Managing CI Fellowship Project Portfolios: Making Projects Count Thrice

Poster Number: P01
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Building Value for Informatics via Education and Training, Leadership Development for Informatics Trainees and Workforce, Clinical informatics organizational models
Primary Track: Leadership and Governance
Programmatic Theme: Organizational Challenges

As CI fellowship programs mature and accumulate projects through fellow involvement over time, continuity can be challenging due to graduation or transitions in ownership. It is difficult for program directors to manage project portfolios, ensuring that project milestones are met, that fellows are learning skills, that return on investment (ROI) is assessed, and that research is disseminated. We created a project inventory and assessment checklist to audit projects, rank priority level and align them with submission deadlines for conferences, internal grants and presentation to leadership. Through this process, we reduced the number of active projects from 19 to 8, submitted internal grants for 4, conference abstracts for 6, and identified 3 for ROI evaluation. We also standardized project organization to centrally manage key documents and deliverables. Project management principles can be used to organize CI Fellowship project portfolios, and optimize educational value, research dissemination, and organization impact.
Speaker(s):
Rachel Wong, MD, MPH, MBA, MS
Stony Brook University, School of Medicine

Author(s):

Barriers and Facilitators of Implementing Value-based Care: The Case of SwissDiabeter

Poster Number: P02
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Care Delivery Models, Change Management, Interoperability, Telemedicine and Telehealth including mHealth, App’s etc, Innovation in Digital Care
Primary Track: Leadership and Governance
Programmatic Theme: Organizational Challenges

Global diabetes care spending reached $966 billion in 2021, highlighting the need for cost-effective strategies like value-based care (VBC). This study explores implementing SwissDiabeter, a VBC-inspired diabetes clinic in Switzerland, informed by lessons from the Dutch Diabeter model. Based on 27 stakeholder interviews, key success factors include leadership, financial restructuring, operational improvements, and digital technologies. Recommendations include redesigning incentives, fostering partnerships, and tracking outcomes anonymously, offering insights for advancing VBC globally and beyond diabetes care.
Speaker(s):
Odile-Florence Giger, Master
HSG

Author(s):

Engaging Academic Medical Centers in Pediatric Digital Health: a Design Process and Model for Innovation

Poster Number: P03
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Driving Digital Equity, Innovation in Digital Care, Co-production/Co-Design, Diversity, Equity and Inclusion, Health IT Standards (USCDI, FHIR®, SMART, etc.), FDA Digital Health and Software as a Medical Device (SaMD)
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Emerging Technology and Technical Infrastructure

Pediatric digital health innovation lags behind adult-focused advancements, hindered by fragmented care systems and insufficient health economic incentives. Despite the long-term returns of investing in children's health, barriers in regulation, privacy, and market dynamics stymie progress. This project presents an educational, project-based approach, leveraging the Stanford Biodesign Process to address pediatric-specific needs. We will present the design process, the resulting digital products, and a framework for attendees to replicate these efforts at their institutions.
Speaker(s):
Aydin Zahedivash, MD, MBA
Stanford University

Author(s):
Paul Schmiedmayer, Postdoctoral Student - Stanford University; Vishnu Ravi, MD - Stanford University School of Medicine; Vasiliki Bikia, PhD - Stanford University; Oliver Aalami, MD - Stanford University School of Medicine, Surgery Dept.;

Exploring Telehealth Outcomes in a Federally Qualified Health Center

Poster Number: P04
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Telemedicine and Telehealth including mHealth, App’s etc, Workflow Efficiency, Big Data, Diversity, Equity and Inclusion
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Informatics-Driven Value-Based Healthcare

This descriptive study analyzes electronic health record data from a Federally Qualified Health Center to explore trends in emergency department and urgent care referrals at the end of telehealth visits. Using descriptive statistics, we examined 303,324 telehealth visits from July 1, 2023 - June 30, 2024, identifying trends in referrals on weekends. This data lays the groundwork for a retrospective cohort study to address health disparities and optimize telehealth services.
Speaker(s):
Efren Aguilar, BS
DGSOM/CDU/AltaMed

Author(s):
Efren Aguilar, BS - DGSOM/CDU/AltaMed; Christina Jung, MD, MPH - Children's Hospital Los Angeles; Eric Lee, MD, FAMIA - AltaMed Health Services; Jeffrey Arroyo, MD - Altamed Health Services;

Challenges of Implementing Social Determinants of Health Screening in a Low Vision Clinic

Poster Number: P05
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Social Determinants of Health, Algorithmic bias and impacts on Health Equity, Population Health, Diversity, Equity and Inclusion, Driving Digital Equity, EHR Implementation and Optimization
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Usability, Efficiency, and User Experience

Patients in a low vision clinic have unique challenges when completing online patient portal questionnaires. We illustrate these challenges with data from our implementation of social determinants of health questionnaires in a unique community ophthalmology clinic that primarily treats low vision patients and compare them to the same tools at a geographically similar primary care clinic.
Speaker(s):
Navin Pathak, MD
UW Health

Author(s):

Impact of a remote second opinion program on changes in recommended treatment plans for patients with cancer across the U.S.

Poster Number: P06
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Innovation in Digital Care, Patient Safety, Telemedicine and Telehealth including mHealth, App’s etc, Care Delivery Models
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Second opinion consultations can help to identify errors in cancer diagnoses and recommend the optimal course of treatment for cancer patients. Yet, patients may experience barriers to accessing these services. The objective of this study was to examine the extent to which a remote second opinion program at a high-volume comprehensive cancer center led to changes in diagnosis and treatment for cancer patients across the U.S.
Speaker(s):
Allison Lipitz-Snyderman, PhD
Memorial Sloan Kettering Cancer Center

