R. Vanderpool, A. Oh, R. Jensen, National Cancer Institute; U. Sarkar, University of California San Francisco; T. Hernandez-Boussard, Stanford University School of Medicine
This panel will provide informatics investigators with information and advice on developing a competitive grant application for submission to the National Cancer Institute (NCI). The panel consists of three NCI program staff who oversee research programs in cancer communication, implementation science, and healthcare delivery, and two NCI-funded investigators whose work focuses on informatics research in the cancer prevention and control context. NCI staff will describe current informatics-relevant funding opportunities, outline key steps in the grant submission and review process, provide tips on writing a strong application, and highlight examples of NCI-funded informatics research. Two NCI grantees will then engage in a moderated discussion regarding their personal experience navigating the NCI funding process. Attendees will have ample time to ask questions and should leave the session with an understanding of the NCI grant funding process, insight into funding priorities in informatics, and a set of recommendations for writing strong grant proposals.
S. Collins Rossetti, A. Moy, Columbia University; M. Kang, Brigham and Women's Hospital; J. Schwartz, K. Cato, Columbia University
Recent national recommendations, calls to action, and federal strategic plans highlight that reducing EHR burden is a priority of government agencies, and aligned with the Quadruple Aim. Overall, there are a lack of measures and established approaches to systematically describe and quantify the problem or to support efforts to reduce documentation burden. This panel will present the state of the literature and a suite of potential approaches for leveraging mixed-methodologies to reduce documentation burden, including EHR data-driven analyses and engagement with end-users. In addition to open discussion, the questions to stimulate discussion among attendees will relate to feasibility and scalability of measurement approaches, measurement limitations, and implications to inform interventions aimed at reducing documentation burden.
S. Garcia, Office of the National Coordinator for Health Information Technology; T. Kannampallil, Washington University School of Medicine; J. Hellewell, Intermountain Healthcare; V. Nguyen, Stratametrics; C. Thompson, Clinovations GovHealth
Synthetic health data can reflect the characteristics of a population of interest and be a useful resource for researchers, health information technology (health IT) developers, and informaticists. Researchers and developers often depend on anonymized data to test theories, data models, algorithms, or prototype innovations but may be required to aggregate, de-identify, or analyze data before it can be used. Additionally, high quality health data can be difficult to access because of cost, patient privacy concerns, or legal restrictions. The risk of re-identification of anonymized data is high and impossible to eliminate, especially for rare medical conditions. Interoperability issues impede gathering data from different resources for robustly testing analysis models, algorithms, or developing software applications. Synthetic health data helps address these issues and speeds the initiation, refinement, and testing of innovative health and research approaches. Capitalizing on this opportunity, the Office of the National Coordinator for Health Information Technology (ONC) is leading an effort to enhance an open-source synthetic data engine to accelerate research. This panel will discuss ONC efforts to support the generation of synthetic health data for research. Additionally, panelists will share their experiences using various types of synthetic health data along with associated benefits and limitations.
E. Gallego, EMI Advisors; W. Suarez, Kaiser Permanente, HL7 International; A. Taylor, Office of the National Coordinator for Health Information Technology; J. Lugo, Health and Human Services Administration for Community Living
The influence of social determinants on health outcomes is increasingly recognized in emerging payment reform programs, federal and state-based policies, and information technology initiatives. The growing awareness of how Social Determinants of Health (SDOH) shapes health has contributed to efforts to address actionable socioeconomic risk factors through the health care delivery system. However, the ability to document and address these social risk factors in clinical settings is hampered by the lack of standards available to code and exchange the data. This panel will introduce the Gravity Project, initiated in May 2019 by the Social Interventions Research and Evaluation Network (SIREN). The presentation will highlight the project’s approach to consensus-driven development of coded data elements needed for interoperability. Learning objectives included understanding the value proposition for standardizing SDOH data to support documentation activities within an EHR, interoperable electronic data exchange, and aggregation across clinical, community-based, research, and population-health systems.
S. Madhavan, Georgetown University; J. Guinney, SAGE Bionetworks; M. Haendel, Oregon Health & Science University; A. Wagner, Washington University; R. Thangudu, ESAC Inc.
Cancer research is on the precipice of transformational advancements in understanding tumor biology, disease etiology, risk stratification, and pathways to novel treatments. Yet this field has been stymied by siloed efforts to meaningfully collect, interpret and aggregate disparate data types from multiple high-throughput modalities to support clinical care. While individual data sources such as Proteomic Data Commons (PDC) and Human Tumor Atlas Network (HTAN) are organizing large-scale, patient-derived experimental and associated clinical data, the Cancer Research Data Commons (CRDC) is developing components such as the Center for Cancer Data Harmonization (CCDH) and Cancer Data Aggregator to harmonize datasets from individual nodes to bring analytical tools and computational power closer to the data. Once harmonized, these datasets have to be informed through interpretation efforts such as through ClinGen and the GA4GH. This timely panel will discuss efforts to harmonize novel data sets to drive cancer precision medicine through global data sharing.
A. McCoy, Vanderbilt University Medical Center; S. Dutta, D. Rubins, Harvard Medical School, Brigham and Women’s Hospital; M. Tobias, Phrase Health; A. Wright, Vanderbilt University Medical Center
With the near universal adoption of commercially developed electronic health records, local clinical decision support (CDS) implementations are now ubiquitous, and evolving standards provide a glimpse into an increasingly interoperable state for CDS. Despite the increasing ease to incorporate CDS like alerts and order sets into workflows, many informatics teams have trouble ensuring their build is useful due to the investment of effort and resources required. Novel, data-driven approaches are necessary for effectively evaluating and optimizing CDS to improve desired outcomes. This didactic panel of experts in CDS will describe diverse approaches to evaluating CDS across multiple healthcare organizations, including comprehensive measurement, proactive monitoring, capture of provider feedback, and analytics and governance across institutions.
