Chrysler, Network for Public Health Law; C. Ishikawa, Kahuina Consulting, LLC
The goal of IZ Gateway project is to simplify the process for a multistate provider organization to submit immunization information to the appropriate immunization information system (IIS) without requiring the provider organization to create individual point-to-point connections with each IIS in the regions they operate in. This will reduce the cost for multistate provider organizations to send immunizations to the appropriate IISs and will improve the completeness of information in IIS.
The project team conducted an environmental scan to inform the development of clear technical requirements necessary to support simplified provider reporting to immunization registries across jurisdictional boundaries (without the need to make multiple connections). Organizations such as large multistate public and private provider organizations, hospital groups, adult focused care groups, large and independent pharmacy chains, and other entities involved in sending and receiving immunization data across jurisdictional boundaries were interviewed to inform this project.
N. Elhadad, I. Urteaga, Columbia University; A. Callahan, Stanford University; J. Reps, P. Ryan, Janssen Research and Development
Recent work in machine learning and modeling from healthcare data has shown much promise in leveraging large datasets to infer signal and predict outcomes of interest. With increasing numbers of such data sources available via collaborative efforts such as those created by the Observational Health Data Science and Informatics (OHDSI) network, it is now possible conduct clinical studies at an unprecedented scale. In this didactic panel, we will focus on another contribution of such data networks beyond the increased power of larger sample size, namely learning across multiple data sources as a way to improve the robustness of the learned models and increase the reliability of derived evidence. Our panel participants are junior and mid-career researchers in the field of machine learning for healthcare. We will (1) review the OHDSI network as one example of such a data network, (2) describe use cases in predictive analytics and an on-demand informatics consult service to distill clinical evidence from such data networks, where increased reliability is achieved through learning across different types of observational data, (3) describe current efforts in facilitating external model validation in OHDSI, and (4) illustrate current efforts in machine learning and specifically transfer learning to increase model robustness.