Author(s):
Allison Lipitz-Snyderman, PhD - Memorial Sloan Kettering Cancer Center; SUSAN CHIMONAS, PhD - Memorial Sloan Kettering Cancer Center; Lauren Klein Levine, DrPH - Memorial Sloan Kettering Cancer Center; Brendan Raftery, MS - Memorial Sloan Kettering Cancer Center; Cole Manship, MPH - Memorial Sloan Kettering Cancer Center; Sergio Giralt, MD - Memorial Sloan Kettering Cancer Center; Tiffany Traina, MD - Memorial Sloan Kettering Cancer Center; Steven Sugarman, MD - Memorial Sloan Kettering Cancer Center; Benjamin Roman, MD MHHP - Memorial Sloan Kettering Cancer Center;

Advancing Healthcare Innovation: Insights from Key Stakeholders at the Cedars-Sinai Accelerator

Poster Number: P07
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Cross-organization Partnerships including Public-private Partnerships, Disruptive and Innovative Technologies, Cross Setting Collaboration
Primary Track: Industry and Commercial Partnership
Programmatic Theme: Organizational Challenges

Healthcare accelerators help advance innovation by connecting startups with key stakeholders within the health system. Limited descriptive studies exist covering this topic. We conducted a rapid qualitative analysis of eight focus group interviews with 14 key stakeholders across eight departments at Cedars-Sinai focusing on technology valuation, adoption, and integration. Three key themes emerged: investing in problem understanding, creating a balanced scorecard for value analysis, and leveraging early and regular stakeholder engagement. Our findings share insights into the complex dynamics between startups and health systems in advancing innovation.
Speaker(s):
Randy Tsai, BS, BA
David Geffen School of Medicine at UCLA

Author(s):
Jin Sol Lee, MD, MPH - David Geffen School of Medicine at UCLA/Harbor UCLA Medical Center/Cedars-Sinai Medical Center; Jim Laur, JD - Cedars-Sinai Medical Center; Nirdesh Gupta, PhD - Cedars-Sinai Medical Center; Matthew Sakumoto, MD - Sutter Health;

Assessing Documentation Burden after Centers for Medicare and Medicaid Services Billing Changes for Medical Student Notes

Poster Number: P08
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Documentation Burden, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Clinician Burnout, Usability and Measuring User Experience
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This study evaluates the impact of CMS billing changes that allowed medical student notes to be used as the billable clinical note. We analyzed documentation burden by comparing medical student note length and writing time before and after implementation and attending attestation length and writing time on medical student compared to resident notes after implementation. Results showed a significant increase in attested student notes after the change without additional documentation burden for attendings or students.
Speaker(s):
Catherine Blebea, MD
University of California San Francisco

Author(s):
Catherine Blebea, MD - University of California San Francisco; Aris Oates, MD - UCSF Health;

Quality Improvement in Insulin Prescribing at Hospital Discharge

Poster Number: P09
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Adaptive Clinical Decision Support, Workflow Efficiency, Quality Measures and eCQMs / Quality Improvement
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

At Stony Brook University Hospital, we developed an order set to provide embedded clinical decision support (CDS) into the EHR ordering workflow at the point of prescription to the pharmacy across the SBUH system. The order set provided clinicians information regarding insulin types and the corresponding medical supplies. After implementation and education, there was an absolute decreased number of insulin administration and hypoglycemia prevention errors and improved perceptions of insulin prescribing complexity.
Speaker(s):
Jonathan Ambut, MD
Stony Brook University Hospital

Author(s):
Jonathan Ambut, MD - Stony Brook University Hospital; Adam Slavick, MD - Stony Brook University; Alexandra Rusz, MD - Stony Brook Medicine; Brooke Learned, DO - Stony Brook University; Mishelle Centeno Gavica, MD - Stony Brook University; Julie Kim, MD - Stony Brook University; Apurva Shah, MD - Stony Brook University; Eileen Keck, BSc - Stony Brook University/SBMIT; Mathew Tharakan, MD - Stony Brook Medicine; Marina Charitou, MD - Stony Brook University; Alan Chang, MD - Stony Brook University; Rachel Wong, MD, MPH, MBA, MS - Stony Brook University, School of Medicine; Veena Lingam, MBBS - Moffitt Cancer Center-USF;

Charting the Course: Mapping Nursing Audit Logs to SNOMED CT

Poster Number: P10
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Clinician Burnout, Documentation Burden, Health IT Standards (USCDI, FHIR®, SMART, etc.)
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Electronic health record (EHR) audit log data capture nursing activity at the user level and have the potential to provide insights into nursing actions. However, there is little standardization in EHR audit logs and their correlation to nursing tasks has not been explored. We extracted nursing EHR interactions from the audit log metadata and mapped them to a standardized reference terminology, SNOMED CT, to promote the reuse of interoperable nursing data.
Speaker(s):
Victoria Tiase, PhD, RN, NI-BC, FAMIA, FAAN, FNAP
University of Utah

Author(s):
Katherine Sward, PhD - University of Utah; Julio Facelli, PhD - Facelli;

Standardizing Inpatient Psychiatric Documentation Templates Using a Combination of Generative AI and Clinician Consensus

Poster Number: P11
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Documentation Burden, Artificial Intelligence/Machine Learning, Workflow Efficiency
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This study addresses the challenge of voluminous and inconsistent inpatient psychiatric documentation post widespread EHR adoption across a large multi-center hospital system. Despite the independent development of documentation templates across each campus, we harnessed Generative AI to streamline and standardize templates. This involved merging templates from all campuses across different clinical scenarios, dissecting existing notes into common sections, and synthesizing an amalgamated note template using a secure, HIPAA-compliant version of ChatGPT 4o. The AI-generated templates were then collaboratively reviewed and revised by psychiatrists from each service. The standardized templates resulted in a reduction of word and character count. This AI-driven, collaborative process not only streamlined documentation workflows but also improved the overall quality and usability of psychiatric notes by ensuring clinical concensus. Our results underscore the potential of generative AI in enhancing clinical documentation standardization within large multi-center institutions.
Speaker(s):
Conner Polet, MD MBA
NYU Langone Health

Author(s):
Jonah Feldman, MD, FACP - NYU Langone Health;

Standardizing the Combat Trauma Registry: Insights and Lessons Learned

Poster Number: P12
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Big Data, Coordination Across the Continuum of Care, Data Governance, Data Science, Data Visualization, Learning Health System
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Problem: Having accurate and data is paramount for maximizing patient outcomes. The variety of data sources, formats and timing inconsistencies create a loss of granularity, misrepresentation of semantic meaning, and syntactic collision. A common data model to aggregate and standardize complex data from a variety of sources would negate these interoperability concerns and improve trauma outcomes.