M. Reading Turchioe, N. Benda, Weill Cornell Medicine; L. Grossman Liu, Columbia University; F. Wang, Weill Cornell Medicine; K. Miller, MedStar Health
Although the predictive accuracy of machine learning models in healthcare has dramatically increased, lack of interpretability continues to hinder their usefulness in clinical decision-making. User-centered design aligns a tool with an end user’s preferences, needs, and knowledge. Therefore, conducting user-centered design when developing tools for presenting predictive models may improve interpretability, clinician trust, and clinician’s ability to explain the model to patients. This panel will discuss several projects in which user-centered design is being used to create and implement tools for presenting predictive models to clinicians across a range of settings and clinical scenarios. Tools include visualizations, dashboards, and specific mechanisms of the models themselves. The panel will discuss the benefits and drawbacks of different tools and user-centered design methods used in this research. The audience will engage in discussions about the unique new challenges of current machine learning-based predictive models, using traditional risk calculators as a counterpoint for discussion.
S. Meystre, Medical University of South Carolina; J. Silverstein, University of Pittsburgh School of Medicine; G. Savova, Boston Children's Hospital; V. Petkov, National Cancer Institute; B. Malin, Vanderbilt University Medical Center
Large quantities of patient clinical data are becoming available in an electronic format, generated by the fast- growing adoption of electronic health record (EHR) systems in the U.S. This growth creates tremendous potential but also a growing concern for patient confidentiality. Secondary use of this 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 protect patient data confidentiality. A substantial amount of clinical data in the EHR are represented as narrative text and de-identification of such data 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 or even improved protection. However, despite these advances, automatic de- identification of clinical text is not commonly used and accepted. This panel will focus on automatic de- identification of EHR text with perspectives from various stakeholders, reporting on a workshop organized by the National Cancer Institute on February 25-26, 2020, and supported by the Cancer Moonshot Initiative. Discussions will aim at broad sharing of opinions, ideas, advice, and practical experiences with clinical text de-identification.
D. Tao, Saint Louis University; D. Wei, Stockton University; M. Sordo, 3MGH Institute of Health Professions, Harvard Medical School
Work-life balance issues have been reported as significantly impacting career development, especially for female professionals in health informatics. The proposed panel seeks to provide multiple perspectives on work-life balance issues and suggest potential solutions specifically targeted at early- and mid-career professionals in health informatics. Different working environments, multiple roles, different career stages, in addition to gender and ethnicity might intersect with these issues. The panel presentations and discussions will be helpful to identify work-life balance needs and provide valuable guidance for administrative plans to support women in health informatics. After participating in this session, the learners should be better able to:
- - Understand the current work-life balance issues faced in the health informatics community, especially for women.
- - Learn supporting strategies and measures that could be applied to individual’s workplaces to engage organizational leaders and create a supportive working environment.
E. Orenstein, Emory University, Children's Healthcare of Atlanta; J. Chaparro, Nationwide Children's Hospital; E. Webber, Indiana University; N. Muthu, Children's Hospital of Philadelphia
Influenza vaccine among children presenting to acute care is poor despite its effectiveness at preventing influenza-related illness, hospitalization, and death. Acute care visits are an important opportunity to improve vaccine coverage. Clinical decision support (CDS) has shown promise to increase influenza vaccine uptake. However, promoting health maintenance interventions such as vaccines in acute care settings requires navigating unique sociotechnical constraints to be effective. In this panel, we will review three institutions’ approaches to CDS for influenza vaccine uptake in emergency departments, urgent cares, and inpatient settings. We will illustrate common strategies, describe the reasons behind local customizations, and identify generalizable lessons that can be applied to other health maintenance interventions in acute care settings.
C. Hsiao, Agency for Healthcare Research and Quality; D. Marinac-Dabic, U.S. Food and Drug Administration; S. Kim, Centers for Disease Control and Prevention; A. Dobes, Crohn's & Colitis Foundation; S. Lumsden, U.S. Department of Health and Human Services
Under the Office of the Secretary Patient-Centered Outcomes Research Trust Fund (OS-PCORTF) portfolio, the Assistant Secretary for Planning and Evaluation (ASPE) supports and coordinates collaborative projects intended to build data capacity and infrastructure for patient-centered outcomes research (PCOR). This panel features three OS-PCORTF projects that have developed innovative technology to optimize researchers’ access to patient data from various sources including registries, electronic health records, and patient-reported outcomes. An extramural PCOR researcher will highlight work that has been built off an OS-PCORTF project’s technology. Innovative tools and technologies produced by these projects include a surveillance network to capture timely data for maternal, infant, and child health outcomes; mobile applications to capture patient-provided information; and a registry network for women’s health technologies. These projects demonstrate the value in creating, testing, and disseminating emerging technologies to make data more accessible for clinical decision-making and research.
P. Hoffman, Yale Child Study Center; E. Webber, Indiana University Health, Regenstrief Institute, Inc., Indiana University School of Medicine; J. Scott, University of Washington; J. Hron, Boston Children’s Hospital; T. Cullen, Regenstrief Institute, Inc., Indiana University School of Medicine
COVID-19 is a highly contagious respiratory pandemic that started in China and migrated to the U.S. in 2020. During the outbreak, telehealth services have gained prominence as mechanisms for delivering patient care in an effort to minimize community spread of illness and potential exposure in health care environments. The panel will describe the various settings where telemedicine has been implemented and lead a discussion of the progress as well as the challenges and lessons learned from these implementations. The benefits and limitations of the expansion of telemedicine will be presented.
