Implementation of Asynchronous, Virtual Quality and Safety Huddles

Presentation Time: 01:30 PM - 01:45 PM

Abstract Keywords: Collaborative Workflow Systems, Implementation Science and Deployment, Reproducible Research Methods and Tools
Primary Track: Clinical Research Informatics
Programmatic Theme: Real-World Evidence in Informatics: Bridging the Gap between Research and Practice

Health care delivery processes have been engineered to provide safe, timely, efficient care to patients. Quality Improvement and Patient Safety (QIPS) champions, who are often busy clinicians, ensure that frontline staff follow processes aligned with organizational goals. They rely on tools for timely detection and interventions to close care gaps. Huddles are commonly utilized tool to address care gaps and brainstorm ways to prevent future events. Barriers such as scheduling conflicts, space, etc., often impede synchronous huddles. We utilized electronic health record (EHR) and robotic process automation (RPA) to assist QIPS with detection and communication with frontline providers. At NYULH, we implemented virtual quality and safety huddles (VQuaSH) via RPA. Our leadership identified safety concerns for VQuaSH. We measured improvements in First Contact Provider (FCP) coverage and compliance with admission medication reconciliation. FCP coverage increased from 81.5% to 88.2%. Admission medication reconciliation compliance increased from 85.0% to 93.0%. Effective implementation of this tool requires clear alignment to organizational initiatives, identified QIPS, and identified staff to act. This is an efficient technology that can be applied to a variety of quality and safety improvements within healthcare.
Speaker(s):
Ulka Kothari, MD
NYU Langone Medical Center

Author(s):
John Will, MPA - NYU Langone Health; Christopher Sonne, MD - NYULMC; Thomas Roncoli, BA - NYU Langone Health; Ajay Mansukhani, BA - NYU Langone Health; Hardev Randhawa, BA - NYU Langone Health; Jonah Feldman, MD, FACP - NYU Langone Health; Paul Testa, MD, JD, MPH - NYU Langone Health;

Impact of Stressful Life Events on Preventive Colon Cancer Screening

Presentation Time: 01:45 PM - 02:00 PM

Abstract Keywords: Social Determinants of Health, Natural Language Processing, Outcomes Research, Clinical Epidemiology, Population Health, EHR-based Phenotyping, Real-World Evidence and Policy Making
Primary Track: Clinical Research Informatics
Programmatic Theme: Harnessing the Power of Large Language Models in Health Data Science

Background: Increase in early onset colorectal cancer makes adherence to screening a significant public health concern with various social determinants playing a crucial role. Stressful life events, such as divorce, marriage, or sudden loss of job, have a unique position among the social determinants of health.
Methods: We applied a large language model (LLM) to social history sections of clinical notes in the health records database of the Medical University of South Carolina to extract recent stressful life events and assess their impact on colorectal cancer screening adherence. We used pattern-matching regular expressions to detect a possible signal in social histories and ran LLM four times on each social history to achieve self-consistency and then used logistic regression to estimate the impact of life events on the probability of having an EHR code related to colorectal cancer screening.
Results: The LLM detected 380 patients with one or more stressful life events and 5,344 patients without. The events with the most negative impact on screening were incarceration (OR 0.26 95% CI 0.08-0.88), becoming homeless (OR 0.18 95% CI 0.02-1.38), separation spouse/partner (OR 0.32 95% CI 0.07-1.42), getting married/partnered (OR 0.59 95% CI 0.22-1.63). Death of somebody close to the patient, excluding spouse, increased the chance of screening (OR 1.21 95% CI 0.71-2.05).
Conclusion: Our findings suggest that stressful life events might have an unexpected impact on screening, with some events, such as experiencing somebody’s death, acting as facilitators to screening.
Speaker(s):
Dmitry Scherbakov, PhD
Medical University of South Carolina

Author(s):
Paul Heider, PhD - Medical University of South Carolina; Ramsey Wehbe, MD, MSAI - Medical University of South Carolina; Leslie Lenert, MD - Medical University of South Carolina; Alexander Alekseyenko, PhD, FAMIA - Medical University of South Carolina; Jihad Obeid, MD - Medical University of South Carolina;

Identifying Accurate and Racially Unbiased Lung Cancer Screening Eligibility Mechanisms for Black and White Smokers with Different Baseline Incidence Rates

Presentation Time: 02:00 PM - 02:15 PM

Abstract Keywords: Fairness and Disparity Research in Health Informatics, Machine Learning, Generative AI, and Predictive Modeling, Data-Driven Research and Discovery, Measuring Outcomes
Primary Track: Data Science/Artificial Intelligence
Programmatic Theme: Health Data Science and Artificial Intelligence Innovation: From Single-Center to Multi-Site

This study involved a retrospective analysis of data from the PLCO (Prostate, Lung, Colorectal, and Ovarian) cancer screening trial to develop a framework for accurate and racially unbiased Lung cancer screening eligibility identification by considering significant differences in Black and White smokers’ baseline incidence rates. The framework integrated the bias mitigation approach with machine learning techniques and showed great promise in mitigating racial biases in lung cancer screening, leading to more equitable healthcare outcomes.
Speaker(s):
Piyawan Conahan, Ph.D.
Moffitt Cancer Center