Methods: The totality of the DODTR’s 150,000 records were summarized and exported into the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). Next a high-level mapping from the DODTR to standard OMOP CDM V5.4 tables was designed. Code mapping from source data to standard OMOP concepts was performed with 3 clinical informaticist.

Results: The DODTR has 9100 unique ICD 9 and 10 Diagnosis and Procedure terms. 4807 (53%) had a match score of 0.9 or greater. We reviewed initial mappings and adjudicated of 8193 (90%) terms. 90% of ICD10 and 71.9% of ICD9 unique terms were mapped, however, both were mapped to at least 98% based on frequency. Equivalency showed that 5317 (58%) of ICD10 and 5102 (69.6%) mapped to ‘Equal’. ICD9 had only 138 (1.9%) wider terms while ICD10 had 2792 (31%) indicating that the granularity of ICD10 codes was lost when moving to SNOMED. Quality assurance, imputation, and custom codes were identified for future refinement.

Conclusion: This is the first described mapping from a trauma registry to the OMOP CDM. Our frameworks provides a reliable, repeatable and generalizable strategy to convert a trauma registry into the OMOP CDM.
Speaker(s):
Darshan Thota, MD
United States Navy

Author(s):
Jonathan Stallings, PhD - Joint Trauma System; Michael Shiels, RN - Joint Trauma System; Jonpaul Trossi, RN - Joint Trauma System; Jennifer Gurney, MD - Defense Health Agency;

The Navigator’s Path: Charting a Course for Improved Substance Use Disorder Management Focused on Healthcare Effectiveness Data and Information Set (HEDIS) Measures

Poster Number: P13
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Workflow Efficiency, Clinical Process Automation, Documentation Burden
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This project introduces a systematic approach to enhance the management of substance use disorder patients in emergency departments through the Substance Use Navigator (SUN) initiative. By implementing structured electronic health record workflows and standardized documentation, we improved patient identification, tracking, and hospital reimbursement potential. The use of clinical decision support and data-tracking tools enhances patient care, with initial results indicating increased efficiency and reduced manual errors. Further research is needed to address patient follow-up complexities.
Speaker(s):
Roderick Eguilos, DO
University of California, Irvine

Author(s):
Graham Stephenson, MD - UC Irvine Medical Center; Monil Patel, MD - UC Irvine Health; Lindsey Spiegelman - UC Irvine Medical Center; Bharath Chakravarthy, MD, MBA - UC Irvine School of Medicine;

Utilizing Se(CER)t (MSG)s to Capture Discharge Diagnoses

Poster Number: P14
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Workflow Efficiency, Documentation Burden, Interprofessional Collaboration
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Clinical documentation queries are directly related to clinician burnout. Through collaboration with ancillary medical staff (Registered Dieticians, Physical and Occupational Therapists), we better captured relevant diagnoses (cachexia, physical debility diagnoses), and then utilized EPIC CERMSG smartlinks to automatically integrate them into the primary physician's discharge summary. By leveraging the expertise of specialized healthcare providers, we streamlined documentation workflows while minimizing disruption, ultimately improving the quality of clinical documentation and decreasing the burden on providers.
Speaker(s):
Thalia Nguyen, MD
UCLA Health

Author(s):
Thalia Nguyen, MD - UCLA Health;

Clinical Note Sectioning with Fine-Tuned LLMs: A Flexible, Institution-Agnostic Approach

Poster Number: P15
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Documentation Burden, Artificial Intelligence/Machine Learning, Workflow Efficiency
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Accurate sectioning of clinical notes is crucial for information extraction but often requires manual chart reviews. This study evaluated four large language models (LLMs) for identifying key sections (history of present illness, interval history, and assessment & plan) in oncology notes. Fine-tuned models like Llama 3.2 (3B) performed comparably to GPT-4o, achieving high precision (0.72) and recall (0.79). Results demonstrate fine-tuned LLMs as a scalable, data-efficient solution for robust note sectioning.
Speaker(s):
Joshua Davis, BS
Dana-Farber Cancer Institute

Author(s):
Joshua Davis, BS - Dana-Farber Cancer Institute; Thomas Sounack, MS - Dana-Farber Cancer Institute; Kate Sciacca, NP - Dana-Farber Cancer Institute; Jessie Brain, CNP - Dana-Farber Cancer Institute; Nicole Agaronnik, BS - Dana-Farber Cancer Institute; Charlotta Lindvall, MD PhD - Dana-Farber Cancer Institute;

IFRLex: an Inpatient Fall-Reporting Lexicon for Identifying Electronic Self-Reporting Information from Clinical Notes

Poster Number: P16
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Patient Safety, Clinical Process Automation, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Data Science, Artificial Intelligence/Machine Learning, Documentation Burden
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Recognizing that clinical events occurring during hospitalization are documented in detail as text in nursing notes and relevant contextual information, we developed an inpatient fall-reporting lexicon (IFRLex) based on NLP to identify incident elements required for reporting falls electronically. This approach can reduce the reporting burden on nurses and streamline the self-reporting process. A data-driven preliminary evaluation of the IFRLex was also conducted.
Speaker(s):
INSOOK CHO, PhD
Inha University

Author(s):
Hyunchul Park, MBA - aSSIST University; Byeong Sun Park, MS - Inha University; Hyekeyong Shin, MS - Inha University;

Comparing pre-visit anxiety & depression screener delivery methods and enhancing EHR-REDCap integration for quality improvement & research

Poster Number: P17
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Patient-Generated Data / Patient Reported Outcomes (PROs), Learning Health System, Consumer and Patient Engagement
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Given time limitations for clinicians to conduct recommended mental health screening and inefficiency of manual data collection by clinicians for quality improvement or research, we tested text versus email messaging to encourage patient pre-visit completion of mental health screening and developed a REDCap Kit API module to facilitate screener result data collection from Epic EHR flowsheets. Text messages had higher response rates than email, and the REDCap module improved accuracy and efficiency of data collection.
Speaker(s):
Heidi Munger Clary
Wake Forest University School of Medicine