N. Shimpi, K. Williams, A. Alekseyenko, Medical University of South Carolina
Opioid epidemic is a major public health concern in the United States. The Centers for Disease control estimates that nationally, 115 deaths occur each day from opioid overdose. The purpose of this panel is to provide an overview of implementation strategies and current initiatives that are undertaken for the establishment of informatics infrastructure for opioid management. Our goal is to promote informatics approaches to achieve best practices for opioid management facilitate reduction of adverse outcomes currently associated with opioid utilization.
T. Schleyer, Regenstrief Institute, Inc.; M. Hightower, University of Utah; C. Harle, University of Florida Health; A. Landman, Brigham Health; R. Rudin, Rand Corporation
The advent of Fast Healthcare Interoperability Resources (FHIR) and its inclusion in regulations mandated by the 21 st Century Cures Act have raised expectations that electronic health records (EHRs) will become platforms for a myriad of clinically useful apps. Similar to the Apple App Store or Google Play Store, such platforms are expected to support app ecologies that allow healthcare organizations to meet their health IT needs in a much more customized and granular fashion than previously possible. Although the FHIR standard is likely to reduce barriers to the technical integration of third-party apps with EHRs, challenges to achieving this vision remain. Healthcare delivery organizations will need to manage relationships with app developers, integrate the apps into workflows, address security risks, monitor that apps work as intended, and evaluate investments in apps in light of strategic priorities. Healthcare delivery systems are beginning to encounter these challenges but have little, if any, guidance on how to master them. This panel, composed of a biomedical informatics researcher, Chief Medical Information Officer, Chief Research Information Officer, Chief Information Officer, and an information scientist will offer insights into how the app ecology is expected to evolve, whether and how it will meet expectations of accelerated innovation, and how its challenges can be met. Attendees will hear about how different types of healthcare delivery systems are approaching this new world of EHR-integrated apps, and what can be learned from the experience of other industries as they have evolved to leverage application programming interfaces.
A. Samarth, Clinovations GovHealth; K. Chaney, Office of the National Coordinator for Health Information Technology; D. Levin, Datica; A. White, Cerner; R. Sahu, 1upHealth
The emerging health application programming interface (API) ecosystem brings opportunities to enable electronic health data sharing for care and research. However, it is important to ascertain how the current infrastructure can deliver on this opportunity and what barriers could impede uptake. Accordingly, the Office of the National Coordinator for Health Information Technology (ONC) led collaborative efforts to understand the availability and use of Health Level Seven (HL7 ®) Fast Health Interoperability Resource (FHIR ®) APIs and third-party mobile applications (apps) for sharing of electronic health data. APIs leveraging FHIR ® resources are available to extend capabilities of electronic health records, meet requirements of ONC’s and the Centers for Medicare & Medicaid Services final regulation, and meet the goals of the 21st Century Cures Act. This panel will discuss real-world experiences leveraging APIs and third-party apps to exchange electronic health data for use by patients and providers, including participants in the Sync for Science pilot conducted with the All of Us Research Program. Panelists will bring API implementation perspectives from government, providers, researchers, and health IT and app developers; health system preferences and policies; and EHR vendor and app developer requirements that extend beyond implementation of standards for interoperable data exchange.
S. Labkoff, Multiple Myeloma Research Foundation; L. Rozenblit, Prometheus Research; A. Elbert, Cystic Fibrosis Foundation; R. Belenkaya, Memorial Sloan Kettering
The cornerstone of personalized medicine approaches to rare diseases study is aggregating enough data to study. For diseases with limited populations, identifying sufficient patients at scale is just one major challenge. Given the advancements in personalized medicine, the need for well-curated data sets is critical. Though there is wide adoption and use of EHRs, challenges remain in aggregating enough of the right, aligned data to create study cohorts. There remain challenges from many perspectives, including standards implementations, legal constraints due to antiquated laws, and a lack of common data models. This panel will discuss many of the obstacles and solutions found in creating registries from an informatics, regulatory, legal and standards perspective. Each speaker will provide their perspectives on aggregating data sets (EHR, genomics, immunologic, patient-reported outcomes, and claims), how the health information technology infrastructure created for the EHR world helps or hinders these efforts, and the ramifications from the use of standard data models and terminologies. This panel will discuss the challenges of creating registries for rare diseases to enable personalized medicine clinical research in the context of a post-adoption EHR world. Each speaker has real-world experience in building or managing major aspects of such registries.
E. Pfaff, University of North Carolina at Chapel Hill; D. Gabriel, D. Jiao, S. Hong, C. Chute, Johns Hopkins University
The Next Generation Data Harmonization Core is a component of the National Center for Advancing Translational Sciences (NCATS) National Center for Data to Health (CD2H). This collaborative, national multidisciplinary group focuses on advancing methods and tools that support utilization of data with sources that span the spectrum of translational research domain utilized by Clinical and Translational Science Awardees (CTSAs). Many CTSA informatics organizations face resource bottlenecks associated with the requirement to support data transformation between multiple research databases and common data models (CDMs). The panel consists of representatives from the CD2H Next Generation Health Open Terminology (HOT) Ecosystem working group whose purpose is developing solutions to this problem. Results of data harmonization, transformation innovation, tool development and other supporting activities utilizing Health Level Seven’s Fast Healthcare Interoperability Resources (HL7 FHIR) application programming interfaces (APIs) are discussed.
J. Adler-Milstein, University of California San Francisco; J. Chen, Stanford University; M. Wang, University of California San Francisco; M. Hribar, A. Rule, Oregon Health & Science University
EHR systems generate log data, which can be productively employed to improve the efficiency, quality, and safety of care processes and outcomes. There is growing consensus on the value of EHR log data and the need for consistent approaches to working with them, including methods and measures. Following highly successful AMIA panels – the first in 2018 that introduced audit log data and its potential uses, and the second in 2019 that described approaches to achieve consistent methods and measures, our panel will present an initial cohort of leading-edge, large-scale projects using audit log data to tackle timely health system performance challenges. Projects were selected to illustrate how audit log data can support a range of research topics and methods. The panel will also feature a discussant-led conversation about the opportunities, limitations, and methodological challenges emerging from real-world applications 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 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.
