AMIA 2021 Virtual Day Program

The National Institutes of Health recognizes the importance of biomedical informatics innovations in accelerating data-driven discovery in biomedical science, clinical care and patient self-management. Of key importance is accessing information in the electronic health record. In the past three years, the NIH has expressed its public commitment to encourage researchers to adopts informatics innovations such as common data elements (CDEs), the Fast Healthcare Interoperability Resource (FHIR) standard, and the United States Core Data for Interoperability (USCDI). Policy statements are most effective when they are accompanied by operational procedures and vocal leadership direction. The purpose of this panel is to demonstrate the strength of informatics know-how among the leadership of the NIH, identify existing commitments by the NIH to ensure accelerate the research value of clinical health records, and to glean insights of strategies the NIH should consider as they develop the informatics infrastructure for the future of biomedical discovery.

The increasing focus of AI on the frontier of clinical applications has led to the accumulation of evidence for tis clinical decision support across a wide range of specialties such as radiology, oncology, ophthalmology, and critical care. In recent years, FDA has published action plans for regulating AI/ML-Based Software as a Medical Device (SaMD) and approved more than 160 medical AI products5. Meanwhile, due to the sensitivity of clinical AI in producing unintended consequences such as incorrect diagnoses, unnecessary treatments, and racial disparities, there is also a growing concern on how to produce safe, reliable, and trustworthy AI for adaptive clinical decision support (Adaptive CDS) systems. In the emerging field of AI bias studies, the growing concern has yielded an increasing body of literature exploring the methodology of estimating the impact of bias and coming up with counter strategies. However, in real-life scenarios, oftentimes the implementation of such applications requires experience on the trenches for how best to embed the actionable insights in clinical workflows and warrant behavioral change and sustainable adoption from clinicians. In order to fulfill the potential of applying health AI to ensure patient safety in adaptive CDS systems, in this panel we have gathered top-notch thought leaders who are currently working on improving the use of AI/ML insights in clinical workflow for patient safety. The panel will be a fact-full discussion, with concrete examples of studies reflecting on what principles have we learned so far, as well as questions that need to be solved next. It will also report on the active on-going experiments that are expected to lead us to the next level of understanding towards a systematic framework.

Cultivating a diverse and inclusive workforce is critical for the biomedical informatics profession, both from a perspective of fairness and equity and to avoid technology that fails to address concerns of significant portions of the population or that introduces unintentional biases. However, having diversity and representation in the informatics workforce requires intentionality and deliberate action to combat systemic inequalities and requires intentional effort to ensure inclusive professional environments. This challenge is not unique to biomedical informatics, but actions are required in order to continue our work of transforming health and healthcare. This panel reports on the experiences of informaticians and data scientists across academic and industry settings in creating diverse and inclusive environments. The learning objectives include: 1) describe why a diverse workforce and inclusive working environments are critically important goals for the informatics field; 2) gain an understanding of lessons learned and best practices for broadening participation in informatics; and 3) learn about resources to support efforts to broaden participation in informatics.

Reducing Physicians' Cognitive Load During Chart Review: A Problem-Oriented Summary of the Patient Electronic Record

• Jennifer Liang, IBM Research

Characterization of Electronic Health Record Documentation Shortcuts: Does the use of SmartPhrases increase efficiency in the Emergency Department?

• Rimma Perotte, Hackensack University Medical Center

Automatic Assignment of Radiology Examination Protocols Using Pre-trained Language Models with Knowledge Distillation

• Wilson Lau, University of Washington

Identifying ARDS using the Hierarchical Attention Network with Sentence Objectives Framework

• Kevin Lybarger, University of Washington

Comparing Deep Learning and Conventional Machine Learning Models for Predicting Mental Illness from History of Present Illness Notations

• Ingroj Shrestha, The University of Iowa

Simulating Screening for Risk of Childhood Diabetes: The Collaborative Open Outcomes tooL (COOL)

• Mohamed Ghalwash, IBM

Human Factors Considerations in Transitions in Care Clinical Decision Support System Implementation Studies

• Erin Kennedy, University of Pennsylvania

Outpatient Portal Use in Pregnant Women for Blood Glucose Management

• Priti Singh, The Ohio State University

Impact of a Clinical Decision Support Tool for Cardiovascular Preventive Care in Community Health Centers: Randomized Trial Results

• Rachel Gold, KPCHR, OCHIN

Systematic replication of smoking disease associations in the All of Us Research Program

• David Schlueter, NHGRI

Controversial Trials First: Identifying Disagreement Between Clinical Guidelines and New Evidence

• Florian Borchert, Digital Health Center, Hasso Plattner Institute

Discovering Associations between Social Determinants and Health Outcomes: Merging Knowledge Graphs from Literature and Electronic Health Data

• Yoonyoung Park, IBM Research

Identifying Sleep-Related Factors for Cognitive Function in a Hispanics/Latinos Cohort: A Dual Random Forest Approach

• Xiaojin Li, The University of Texas Health Science Center at Houston

Augmenting Barbershop Initiative Health Promotion with Wearables in African-American Populations