Author(s):
Heidi Munger Clary - Wake Forest University School of Medicine; Beverly Snively, PhD - Wake Forest University School of Medicine; Cody Hudson, MS - Wake Forest University School of Medicine; Joseph Criscitiello, BA - Wake Forest University School of Medicine; Umit Topaloglu, PhD - National Cancer Institute;

Description and Preliminary Outcomes from a Centers of Excellence Model to Enable Pharmacist-led Health Services on a Nationwide Scale

Poster Number: P18
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Workflow Efficiency, Coordination Across the Continuum of Care, Medication Adherence
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This study highlights the development and implementation of an informatics-enabled Centers of Excellence workflow model within retail pharmacies across the United States. First use cases resulting from the program focus on expanding pharmacist-led medication reviews and pharmacy technician-led telephonic outreach to improve patient care and address gaps in care. Early outcomes demonstrate high patient engagement, closure of care gaps, and increased provider satisfaction, advancing healthcare equity nationwide.
Speaker(s):
Eleanor Beltz, PhD, ATC
Aetna Medical Affairs, CVS Health

Author(s):
Amanda Zaleski, PhD, MS - Aetna; Wendy Zhao, MPA, MBA - CVS Health; Lindsay Marchand, PharmD - CVS Health; Kelly Jean Craig, PhD - CVS Health; Eleanor Beltz, PhD, ATC - Aetna Medical Affairs, CVS Health; Jinali Desai, PharmD, MPH, MBA - CVS Health;

Usability Assessment of the Rooming Navigator using Epic’s Workflow Analyzer Tool

Poster Number: P19
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Usability and Measuring User Experience, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, EHR Implementation and Optimization, Clinician Burnout, Workflow Efficiency, Human Factors Testing
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This study evaluates the usability of our EHR’s Rooming Navigator using Keystroke Level Modeling (KLM) and the System Usability Scale (SUS) to establish baseline metrics for redesign. Thirty-five users participated, with results highlighting task durations, screen transitions, and mouse/keyboard actions. High SUS scores indicated general satisfaction, with minimal correlation to task time or experience. Findings establish a foundation for optimizing workflow efficiency, demonstrating KLM and SUS as essential tools for usability assessment in healthcare.
Speaker(s):
matthew nudelman, md
UW Health

Author(s):

Preventing Surgical Delays: Leveraging Decision Support and Pharmacy Dispensation Data for GLP-1 and SGLT2

Poster Number: P20
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Medication Adherence, Patient Safety, Innovation in Digital Care
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Patients on GLP-1 receptor agonists and SGLT2 inhibitors require specific pre-procedural instructions for GI procedures, including discontinuing these medications within a designated timeframe. Failure to follow these instructions often leads to procedure cancellations, disrupting care and clinical workflows. To address this issue, we developed a clinical decision support (CDS) tool that integrates pharmacy dispense data with procedural scheduling systems to identify patients on these medications and provide timely alerts for clinicians and care teams.

In a six-month pilot program, the CDS tool reduced procedure cancellations related to medication management errors by 75%. It identified over 90% of patients requiring intervention and ensured timely delivery of pre-procedural instructions, leading to high clinician compliance and improved patient understanding. These results highlight the value of leveraging pharmacy dispense data in CDS systems to address specific procedural challenges, improve care delivery, and enhance patient outcomes.

This session will demonstrate the development, implementation, and outcomes of the CDS tool, providing attendees with actionable insights to design similar interventions in their clinical settings. The project underscores the importance of data-driven solutions in optimizing procedural workflows and preventing care disruptions.
Speaker(s):
Sri Harsha Palakurty, MD
Cedars Siani

Author(s):
Gabriel Labbad, MD - Cedars-Sinai Health System; Dave Frattinger, BS - Epic;

Tracking Episodes of Care to Review Population Health Analytics and Track Unmet Needs for Behavioral Health Patients

Poster Number: P21
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Data Visualization, Population Health, Workflow Efficiency, Documentation Burden, Diversity, Equity and Inclusion, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, 21st Century Cures (including considerations for special populations such as adolescents), Big Data
Working Group: Clinical Informatics Systems Working Group
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Leadership, Advocacy and Policy

Children’s Healthcare of Atlanta’s Behavioral and Mental Health Center identified a gap in the ability to collect discrete data. Development of an Epic SmartBlock to create episodes of care allowed for a better understanding of the evidenced-based treatments providers are recommending, whether patient’s needs are being met or why not. Data analysis will be used for informed decisions of investments, personnel hiring, and legislative lobbying against mental health parity violations in Georgia.
Speaker(s):
Kayla Mays, DNP
Children's Healthcare of Atlanta

Author(s):
Naveen Muthu, MD - Children's Healthcare of Atlanta; Katie Daniel, DNP - Children's Healthcare of Atlanta;

Adoption of the OMOP CDM for Cancer Research using Real-world Data: Current Status and Opportunities

Poster Number: P22
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Big Data, Population Health, Data Privacy and Secondary Use
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

The Observational Medical Outcomes Partnership (OMOP) common data model (CDM) that is developed and maintained by the Observational Health Data Sciences and Informatics (OHDSI) community supports large scale cancer research by enabling distributed network analysis. As the number of studies using the OMOP CDM for cancer research increases, there is a growing need for an overview of the scope of cancer research that relies on the OMOP CDM ecosystem. In this study, we present a scoping review of the adoption of the OMOP CDM for cancer research, aiming to identify 1) the extent of cancer data analysis utilizing the OHDSI/OMOP CDM, 2) the current development of OHDSI/OMOP CDM as an ecosystem infrastructure for cancer research, and 3) challenges and opportunities from the above two themes for potential future investigations. In total, 49 unique articles were included, with 30 for the data analysis theme and 20 for the infrastructure theme, with 1 article belonging to both themes. This review revealed that while the OMOP CDM ecosystem has undergone critical development that is sufficient to support cancer research, ongoing model development and iteration remains needed to fulfill additional research data needs. Expanding research topics, underscoring rare cancer, integrating more diverse types of data sources, improving data quality, adopting advanced analytics methodology, including in-depth phenotypic data through NLP, and increasing multisite evaluations present important opportunities for future research to facilitate secondary usage of observational data.
Speaker(s):
Liwei Wang, MD, PhD
UTHealth