G. Hripcsak, Columbia University, NewYork-Presbyterian Hospital; D. Blei, E. Bareinboim, Columbia University; M. Schuemie, Observational Health Data Sciences and Informatics, University of California Los Angeles, Janssen Research and Development; L. Zhang, Columbia University
Observational healthcare data, such as administrative claims and electronic health records (EHRs), are promising data sources for determining treatment effects and discovering causal relationships between treatments and clinical features. However, systematic errors (e.g. confounding and selection bias) inherent in the data generation process can introduce unwanted correlations and lead to invalid causal estimates if not handled properly. Besides the challenges residing in the data, the lack of reproducibility among studies further raises the concerns over the validity of results from observational studies. Our panel of informatics leaders, statisticians and computer scientists, including several who are world-famous for their work in casual inference, will jointly discuss how the rigor of observational studies can be improved through recent advances in method development across scientific fields. We will 1) introduce methods for handling confounding and selection bias on large-scale health data, 2) demonstrate how to evaluate the performance of causal methods across the Observational Health Data Sciences and Informatics (OHDSI) network, 3) introduce pioneering works in causal inference for handling unobserved confounding and discovering causal structure from observational data.
H. Liu, Mayo Clinic; X. Jiang, The University of Texas Health Science Center at Houston; S. Pakhomov, University of Minnesota; C. Weng, Columbia University; H. Xu, The University of Texas Health Science Center at Houston
Over the last few decades, the advancement of digitalization in healthcare and scientific research has generated a large amount of data valuable for clinical and translation research. For example, the wide adoption of electronic health record (EHR) system has established large practice-based cross-sectional and longitudinal datasets. The digitalization of clinical research makes literature and study protocols computationally accessible resources for scientific discovery. However, much of the information in those resources is embedded and cannot be directly used. Therefore, Natural Language Processing (NLP), which can extract structured information from text, has received great attention and has played a critical role in leveraging those existing data resources for clinical and translational research. However, the adoption of clinical NLP has been limited due to privacy concerns or lack of portability of NLP systems across different types of clinical text, which indicates challenges and opportunities for designing the next generation of clinical NLP systems. With the advancement of deep learning and privacy-preserving computing, we seek alternative ways in addressing those issues.
A. Alekseyenko, L. Lenert, Medical University of South Carolina
The turn of the decade has challenged all aspects of society with an unprecedented challenge of an emerging healthcare system crisis caused by a novel pathogen. Even before the first official case have been registered in the U.S., COVID-19 has stressed the administrations with concerns about strategies to avoid capacity overrun. At the same time informaticians have received the challenge as an opportunity to demonstrate the utility of solutions and expertise they have accumulated over preceding decades in facing and deflecting exactly this type of crisis. This panel brings together informatics experts at the Medical University of South Carolina to discuss how the COVID-19 challenge has been met by their health system and exactly what steps amounted to a successful comprehensive response by these informatics practitioners.
S. Bernstein, Agency for Healthcare Research and Quality; J. Osheroff, TMIT Consulting, LLC; M. Michaels, Centers for Disease Control and Prevention; B. Alper, EBSCO Clinical Decisions; B. Middleton, Apervita
The AHRQ evidence-based Care Transformation Support (ACTS) initiative has developed a shared vision for health IT-enabled, evidence-informed care delivery and transformation, and a stakeholder-driven Roadmap for broadly achieving this vision. Over 150 individuals from 80+ organizations worked together to create this Roadmap for developing learning health systems that achieve the quadruple aim, including fully leveraging resources from AHRQ and others. This panel will outline AHRQ’s goals, context and process for the initiative, the Roadmap and future vision, as well as the perspectives of 3 participants on implications for their organization and stakeholder groups, and steps they are taking to execute it. Panelists represent CDS and health IT suppliers, federal agencies, standards initiatives, and care transformation experts. After panelist presentations, the second session half will be interactive discussion with attendees (likely including other ACTS participants) to surface opportunities to leverage Roadmap execution efforts to accelerate care transformation within and across organizations.
S. Pitts, Johns Hopkins University School of Medicine; M. Chui, University of Wisconsin-Madison, School of Pharmacy; C. Oltman, NCPDP Foundation; T. Akinwale, Walgreens Co.
Despite implementation of health information technology targeting medication safety, ambulatory adverse drug events (ADEs) prompt over four million people to seek medical care and result in $8 billion in health care expenditures annually. Communication between prescribers and pharmacies is critical to prevent ADE. The National Council for Prescription Drug Programs’ (NCPDP) SCRIPT standard supports the functionality to send electronic prescription cancellations from EHRs to pharmacies, known as CancelRx. Our panel will present novel research and practical experience on CancelRx implementation from the perspectives of two health systems, a large retail pharmacy, and the NCPDP.
E. Scheufele, IBM; M. Ball, University of Texas; J. Cooper, Healthcare Management Systems; J. Murphy, IBM; M. Palchuk, TriNetx
Individuals with biomedical informatics training are increasingly being sought by technology companies pursuing opportunities in healthcare. In this panel, an experienced group of biomedical informaticians describe the wide variety of roles in the healthcare technology industry including product development, scientific evaluation, and executive leadership. Participants will address the transition from academics to industry, the tension in balancing business priorities with scientific rigor, and the keys to success in an industry career. The advantages and challenges of pursuing a career in industry in a rapidly changing global environment will be discussed. Finally, panelists will provide recommendations on what to expect as the landscape for biomedical informatics career opportunities continues to evolve.