• Lisa Anne Bove, UNCW

The impact of COVID-19 pandemic on health information sharing and patient-generated health data: findings from the Health Information National Trends Survey 2020

• Zidan Wang, Northwestern University

Trust and credibility of information sources related to COVID-19 among high-risk diverse adults at the onset of the New York City outbreak: A cross-sectional survey conducted via a community health portal

• Rita Kukafka, Columbia University, Department of Biomedical Informatics

Detecting Fine-Grained Emotions on Social Media during major Disease Outbreaks: Health and Well-being in London before and during the COVID-19 Pandemic

• Olanrewaju Aduragba, Durham University

On the explainability of hospitalization prediction on a large COVID-19 patient dataset

• Ivan Girardi, IBM

Using informatics framework to identify barriers and enablers for practice facilitation in primary care quality improvement

• Jiancheng Ye, Northwestern

Feasibility of using machine learning on insurance claims to identify correlates of lower extremity amputation in insured adults with type 2 diabetes who initiate treatment with sodium-glucose co-transporter 2 (SGLT-2) inhibitors

• Yuan Luo, Northwestern University at Chicago

Leveraging the MedDRA Biomedical Terminology and Weak Labeling for Mental Health Symptom Surveillance from patient notes

• Marie Humbert-Droz, Stanford University

Semantic Expansion of Clinician Generated Data Preferences for Automatic Patient Data Summarization

• Ashutosh Jadhav, IBM Research

Hybrid Ensemble-Rule Algorithm for Improved MEDLINE® Sentence Boundary Detection

• Daniel Le, NIH/NLM/LHC

Outcome Prediction from Behaviour Change Intervention Evaluations using a Combination of Node and Word Embedding

• Martin Gleize, IBM Research

TeamTat: A Collaborative Text Annotation Tool for Creating Gold-Standard Corpora/NLM-Chem, a Gold-Standard Corpus for Chemical Entity Recognition in PubMed Full Text Literature

• Rezarta Islamaj, National Library of Medicine/ National Institutes of Health

Feasibility of Mobile and Sensor Technology for Remote Monitoring in Cancer Care and Prevention

• Susan Peterson, MD Anderson Cancer Center

Older adults’ personal health information management: The role and perspective of various healthcare providers

• Alyssa Bosold, University of Washington

Children's Designs for the Future of Telehealth

• Erin Beneteau, University of Washington

iCardia4HF (Phase I clinical trial): Feasibility and efficacy of using commercially available mHealth technologies and tailored text messages to improve self-care in patients with heart failure

• Spyros Kitsiou, University of Illinois at Chicago

Challenges to Public Health Reporting Experienced by Hospitals

• Chelsea Richwine, ONC

Extraction of Electronic Health Record Data from FHIR for Automated Breast Cancer Risk Assessment

• Julia McGuinness, Columbia University Irving Medical Center

Enhancing the IDEAS Framework with Ontology: Designing Digital Interventions for Improving Cancer Patients' Wellbeing

• Mor Peleg, University of Haifa

Semantic Search for Large Scale Clinical Ontologies

• Hoa Ngo, The Australian E-Health Research Centre

Towards more patient friendly clinical notes through language models and ontologies

• Francesco Moramarco, Babylon Health

Can ICD-11 Replace ICD-10-CM for Morbidity Coding in the U.S.?

• Kin Wah Fung, US National Library of Medicine

Practical Perfusion Quantification in Multispectral Endoscopic Video: Using the Minutes after ICG Administration to Assess Tissue Pathology

• Jonathan Epperlein, IBM Research Europe

A Comparison of Exhaustive and Non-lattice-based Methods for Auditing Hierarchical Relations in Gene Ontology

• Rashmie Abeysinghe, University of Texas Health Science Center at Houston

First-line drug resistance profiling of Mycobacterium tuberculosis: a machine learning approach

• Stephanie Muller, IBM

On Predicting Recurrence in Early Stage Non-small Cell Lung Cancer

• Mohan Timilsina, Data Science Institute, National University of Ireland Galway

Machine Learning Predictability of Clinical Next Generation Sequencing for Hematologic Malignancies to Guide High-Value Precision Medicine

• Grace Y.E. Kim, Stanford University

Machine Learning for Predicting Hospital Acquired Pressure Injuries in ICU Patients: From Explainable AI to Ensemble Super Learners

• Jenny Alderden, University of Utah College of Nursing

Colonizing Microbiome as a Determinant of COVID-19 Outcome: A Pilot Study

• Alexander Alekseyenko, Medical University of South Carolina

COVID-19 Trial eSource Data Capture Efficiencies using OneSource

• Adam Asare, UCSF / Quantum Leap Health Care

Challenges in Designing A Notification System for Electronic Medical Records: an Exploration of User Expectations

• Katherine Blondon, University Hospitals of Geneva

Development of a multi-class deep-learning algorithm capable of diagnosing many histopathologic entities in digital pathology