Author(s):

Implementing Ambient AI Scribes for Primary Care Provider Wellbeing at a Federally Qualified Health Center

Poster Number: P23
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Ambient documentation, Change Management, Clinician Burnout
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Organizational Challenges

To explore optimal implementation techniques for Ambient Artificial Intelligence (AI) Scribes in a real-world Federally Qualified Health Center (FQHC), we piloted Nabla Copilot with 32 providers. The largest barrier to adoption was simply completing the initial platform sign-in/setup and customization. Future implementations should include hands-on platform setup training, in the live AI Scribe environment. Note quality was high and acceptable to providers.
Speaker(s):
Margaux Benoit, MSc
Nabla

Author(s):
Matthew Malek, MD MPH - Thundermist Health Center; Margaux Benoit, MSc - Nabla;

Multidisciplinary Creation of Best Practice Alert to Reduce Excessively High Stool Output in the NICU

Poster Number: P24
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Patient Safety, Clinician Burnout, Interprofessional Collaboration
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

This poster highlights the successful implementation of a BPA in a NICU to improve the detection of high stool output episodes. Readers will learn strategies for multidisciplinary collaboration to address patient safety concerns, develop evidence-based thresholds, and optimize CDS tools. The case study demonstrates a 60% reduction in high stool output events and emphasizes balancing alert effectiveness with minimizing fatigue. Participants will leave equipped to design, test, and implement CDS interventions that enhance patient outcomes.
Speaker(s):
Alexa Gilman, MSN, APRN, NNP-BC


Author(s):
Agata Nytko, BA - Ann & Robert H. Lurie Children's Hospital of Chicago; Gustave Falciglia, MD - Northwestern University; Roderick Jones, MPH - Ann & Robert H. Lurie Children’s Hospital of Chicago; Abhineet Sharma, MD - University of Nebraska Medical Center; Barbara Fleming, MSN - Ann & Robert H. Lurie Children's Hospital of Chicago; Stephanie Jones, MSN - Ann & Robert H. Lurie Children's Hospital of Chicago; Kalee Ryan, Kalee Ryan, MBA, MSN, RN - Ann & Robert H. Lurie Children's Hospital of Chicago; Naomi Sullivan, Med, MS - Ann & Robert H. Lurie Children's Hospital of Chicago;

Reducing Interruptions from Non-Urgent Secure Texts in the Hospital

Poster Number: P25
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Clinician Burnout, Communication Strategies, Quality Measures and eCQMs / Quality Improvement, Workflow Efficiency
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This single-site quality improvement study aimed to reduce interruptions and alert fatigue in hospitalists from non-urgent messages by leveraging native functionalities in the secure texting platform, including “Busy” presence and “Priority” messages. Results from a 6-week pilot demonstrated significant adoption of the intervention and increase in silently delivered messages. Pilot users reported a significant reduction in hourly interruptions.
Speaker(s):
April Liang, MD
Stanford University

Author(s):
April Liang, MD - Stanford University; Stephen Ma, MD, PhD - Stanford University School of Medicine; Brian Phillips, RN, MSN - Stanford Healthcare; Stav Cullum, MD; Mark Keroles, MD - Stanford University; Ginger Yang, MD - Stanford Healthcare Tri-Valley; WEIHAN Chu, MD - Stanford Health Care; David Svec, MD, MBA - Stanford Health Care Tri-Valley; Lisa Shieh, MD, PhD - Stanford University Department of Medicine; Christopher Sharp, MD - Stanford University School of Medicine;

Pilot Study of Ambient Documentation Technology on Provider Time Allocation and Patient-Provider Interactions

Poster Number: P26
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Ambient documentation, Workflow Efficiency, Documentation Burden, Clinician Burnout
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Artificial Intelligence (AI) scribes are being widely implemented across the US, however the implications on hospital medicine providers' workflows and patient interactions remain unclear. This pilot time motion study shows that median documentation time of female hospitalists decreased with AI scribe use, whereas there was an increase for male hospitalists. Direct patient care time did not change, however provider eye contact time with patients increased, suggesting improved quality of patient-provider interactions.
Speaker(s):
Sharmila Tilak, MD
Brigham & Women's Hospital

Author(s):

Addressing Incomplete Inhalant IgE Panels Through User-Centered Design

Poster Number: P27
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Usability and Measuring User Experience, Quality Measures and eCQMs / Quality Improvement, Workflow Efficiency
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Incomplete inhalant IgE panels can lead to treatment failures in allergen immunotherapy (AIT). A user-centered SmartSet was developed in the electronic health record (EHR) to guide clinicians in identifying and supplementing missing test components based on the payor-assigned laboratory. Usability testing with allergy clinicians revealed positive feedback, emphasizing workflow integration and tailored communication strategies. This intervention holds promise for improving diagnostic completeness and AIT outcomes, with further evaluation planned to assess real-world impact.
Speaker(s):
Idil Ezhuthachan, MD, MS
Children's Healthcare of Atlanta/Emory University

Author(s):
Alexis Carter, MD - Children's Healthcare of Atlanta; Swaminathan Kandaswamy, PhD - Emory University School of Medicine;

Optimizing Pre-hospital Blood Transfusion in Austere Environments: A Validation Study of Machine Learning Models