Author(s):
Lary A. Robinson, M.D. - Thoracic Oncology Program (Surgery), H. Lee Moffitt Cancer Center & Research Institute; Trung Le, Ph.D. - Department of Industrial and Management Systems Engineering, University of South Florida; Gilmer Valdes, Ph.D. - Department of Machine Learning, H. Lee Moffitt Cancer Center & Research Institute; Matthew B. Schabath, Ph.D. - Department of Cancer Epidemiology, H. Lee Moffitt Cancer Center & Research Institute; Margaret M. Byrne, Ph.D. - Department of Health Outcomes and Behavior, H. Lee Moffitt Cancer Center & Research Institute; B Lee Green, PhD; Issam El Naqa, PhD; Yi Luo, PhD - H. Lee Moffitt Cancer Center & Research Institute;

Feasibility of Automated Precharting using GPT-4 in New Specialty Referrals

Presentation Time: 02:15 PM - 02:30 PM

Abstract Keywords: Machine Learning, Generative AI, and Predictive Modeling, Informatics Research/Biomedical Informatics Research Methods, Implementation Science and Deployment
Primary Track: Data Science/Artificial Intelligence
Programmatic Theme: Harnessing the Power of Large Language Models in Health Data Science

This study evaluates the feasibility of using GPT-4 to automate precharting for specialty referrals, focusing on new patients referred to an otolaryngology clinic for nasal congestion. We describe the design decisions and strategies tested in creating this precharting utility, including methods for prompt design and token limit handling. Through iterative testing and building, our tool 95.0% agreement with physician consensus in a small retrospective test sample. Results from a small-scale pilot showed favorable feedback of summaries in a real-world clinical setting, though there was a discrepancy between high intention to use the summary but lower perception of time savings. Our results demonstrate that automated pre-charting with accuracy and clinical relevance can be feasible with large language models such as GPT-4. Our design features can inform the development of vendor chart summarization solutions
Speaker(s):
April Liang, MD
Stanford University

Author(s):
Juan Banda, PhD - Stanford Health Care; Thomas Savage; Abby Pandya, MS, MBA - Stanford Health Care; Rebecca Carey, MBA - Stanford Health Care; Uchechukwu Megwalu, MD, MPH - Stanford University; Michael Chang, MD - Stanford University; Dev Dash - Stanford; Conor Corbin; Aditya Sharma, BSE - Stanford Health Care; Rahul Thapa, Master's - Stanford University; Nikesh Kotecha; Nigam Shah, MBBS - Stanford University; Jonathan Chen, MD, PhD - Stanford University Hospital; Jennifer Lee, MD - Stanford University;

Disparate Model Performance and Stability in Machine Learning Clinical Support for Diabetes and Heart Diseases

Presentation Time: 02:30 PM - 02:45 PM

Abstract Keywords: Fairness and Disparity Research in Health Informatics, Machine Learning, Generative AI, and Predictive Modeling, Clinical Decision Support for Translational/Data Science Interventions, Informatics Research/Biomedical Informatics Research Methods, Social Determinants of Health, Ethical, Legal, and Social Issues
Primary Track: Data Science/Artificial Intelligence
Programmatic Theme: Digital Health Technologies for Patient Research

Machine Learning (ML) algorithms are vital for supporting clinical decision-making in biomedical informatics. How-
ever, their predictive performance can vary across demographic groups, often due to the underrepresentation of his-
torically marginalized populations in training datasets. The investigation reveals widespread sex- and age-related
inequities in chronic disease datasets and their derived ML models. Thus, a novel analytical framework is introduced,
combining systematic arbitrariness with traditional metrics like accuracy and data complexity. The analysis of data
from over 25,000 individuals with chronic diseases revealed mild sex-related disparities, favoring predictive accuracy
for males, and significant age-related differences, with better accuracy for younger patients. Notably, older patients
showed inconsistent predictive accuracy across seven datasets, linked to higher data complexity and lower model per-
formance. This highlights that representativeness in training data alone does not guarantee equitable outcomes, and
model arbitrariness must be addressed before deploying models in clinical settings.
Speaker(s):
Ioannis Bilionis, MsC, Electrical and Computer Engineer
Adhera Health and Universitat Pompeu Fabra

Author(s):
Ioannis Bilionis, MsC, Electrical and Computer Engineer - Adhera Health and Universitat Pompeu Fabra; Ricardo C. Berrios, Economics and Marketing - Adhera Health; Luis Fernandez-Luque, Informatics - Adhera Health; Carlos Castillo, Computer Science - ICREA and Universitat Pompeu Fabra;

Generating Computable Phenotype Intersection Metadata Using the Phenoflow Library

Presentation Time: 02:45 PM - 03:00 PM

Abstract Keywords: EHR-based Phenotyping, Knowledge Representation, Management, or Engineering, Natural Language Processing
Primary Track: Clinical Research Informatics
Programmatic Theme: Harnessing the Power of Large Language Models in Health Data Science

Researchers must be supported in reusing computable phenotypes. This can be done by identifying intersections when multiple definitions exist for the same conditions (definition multiplicity). We analysed the standardised phenotypes present in the Phenoflow library to first understand the extent of definition multiplicity, before identifying the extent of intersections between definitions. We found that although definition multiplicity is extensive, many definitions still contain unique elements which can be shared with researchers as phenotype metadata.
Speaker(s):
Vasa Curcin, PhD
King's College London

Author(s):
Martin Chapman - King's College London; Luke Rasmussen, MS, FAMIA - Northwestern University; Jennifer Pacheco, MS - Northwestern University, Center for Genetic Medicine; Laura Wiley, PhD - University of Colorado; Vasa Curcin, PhD - King's College London;