T. Campion, Weill Cornell Medicine ; F. Shaya, University of Maryland School of Pharmacy; D. Robins, R. Brody, J. Finkelstein, Icahn School of Medicine at Mount Sinai
Electronic consent (e-consent) is increasingly being used as means to facilitate patient engagement in clinical research. Recent reports demonstrated significant potential of e-consent in catalyzing patient-oriented translational clinical research. However, adoption of e-consent for research remains low despite evidence of improved patient comprehension, usability, and workflow processes. To accelerate successful adoption of e-consent, the experience and lessons learned at existing e-consent sites must be organized and shared among institutions conducting clinical trials. Panelists from early adopter institutions will compare and contrast their experiences developing, assessing and implementing e-consent in their institutions for clinical trials and universal biobanking consents. Perspectives on the integration of e-consent platforms into commercial EHR, CTMS, IRB and biobanking platforms will be discussed. Key implementation issues that will be addressed include challenges and barriers to developing and implementing e-consent. After participating in this session, participants will be able to formulate optimal approaches for successful adoption of e-consent.
G. Kuperman, Memorial Sloan Kettering; A. Wagner, Dana-Farber Cancer Institute; E. Weiss, Roswell Park Cancer Center; C. DesRoches, OpenNotes.org
Sharing notes with patients is becoming increasingly prevalent and much is being learned. Patients report several benefits when sharing notes is implemented and only a small minority report increased anxiety or worry. Providers are apprehensive at the prospect of sharing notes, but anticipated fears almost never materialize. Implementing the sharing of notes in the oncology setting requires special considerations to be taken into account. For example, the prevalence of serious illness is much higher in oncology and the technical aspects of the care may be more complicated. This panel will include representatives from three cancer centers that have implemented the sharing of notes at their organization. These panelists will relate their experiences and lessons learned. The panel will also a representative from OpenNotes.org, an organization dedicated to the advancing patients’ access to their notes. She will speak on the differences between implementing shared notes in oncology vs. non-oncology settings.
N. Benda, M. Reading Turchioe, R. Masterson-Creber, M. Sharko, J. Ancker, Weill Cornell Medicine
New federal regulations (HealthEData, part of the 21 st Century Cures Act) put patients in control of their health records provides unprecedented access to health data. Yet poor health literacy, numeracy, and technology literacy pose barriers to turning data into actionable health knowledge. Informatics researchers have an obligation to apply findings about literacy, numeracy, and inclusive design to avoid perpetuating inequalities in health and the healthcare system. In this panel, we advocate a work systems approach, in which the health information providers, resources, and contexts surrounding the patients are all considered in the presentation of health data. The panelists will share research aimed at promoting patient knowledge comprehension through visualizations created using a work systems approach. After participating in this session, attendees should be able to describe the core constructs of a work systems framework for inclusive design and apply different approaches for translating health data into knowledge.
Y. Zhang, Weill Cornell Medical College/Cornell University; J. Chen, Stanford University; A. Wright, Vanderbilt University; J. Ancker, Weill Cornell Medical College, Cornell University; M. Tobias, Phrase Health
Throughout its history, clinical decision support (CDS) has generally been rule-driven. With access to larger data sets and greater computational power, healthcare organizations have increasing opportunities to make CDS data-driven. Data-driven approaches allow organizations to develop patient-centered and customizable CDS that are most appropriate based on insights generated from data at each organization. Medical informatics has multiple opportunities to promote and facilitate the design, implementation, and evaluation of data-driven CDS. In this panel, we will focus on CDS for Computerized Provider Order Entry (CPOE) and discuss multiple challenges and opportunities around three areas: how to develop data-driven CDS; how to evaluate; and how to implement and govern to improve health and healthcare operations. After participating in this session, attendees should be able to define data-driven CDS, identify barriers against its development and use, and describe informatics approaches to overcome these barriers through case studies at multiple organizations across the US.
R. Li, Stanford University School of Medicine ; N. Muthu, Children's Hospital of Philadelphia; M. Smith, Stanford University School of Medicine ; S. Kandaswamy, Emory University; J. Chen, Stanford University School of Medicine
Artificial Intelligence (AI), and in particular Machine Learning (ML), has generated much excitement, but has yet to produce significant improvements in healthcare delivery. While the emergence of increasingly accurate ML models has indicated the potential for tasks in healthcare to be supported or even replaced by AI, this shift in the nature work has not yet occurred at scale. This panel will discuss current state barriers to improving healthcare with ML, for implementing ML enabled solutions informed by the socio-technical model for health information technology (HIT), and present two case studies of active ML implementation efforts at two health systems that incorporate perspectives from clinical informatics, data and computer science, process improvement, and human factors engineering. The panel includes a diverse set of viewpoints with expertise in informatics implementation (R.L.), cognitive informatics (N.M.), human factors (S.K.), process improvement (M.S.) and data science (J.C.).
M. Matheny, Vanderbilt University Medical Center, Tennessee Valley Healthcare System VA; S. Luther, James A. Haley VA; W. Roddy, Uniformed Services University of the Health Sciences; C. Brandt, J. Erdos, Yale University
The NIH-DoD-VA Pain Management Collaboratory (PMC) is a consortium of eleven R-01 type pragmatic clinical trials (PCTs) within VA and DoD healthcare settings examining non-pharmacologic strategies for pain management, and a coordinating center (PMC3), tasked to coordinate, share, and harmonize data, information, and study designs across the trials. During the planning phase for each PCT, informatics-focused PMC work groups (WG) were formed to help support and harmonize activities across projects. For this panel, we included representatives from the PMC3 to present key challenges brought forward by the PCTs, and summarize the processes to address these challenges and promote cross-project standardization based on best practices, as well as lessons learned and key successes in the first two years of the consortium. Key challenges include recruitment of patients using the electronic health record; data collection, standardization, and harmonization across multiple sites and organizations; clinical phenotyping in the EHR for pain conditions; and intra- and inter-organization data sharing.