Poster Number: P28
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Care Delivery Models, Data Science, Data Visualization, Disruptive and Innovative Technologies, Learning Health System
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Problem: Hemorrhagic shock is the number one cause of preventable death on the battlefield and requires aggressive treatment with blood products. Determining the optimal thresholds and variables for transfusion in needed to maximize outcomes. In this study, we will validate the accuracy of previously studied civilian machine learning prediction models to find the optimal sensitivity and specificity pre-hospital transfusion variables for OPMED.
Methods: We replicated a study by Zadorozny et al using a Fast Frugal Tree and additional machine learning models to predict blood transfusion in the pre-hospital operational medicine setting. We compared an ensemble of machine learning techniques using a fast frugal trees (FFT) to generate a clinical decision support tool using optimal transfusion cut-offs for systolic blood pressure, pre-hospital lactate, shock index, and Abbreviated Injury Score. We will compare CART, LR, RF and SVM using area under the curve and receiver operator curves and used Youden’s J index to asses accuracy.
Results: Preliminary results showed only a subset of Zadorozny’s 29 variables are captured in our registry. Our final data set resulted in choosing transfused blood, lowest systolic blood pressure, lowest shock index, lowest diastolic blood pressure, heart rate, and injury types. We found a similar level of performance using logistic regression compared to (FFT) and comparison model performance is currently underway.
Conclusion: Given the precarious nature of the deployed environment, accurate mission planning for material solutions such as blood products is critical to maximizing patient outcomes.
Preliminary results show utility in using ML models for transfusion prediction.
Speaker(s):
Darshan Thota, MD
United States Navy

Author(s):
Jonathan Stallings, PhD - Joint Trauma System;

Addressing Organ Procurement Organization (OPO) lack of access to Electronic Health Records (EHRs)

Poster Number: P29
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: EHR Implementation and Optimization, Documentation Burden, Patient Safety, Workflow Efficiency, Usability and Measuring User Experience, Interprofessional Collaboration, Clinician Burnout
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Organizational Challenges

This initiative reduces delays and supports Organ Procurement Organizations (OPOs) by implementing a standardized protocol incorporating simplified Electronic Health Record (EHR) access for regional OPO, streamlined workflow for organ referrals and procurements and integrated documentation and ordering practices. These improvements address EHR-related challenges that hinder OPOs from fulfilling federally mandated obligations and enhance donor identification, timely referrals, and coordination between hospital and OPO teams, ultimately saving clinician time and supporting effective donor organ procurement.
Speaker(s):
Selina Chen, MD, MPH, DipABLM
HPH

Author(s):

Considerations for Implementation of Automatic System-generated Endocrinology Consultations for Hospitalized Patients with Insulin Pumps

Poster Number: P30
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Clinical Process Automation, Patient Safety, Artificial Intelligence/Machine Learning
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

The American Diabetes Association states automated insulin delivery systems should be continued and supported during hospitalization, when clinically appropriate and with proper supervision and recommends consultation of endocrinology for these patients. We are planning to implement Epic’s Instant Orders functionality to ensure hospitalized patients with insulin pumps more reliably get consultations to endocrinology. Interviews with stakeholders have revealed a need for closed loop communication and accuracy of identification methods are important considerations prior to implementation.
Speaker(s):
Sarah Stern, MD
Vanderbilt University Medical Center

Author(s):
Dara Mize, MD, MS - Vanderbilt University Medical Center; Marc Maldaver, MD - Vanderbilt University Medical Center;

Implementation of a Clinical Decision Support Tool to Optimize Adherence to Heart Failure Clinical Guidelines Within a Network of Value-Based Primary Care Clinics

Poster Number: P31
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Population Health, Care Delivery Models, Adaptive Clinical Decision Support
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Heart failure is a condition affecting more than 6 million individuals nationwide. Clinical guidelines articulate best practices in heart failure management; however, opportunities to improve disease management and guideline adherence exist. Clinical decision support (CDS) serves as useful tool for enhancing health-related decisions to standardize care. Investigators present how a tailored CDS tool specific to HF disease management may offer opportunity to address current care gaps in the case of congestive heart failure.
Speaker(s):
Michael Chen, MD
Oak Street Health

Author(s):
Rebecca Anastos-Wallen, MD - Oak Street Health/CVS Health; Katelyn Hilliker, BS - Oak Street Health; Kelsey Cooper Kelsey Cooper, MBA - Oak Street Health/CVS Health; Arti Panchal, MPA - Oak Street Health/CVS Health; Roward Agulto, BS - Oak Street Health/CVS Health; Michael Chen, MD, MBA - Oak Street Health/CVS Health; Julien Bendelac, BA - Oak Street Health/CVS Health; Amanda Zaleski, PhD, MS - Aetna; Kelly Jean Craig, PhD - CVS Health; Ashley Chou, MD, MAPP - Oak Street Health/CVS Health;

Cancer Patients' Messages About Radiology/Pathology Reports: Insights for AI

Poster Number: P32
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Communication Strategies, 21st Century Cures (including considerations for special populations such as adolescents)
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Cancer patients use portals to view reports prior to discussing with physicians, leading to diverse concerns. Analyzing 1 week of messages about pathology/radiology reports viewed within 6 hours of posting uncovered requests for interpretations (24/48, 50%) and implications (14/48, 29%); expressions of concern (5/48, 10%) or relief (3/48, 6%), and perceptions of errors/omissions (3/48, 6%). AI could categorize and triage message patterns to help meet patients' needs pending conversations with providers.
Speaker(s):
SUSAN CHIMONAS, PhD
Memorial Sloan Kettering Cancer Center

Author(s):
Allison Lipitz-Snyderman, PhD - Memorial Sloan Kettering Cancer Center; SUSAN CHIMONAS, PhD - Memorial Sloan Kettering Cancer Center; Gilad Kuperman, MD, PhD - Memorial Sloan Kettering Cancer Center;

Quality Improvement Initiative to Improve Incidental Lung Nodule Detection using Large Language Models

Poster Number: P33
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Workflow Efficiency, Quality Measures and eCQMs / Quality Improvement
Primary Track: AI and Care Outcomes
Programmatic Theme: Organizational Challenges