• Gregory Brown, NIH/NCI

Strategies to Engage Traditionally Marginalized Patients in Patient Portal

• Heidi Carpenter, Vanderbilt University Medical School

Localize the Implementation of FDA List of Confusable Drug Names with the Application of Taiwan National Health Insurance Database: A Study of Text Similarity Computation with Real-World Data

• Ya Lin Chen, Teipei Medical University

Machine Learning for Time to Event Biomedical Research

• Ioana Danciu, Oak Ridge National Lab

An AI Model to Screen for the Novel Coronavirus in Emergency Department Clinical Notes

• John Del Gaizo, Medical University of South Carolina

Keys to a Successful Public Health Information System Implementation: Experience with COVID-19 Vaccine Ordering System Development and Implementation Across the United States

• Roua El Kalach, CDC

Patient-Centered Care Practices in Secure Messaging: Qualitative Analysis of Interviews with VHA Primary Care Teams

• Lynn Garvin, VA Boston Healthcare System

A Comparative Time-to-Event Analysis Across Health Systems

• Mohamed Ghalwash, IBM

Utilizing Advanced Visualization Technology to Study Home Care Challenges

• Denise Goldsmith, NIH/NINR

Comparison of a patient cohort and predictive models derived from local academic medical centers versus a national health database

• Benjamin Goldstein, Duke University

Expanding Exposure Notification Verification Code Issuance in WA State

• Amanda Higgins, Washington State Department of Health

Evaluation of the Effectiveness of an Opioid Reversal Alert at an Academic Medical Center

• Bernard Hsia, UF Health

A clinical decision support system (CDSS) ontology to facilitate portable vaccination CDSS rules: preliminary results

• Xia Jing, Clemson University

Patient-Specific Modeling with Lazy Random Forest (LazyRF)

• Adriana Johnson, University of Pittsburgh

Dissemination and Implementation of a Child Abuse Clinical Decision Support System

• Sundas Khan, Northwell Health

U.S. COVID-19 Surveillance in PCORnet®

• Sheryl Kluberg, Harvard Pilgrim Health Care Institute

Systematic Study of Publicly Available Resources Addressing Legal Data Sharing Issues

• Hye-Chung Kum, The Texas A&M University

Adoption of Telemedicine During the COVID-19 Pandemic: Perspectives of Primary Healthcare Providers

• Joe Lintz, Parker University

Identifying Criteria for Antonym Generation from Collocates in Corpora

• Chris Lu, NIH/NLM/MSC

The Real-World Trend Analysis of Cancer Drugs Empowered by Natural Language Processing

• Meng Ma, Sema4

PainRE-Life: A FHIR Based HUB for the management and support of patients with chronic pain

• Sara Marceglia, Università degli Studi di Trieste

Analyzing COVID-19 Case Report Forms Data from the All of Us Program

• Craig Mayer, National Library of Medicine

Characterizing Outpatient Portal Use among Women Receiving Obstetric and High-risk Pregnancy Care

• Evan Morgan, Ohio State University Medical Center

Optimizing Clinician Alerts for Concerning Symptoms in an Electronic Patient-Reported Outcomes (ePRO) System in Community Oncology

• Amila Patel, Navigating Cancer

Mining for parameters in literature: the case of defining electrical stimulation protocols

• Marco Prenassi, Università degli studi di Trieste

Usability and Accuracy Evaluation of the Fitbit Versa 3 Smart Watch Compared to Research-Grade Devices

• Blaine Reeder, University of Missouri

Comparing Performance of Hip-worn and Wrist-worn Research-Grade Activity Monitors for Daily Activities

• Blaine Reeder, University of Missouri

Digital Methodology for Mobile Clinical Decision Support Development

• Blaine Reeder, University of Missouri

Chronic Medical Conditions effects on Telemedicine Implementation During the COVID-19 Pandemic in Senior US Adults

• Jorge Rodriguez, UIC

Utilizing a Clinical Decision Support Alert to Increase Nasal MRSA PCR Ordering and Reduce Vancomycin Course Duration

• Diana Schreier, Mayo Clinic

Digital Mental Health Priorities of People Affected by Mental Health Problems: A Pan-Canadian Survey

• Nelson Shen, Centre for Addiction and Mental Health

Comparison of Three Phrase Chunking Approaches on Medical Concept Coverage in Clinical Text

• Grace Turner, University of Washington

Detecting Major Depressive Disorder from Clinical Notes using Neural Language Models with Distant Supervision

• Yanshan Wang, University of Pittsburgh

Classifying Infection Risk Following Pediatric Cardiac Surgery

• Kaitlin Williamson, Vanderbilt University Medical Center

Development and Validation of a Survival Score for the Emergency Department in Singapore

• Feng Xie, National University of Singapore

Hand hygiene monitoring by positioning technology utilizing IoT devices

• Keiko Yamashita, Nagoya University Hospital

Improving COVID-19 Vaccine Uptake of Mental Illness Patients using Learning Health System Framework

• Xiaoming Zeng, University of North Carolina at Chapel Hill