W. Chapman, University of Melbourne; A. Grando, Arizona State University; M. Johns, The Monarch Center; G. Savova, Harvard University; M. Hightower, University of Utah Health
To address the leadership gap for women in informatics and digital health, the Women in AMIA Committee launched an intensive 6-month leadership program. Twenty-four women from across the U.S. participated in the hybrid program that included three in-person sessions. We performed evaluations throughout the program to assess effectiveness of the program as it was delivered and to compare confidence and skills before, during, and after the program. In this panel, we will report on the results of our evaluations and will discuss the program from several points of view: 1) sponsors who financially supported women from their teams to attend the program, 2) the program developers and curriculum coaches, and 3) participants of the program.
T. Hernandez-Boussard, J. Lossio-Ventura, Stanford University; A. Syrowatka, Brigham and Women's Hospital; W. Song, Brigham and Women's Hospital, Harvard Medical School; P. Dykes, Brigham and Women's Hospital
Overprescribing opioids represents a public health crisis in the U.S. that challenges classic clinical decision making. Many opioid prescriptions for pain management progress to dependence or addiction. Several guidelines have been developed by national organizations to provide general strategies for opioid prescribing. These guidelines suggest the need to quantify and monitor opioid utilization through electronic clinical quality measures (eCQMs) and large-scale clinical data-based algorithms. However, it is difficult to pull accurate and comparable opioid information from real-world data to develop reliable applications. Moreover, studies related to opioid standardization and quantification do not provide details about their processes and have not made code publicly available for reuse and/or comparison, which has not substantially improved current opioid prescribing practices. The objective of this panel is to: 1) introduce eCQMs for opioid prescriptions and their evaluation in real-world settings; 2) describe the development of a tool to systematically extract and convert opioid prescriptions to a morphine milligram equivalent (MME); and 3) introduce the process of testing the generalizability of the opioid eCQMs and dosage conversion tool across two medical systems: the Brigham and Women's Hospital and Stanford University.
W. Hersh, Oregon Health & Science University; G. Jackson, IBM Watson Health; M. Williams, Geisinger; C. Walsh, Vanderbilt University; D. Dorr, Oregon Health & Science University
Applications of machine learning and artificial intelligence have the ability to transform health and healthcare delivery, yet most systems have been evaluated only with small data sets in artificial settings. This panel reports on the experiences of informatics researchers seeking to test such applications in real-world clinical settings. Each panelist will describe efforts to conduct evaluations in healthcare environments, including the challenges they faced from organizational leadership, clinical workflows, and regulatory boards.
H. Wong, A. Shoaibi, U.S. Food and Drug Administration; C. Reich, IQVIA; K. Moll, IBM; T. Hernandez-Boussard, Stanford University
The Food and Drug Administration (FDA)/Center for Biologics Evaluation and Research (CBER) monitors the safety and effectiveness of biologic products, including vaccines, blood and blood derived products and advanced therapeutics. The 21st Century Cures Act in 2016 directs the FDA to use real-world data to develop real-world evidence to support regulatory decisions. CBER established the Biologics Effectiveness and Safety (BEST) Initiative in 2017 that includes a network of electronic health records (EHR) and linked administrative claims-EHR data sources.
The panel presents an overview of the BEST Initiative and informatics considerations for the secondary use of EHRs. Using case studies in biologics, the panel will discuss ascertainment of exposures and outcomes in EHRs, development of semi-automated tools for EHR-based chart review and adverse event reporting, and interoperability and portability of computable phenotypes across EHRs. This interactive session will engage participants to understand the goals and challenges of biologics surveillance using EHRs.
K. Kawamoto, University of Utah; B. Rhodes, Dynamic Content Group; E. McPeek Hinz, Duke University Health System; J. Malinowski, Cerner; D. Hurwitz, Allscripts
Commercial electronic health record (EHR) systems are continuing to expand support for standards-based clinical decision support (CDS) frameworks in their general-release software. Adoption of the HL7 Clinical Decision Support (CDS) Hooks standard, as well as the HL7 Clinical Quality Language (CQL) standard for both local and remote decision support. The US Core Fast Healthcare Interoperability Resources (FHIR) profiles for data access continues to serve as a foundation for this interoperability. This panel will describe the next phase of the work of 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. This includes broadening existing pilot implementations; new pilot sites and technology environments; and collaboration with related decision support and quality measurement projects.
T. Leung, Maastricht University; J. Ancker, Weill Cornell Medical College; J. Cimino, University of Alabama-Birmingham School of Medicine; H. Ross, WittKiefer Consultants; H. Wu, Indiana University-Purdue University Indianapolis
Writing letters of support in academia is a vital skill to sponsor protégés, advance the careers of mentees and colleagues, and promote diversity, equity, and inclusion in informatics and science in general. Letters are an essential part of communicating the value and qualifications of an informatician, and the traits of that individual must be highlighted in a way that focuses on their abilities as a leader, awardee, and deserving candidate – and avoid language about gendered or stereotyped characteristics that can be harmful for their achievement.
In this panel, the speakers will introduce the art of letter writing, drawing from evidence-based practices and individual experiences with writing and reviewing letters of recommendation for colleagues’ promotion, career advancement, award conferral, job applications, and more. The objectives of this panel are to: (1) learn about the art of letter writing (2) reflect on one’s own letter writing; (3) learn how to write a stellar letter of support; and (4) begin writing the next letter of support. Participants are strongly encouraged to bring a one-page sample or excerpt of a prior letter of support they have written (with removal of any identifiable personal or institutional information of the individual for whom the letter is written).