At Stony Brook University Hospital, we created an Large Language Model data pipeline for lung nodule detection of radiographic reports, with the goal of developing an incidental lung nodule (ILN) detection clinical workflow over a week time period. After implementation, we found that 120/250 (48%) were not known to the Lung Cancer Evaluation Center that manage ILN. We determined that clinical and staff resources may require expansion to support additional patient volumes.
Speaker(s):
Jonathan Ambut, MD
Stony Brook University Hospital

Author(s):
Karthik Bharadwaj, MD - Stony Brook University; Dipika Rana, MD - Stony Brook Medicine; Neil Patel, MD, MBA - Stony Brook University; Janos Hajagos, Ph.D. - Stony Brook University; Aishwarya Sinhasane, MS - Stony Brook University; Rachel Wong, MD, MPH, MBA, MS - Stony Brook University, School of Medicine; Mary Saltz, MD - Stony Brook Medicine; Tahsin Kurc; Anish Desai, MD - Stony Brook University; Ankit Dhamija, MD - Stony Brook University; Carolyn Franson, Patient Navigator - Stony Brook University; Denise Albano, DNP - Stony Brook University; Mathew Tharakan, MD - Stony Brook Medicine; Sahar Ahmad, MD - Stony Brook University; Joel Saltz, MD, PhD - Stony Brook University Health Sciences Center; Veena Lingam, MBBS - Moffitt Cancer Center-USF; Jonathan Ambut, MD - Stony Brook University Hospital;

The Association Between Graves' Disease and Risk of Preterm Births: A Predictive Modeling Approach Using Logistic Regression and Random Forests

Poster Number: P34
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Data Science, Artificial Intelligence/Machine Learning, EHR Implementation and Optimization
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Graves' disease significantly increases the risk of preterm birth. Using data from the All of Us research database, we developed predictive models to assess risk factors, including logistic regression and Random Forests. A hybrid model combining both approaches demonstrated improved recall and precision. This work highlights the importance of early identification, education, and targeted screening to manage preterm birth risks in women with Graves' disease.
Speaker(s):
Judith Dike, BS
Meharry Medical College

Author(s):

Automated Classification and Segmentation of Mucus Morphology in Nasal Endoscopy Using YOLOv11 for Enhanced Clinical Decision-Making

Poster Number: P35
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Disruptive and Innovative Technologies, Bridging Analytics, Bedside Care, Clinical Documentation, and Education
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

This study evaluates YOLOv11 for segmenting and classifying mucus morphology (crusts, strands, globs) in nasal endoscopy to reduce diagnostic variability. Results show promising classification accuracy (F1 scores: 57.9% for crusts, 46.8% for strands, 43.8% for globs), marking progress toward AI tools that support non-specialists. This foundation aims to improve access to consistent sinonasal assessments, enhancing early diagnosis and treatment in primary care, and helping clinicians make more reliable, timely decisions.
Speaker(s):
Dipesh Gyawali, M.S. in Computer Science
Ochsner Health

Author(s):
Dipesh Gyawali, M.S. in Computer Science - Ochsner Health; Akio Fujiwara, MD - Ochsner Health; Edward D McCoul, MD, MPH - Ochsner Health; Jonathan Bidwell, PhD - Ochsner Health;

Improving Structural Heart Outcomes Utilizing Episode-based Tracking and Tools

Poster Number: P36
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Workflow Efficiency, Care Delivery Models, Clinical Process Automation
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Structural heart patients have complex pre-operative, post-operative, and longitudinal follow up care coordination, national registry reporting, and research requirements. Coordinating and ensuring completion of the studies, appointments, and required metrics is complex and often requires coordination amongst physicians, advanced practice providers, patient care representatives, and in some cases researchers frequently across multiple departments. We developed and implemented an episode-based tracking tool for structural heart procedures with improvement in completion rates of key tasks.
Speaker(s):
Andrew Bullock, MD
Cedars Sinai Medical Center

Author(s):

Supporting Healthcare Planning and Resource Management for Gastritis and Duodenitis Using Large Language Models

Poster Number: P37
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Adaptive Clinical Decision Support, Innovation in Digital Care, Telemedicine and Telehealth including mHealth, App’s etc, Remote Patient Monitoring, Population Health
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

This study examines the use of Large Language Models (LLMs) in healthcare planning and resource management for gastritis and duodenitis (GD) in remote and low-resource settings. Five LLMs generated tailored strategies, evaluated by healthcare professionals on relevance, plausibility, and practicality. Findings suggest that LLMs offer complementary strengths across telemedicine and rural care contexts, enhancing clinical decision-making and public health support in underserved regions.
Speaker(s):
Zhiyi Sun, M.S.
University of Michigan

Author(s):
Zhiyi Sun, M.S. - University of Michigan; Yupei Liu, MD - Renmin Hospital of Wuhan University; Zelin Wang, MS - University of Arizona; Juan Zhou, PhD - The Affiliated Children’s Hospital of Xiangya School of Medicine, Central South University (Hunan Children’s Hospital); Zheng Ma, MA - Harvard University; Boping Duan, BSN - The Affiliated Children's Hospital of Xiangya School of Medicine, Central South University (Hunan Children's Hospital); Hongmei Zhao, MD - The Affiliated Children's Hospital of Xiangya School of Medicine, Central South University (Hunan Children's Hospital);

Factors Related to Smart Health Devices Use Among Minority Older Adults

Poster Number: P38
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Remote Patient Monitoring, Driving Digital Equity, Innovation in Digital Care, Usability and Measuring User Experience, Telemedicine and Telehealth including mHealth, App’s etc, Connected Care in the Home
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Smart health devices (SHD) are increasingly used among older adults. But less than 4% of SHD users are African American and Latinx. We identify mutable factors, including predisposing, enabling, and need factors, that impact the use of SMD among minority older adults through conducting an exploratory interview study. The results highlight the need for an accessible, supportive, and equitable approach to adopt SHD to fulfill its promise in meeting the needs of minority older adults.
Speaker(s):
Jany Sun, BS
Rush Medical College