R. Carroll, K. Wuichet, Vanderbilt University Medical Center; C. Phillips, The Children's Hospital of Philadelphia; M. Holko, National Institutes of Health; A. Heath, The Children's Hospital of Philadelphia
The rise of interconnectivity between data applications and platforms, in both clinical and research contexts, provides an unprecedented opportunity for data sharing, re-use, and application. As increasingly large amounts of data are available for utilization by an ever-expanding audience, the approaches and methodologies for data quality and curation must evolve rapidly to meet the promise of translational and clinical impact. This panel will present experiences from managing these processes in clinical systems and NIH research programs. Specifically, we will focus on experiences across the All of Us Research Program, NIH Common Fund-supported Kids First Data Resource Center, NHGRI AnVIL, the Children’s Hospital of Philadelphia (CHOP) Children’s Cancer Health Informatics Program, and other consortium-based initiatives. Across these initiatives there are similar data lifecycle challenges, encompassing data collection in study-specific and real-world contexts, processes and tooling for curation, and ways to distribute this information for highest utility.
K. Chaney, T. Zayas-Cabán, Office of the National Coordinator for Health Information Technology; P. White, Salt Lake City Veterans Affairs Medical Center
Widespread adoption of electronic health record (EHR) systems and consumer electronics has resulted in large volumes of health-related data potentially available for research. However, realizing the value of these data for research has been slow due to challenges in both the data and the health information technology (IT) infrastructure that supports it. This panel will discuss efforts by the Office of the National Coordinator for Health Information Technology (ONC) to guide the development of a future health IT infrastructure that supports use of electronic health data for research. This work addresses three objectives: (1) articulate a vision for an ideal health information ecosystem that supports research; (2) identify stakeholders’ priorities needed to address challenges within the current ecosystem; and, (3) a Policy and Development Agenda that will contribute to realizing an ideal health information ecosystem in which both the health IT infrastructure and the data it supports are optimized.
A. Boxwala, Elimu; M. Burton, Apervita; H. Head, Optum; V. Kashyap, Optum; D. Sottara, Mayo Clinic
Authoring of content for clinical decision support (CDS) involves translating guideline narratives into a computer- executable format that can be used to create patient care recommendations. The current state of practice of authoring involves a time-consuming and iterative development amongst clinical subject matter experts (SMEs), informaticists and CDS developers. We propose a panel discussion exploring the role of a clinical SME in the design, construction and deployment of CDS Systems.
J. Holmes, University of Pennsylvania; R. Bellazzi, University of Pavia; C. Combi, University of Verona; J. Moore, University of Pennsylvania; N. Peek, University of Manchester
There is a resurgence of interest in artificial intelligence (AI) applications in biomedical domains. There is a concomitant interest in how such applications reach a conclusion, such as a prediction or classification. Given that AI systems in biomedicine can affect a user’s decision about providing patient care or choosing a particular algorithm for mining data, it is critically important for informaticians and computer scientists to create explainable AI systems to address this. This panel will review the history of explainablity in AI, and introduce four areas in which AI is developed, used, and evaluated.
T. Miller, Boston Children's Hospital, Harvard Medical School; D. Dligach, Loyola University; F. Wang, Weill Cornell Medicine; Y. Si, University of Texas Health Science Center; F. Meric-Bernstam, University of Texas MD Anderson Cancer Center
Electronic health records (EHRs) contain a massive amount of personal health data than can be used by artificial intelligence to improve clinical research and patient care. The information in EHRs, however, is scattered across a variety of heterogeneous data types. A glaring need for future applications of AI to EHRs is a way to develop general representations for different data types, and methods for merging them into a complete picture. In this panel discussion, we bring together experts with backgrounds in building representations from different data types, for a much-needed discussion across research areas that are frequently working separately. In the first part of the panel discussion, each participant will give a brief overview of their work and other state of the art work in their area, and the second part of the panel will be devoted to questions from the moderator and the audience about how to bridge the respective representations.
N. Weiskopf, Oregon Health & Science University; M. Greer, University of Arkansas for Medical Sciences; K. Natarajan, Columbia University; C. Thompson, San Diego State University; H. Lehmann, Johns Hopkins University
Electronic health record (EHR) data for are increasingly utilized for quality improvement, clinical research, and other endeavors related to medical care and research. An expected benefit of these data is high generalizability—or external validity—between patient populations. Unfortunately, issues of data bias, i.e., data quality problems that do not occur completely at random, but rather are driven by underlying mechanisms, have the potential to limit external validity. In this panel we will explore the potential causes of data bias in EHR data, discuss approaches for their detection and amelioration, present examples of existing work that implements these approaches, and articulate unmet needs.
P. Hsueh, Viome Inc.; J. Florez-Arango, The University of Texas Health Science Center; C. Kuziemsky, MacEwan University; C. Nøhr, University of Southern Denmark; V. Patel, New York Academy of Medicine; X. Zhu, Yale University
With continued focus on the impetus to move from fee-for-service to value-based care payment models, one major challenge is to enable effective critical response for patient-centered care delivery in the conventional care settings, which commonly limit patient-provider interactions to short visits separated by long time intervals. Recently, a number of alternative care delivery models have emerged to support asynchronous, non-traditional communication modalities between patients and providers beyond in-person clinical visits. This is pertinent not only to chronic and preventive care, but also to other care scenarios that require critical responses, such as the situation we are currently experiencing during the COVID-19 pandemic. The overarching goal of this panel is thus to bring the informatics community together to work on best practices for studying patient-provider communications that consider these issues, as well as the different contexts where communication takes place. The panelists will draw on several international case studies to share lessons learned in the patient-centered design process of emerging care delivery models. The panelists are leading researchers in the field and will not only provide global case studies to illustrate the untapped potential of patient-provider communication, but also specify the research methods and protocols used in the case studies, including the pros and cons of them. In particular, the adoption and evaluation in current practice to handle critical responses such as COVID-19 pandemic will be discussed.
T. Cullen, Regenstrief Institute, Inc., Indiana University; F. Holl, Neu-Ulm University of Applied Sciences, University of Munich; M. Gong, Chinese Academy of Medical Sciences; A. Kanter, Columbia University, Intelligent Medical Objects; P. Chang, National Yang Ming University
The panel members will chronicle the unique informatics needs, challenges, and approaches to health care demands during a pandemic, illustrated through the initial and current responses to the COVID-19 epidemic, in resource-constrained environments. The ability to identify and address challenges and difficult clinical situations with bioinformatics is critical in the effective, rapid delivery of public health and clinical responses during pandemic conditions. Informatic tools need to support important functions, such as contact tracing, to limit the spread of new cases and trials of new diagnostic/therapeutic approaches. We will describe and compare the unique informatics needs, challenges, and solutions, such as analytics and clinical decision support, that have been developed and deployed during epidemic situations in LMIC.
J. Richardson, RTI International; A. Smith, National Cancer Institute; A. Cheville, Mayo Clinic; M. Bass, Northwestern University; M. Hassett, Dana-Farber Cancer Institute
As part of its Cancer Moonshot SM, in 2018 the National Cancer Institute established an initiative to fund a consortium that aims to improve the monitoring and management of patients’ cancer-related symptoms using informatics solutions. The consortium, Improving the Management of symptoms during And following Cancer Treatment (IMPACT), is comprised of three research centers and a coordinating center that are tasked with collecting and sharing symptom data from across the cancer care continuum – at the point-of-care in oncology clinics and via remote settings – and evaluating the effects that cancer-related symptom management tools and data have on patients, through cancer informatics interventions including patient-directed mobile health interventions and symptom-based clinical decision support tools. This panel will highlight their challenges and solutions they have encountered. Learning objectives include understanding current issues with developing and implementing systems for tracking cancer-related symptoms and using decision support tools for effective supportive care and symptom management.
V. Patel, Office of the National Coordinator for Health Information Technology; J. Adler-Milstein, University of California San Francisco; J. Everson, Vanderbilt University Medical Center; D. Kendrick, MyHealth Access Network, University of Oklahoma; M. Berry, Office of the National Coordinator for Health Information Technology
A decade after HITECH, achieving robust health information exchange (HIE) and addressing barriers to interoperability remain a policy priority. Our panel will describe the role of Health Information Organizations (HIOs) and how health IT provisions of the 21 st Century Cures Act—most notably the Trusted Exchange Framework and Common Agreement (TEFCA)—and broader market dynamics are changing approaches to HIE. We will share timely data on how HIOs are responding to these national policy efforts. In addition, we will share data on HIOs experiences with information blocking, one of the key barriers to interoperability.
D. Borbolla, C. Staes, L. Heermann Langford, V. Nguyen, University of Utah
The increase in the adoption of Fast Health Interoperability Resources (FHIR) facilitated by new regulations creates opportunities for informatics innovations. To accelerate innovation, however, we need a trained workforce. In this panel, we will present four experiences with using FHIR to train clinicians, decision-makers in a global context, and nursing informatics and biomedical informatics students. After attending the panel, participants will be able to: a) describe how to train clinicians, students, and decision-makers in a global context in the use of the FHIR standard, and b) incorporate active learning strategies related to standards and interoperability (using FHIR specifically) into their academic training programs.
J. Norton, National Institute of Diabetes and Digestive and Kidney Diseases; A. Goodman, M. Michaels, Centers for Disease Control and Prevention; R. Taylor, Yale University School of Medicine; A. Gonzales, U.S. Department of Health and Human Services
Maximizing the efficiency and utilization of electronic health records (EHRs) is one of the Department of Health and Human Services’ (HHS) key priorities. This panel will highlight four projects funded under the Office of the Secretary Patient-Centered Outcomes Research Trust Fund (OS-PCORTF), that are developing promising solutions to enhance EHR capacity to support data collection from multiple healthcare settings including research-based settings, health care surveys, emergency departments, and community health clinics. These projects focus on priority populations including those with multiple chronic conditions, children with obesity, and patients with health outcomes related to opioid use disorder. Panelists will each provide an overview of their project, describe the projects’ contributions to care coordination, and discuss informatics-related opportunities to care coordination.
F. Reza, P. Podila, M. Antoine, S. Karki, A. Ojo, Centers for Disease Control and Prevention
The Centers for Disease Control and Prevention (CDC) Public Health Informatics Fellowship Program (PHIFP) Info-Aids provide service-learning opportunities for informatics professionals to transform data into public health action. Info-Aids are short-term technical assistance provided by PHIFP fellows to requesting agencies that meet crucial informatics needs. This panel will showcase Info-Aids that provided informatics expertise in federal, state, local, territorial, and global settings and include the following projects: (a) From Big Data to Big Decisions on Medicaid Reporting, Integration, and Expansion for Public Health in Utah; (b) Value of a Centralized Registry in Low Resource Settings: Acute Rheumatic Fever and Rheumatic Heart Disease Prevention and Control in American Samoa; (c) A Quest For New Data Sources To Improve National Immunization Coverage: Data Source Comparison Dashboard; (d) Data — the prevention and treatment tool for The Carter Center’s Chad Guinea Worm Eradication Program; and (e) NMRS Monitor: Nigeria's journey towards successful EMR implementation.