Author(s):
Yanjun Dong, MA - University at Albany, State University of New York; Jimmie Boliboun - Rush University Medical Center; Katherine Koo; Valeria Vazquez-Trejo, BS - Rush Medical College; Michael Cui, MD - Rush University Medical Center; Victoria Rizzo, Ph.D., LCSW-R, FNAP, FNYAM - University at Albany, State University of New York; Jeannine Rowe, PhD, MSW - University of Wisconsin-Whitewater;

Note Quality Between Two Different Ambient Scribe Vendors and ChatGPT

Poster Number: P39
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Ambient documentation, Usability and Measuring User Experience, Documentation Burden, Disruptive and Innovative Technologies
Primary Track: AI and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

This study systematically compares notes from two different vendors of AI-assisted scribes, highlighting differences in clinical accuracy, readability, and overall quality. Using simulated patient-doctor interactions based on standardized clinical cases, cosine similarity found the notes were mostly similar between vendors, but with significant differences in hallucinations and being succinct.The study found that modified PDQI-9 scoring and cosine analysis are useful methods in comparing notes generated by AI-assisted scribes.
Speaker(s):
Nina Zhu, MD
UCLA

Author(s):
Nina Zhu, MD - UCLA; AAron Chin; Thalia Nguyen, MD - UCLA Health; Kevin Truong, MD, MBA - UCLA; Eric Cheng, MD, MS - UCLA;

Leveraging Digital Platforms to Monitor and Nudge Compliance with Clinical Practice Guidelines

Poster Number: P40
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Building Value for Informatics via Education and Training, Learning Health System, Care Delivery Models, Change Management, Adaptive Clinical Decision Support, Usability and Measuring User Experience
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

The Children’s Hospital of Philadelphia Newborn Care Network (CNBCN) has created evidence and consensus-based guidelines to standardize care provided in network NICUs. Optimal practice adherence to these guidelines is predicated on the provider’s knowledge of the guidelines’ content, their accessibility, and their quality. Our team set out to create and test the feasibility and impact of a novel mobile technology tool called NUDGE (Nurturing Uptake of Data Driven Guidelines using E-Solutions) in supporting guideline adherence.
Speaker(s):
John Chuo, MD, Master in Clinical Informatics, IA
University of Pennsylvania, Children’s Hospital of Philadelphia

Author(s):
Jacob DeMarino, MHA - Children's Hospital of Philadelphia; Crystal Bass, MD - Children's Hospital of Philadelphia; Melissa Schmatz, MD, MHQS - Children's Hospital of Philadelphia; Sharath Chowdawarapu, MD - Children's Hospital of Philadelphia; Purvi Kapadia-Jethva, MD - Children's Hospital of Philadelphia; Alex Ruan, MD - Children's Hospital of Philadelphia;

Description and Real-World Impacts of an Informatics-Enabled Population Health Program in Medicare Advantage Members with Gaps in Care

Poster Number: P41
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Population Health, Big Data, D2C and B2C Strategies
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

We describe outcomes from a proprietary, payor-agnostic, population health management program enabled by a clinical decision support system tool. Building on a rich and diverse data foundation, transdisciplinary expertise, tools, and capabilities are operationalized to effectively deliver near real-time, actionable, contextualized, personalized evidence-based outreach to end users, at scale. Future research will quantify the impact of tailored co-interventions on individual and population health outcomes.
Speaker(s):
Amanda Zaleski, PhD, MS
Aetna

Author(s):
Amanda Zaleski, PhD, MS - Aetna; Raj Patel, MD - CVS Healthspire Payor Solutions/CVS Health; Minming Li, PhD - CVS Healthspire Payor Solutions/CVS Health; Aurel Iuga, MD - CVS Healthspire Payor Solutions/CVS Health; Kelly Jean Craig, PhD - CVS Health; Eleanor Beltz, PhD, ATC - Aetna Medical Affairs, CVS Health; Dorothea Verbrugge, MD - CVS Health; Lia Rodriguez, MD - CVS Healthspire Payor Solutions;

Insights From Validation of EPIC’s New Sepsis Alert Model

Poster Number: P42
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Workflow Efficiency, Adaptive Clinical Decision Support, Patient Safety, Quality Measures and eCQMs / Quality Improvement
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

This study validates EPIC’s new Sepsis Alert Model (V2) against the original model (V1), focusing on reducing alert fatigue and clinical workflow impact. A retrospective analysis of 9784 non-sepsis and 660 sepsis encounters revealed significant reduction in positive alert rates with V2 (3.3% to 1.2% for non-sepsis and 26.1% to 6.8% for sepsis encounters). Spearman correlation and chi-square tests confirmed statistical significance. Results highlight improved alert precision, though education on clinical implications is essential.
Speaker(s):
Awais Farooq, MD
University of Illinois at Chicago

Author(s):

Evaluating AI Models for Generating FHIR Proteomic Test Cases

Poster Number: P43
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Health IT Standards (USCDI, FHIR®, SMART, etc.), Artificial Intelligence/Machine Learning, Precision Health and Genomics, Interoperability
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

AI models—ChatGPT, Claude, LLaMA, and Gemini—were evaluated for their ability to generate FHIR test cases for proteomic data representation, focusing on MolecularSequence and its proposed successor, MolecularDefinition. Outputs were assessed for FHIR compliance, inclusion of key attributes, and adherence to the Genomics Reporting Implementation Guide. While models effectively captured essential elements like mutation type and observed sequences, none achieved full compliance, revealing gaps in their ability to handle evolving FHIR standards. These findings highlight the importance of refining AI models and collaborating with the HL7 Clinical Genomics Work Group to advance reliable proteomic data representation.
Speaker(s):
Tia Pope, Ph.D. Candidate
North Carolina A&T State University

Author(s):
Ahmad Patooghy, PhD - North Carolina A&T State University;

Variability in Probabilistic Word Interpretation in Large Language Models

Poster Number: P44
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Disruptive and Innovative Technologies, Innovation in Digital Care
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

We examine how large language models (LLMs) interpret probabilistic terms (e.g., ‘likely,’ ‘possibly’), highlighting variability that may impact clinical communication. We presented four LLMs with common probabilistic words and asked each to convert to numerical interpretations. We found that terms at probability extremes were most consistent, while others varied widely, but variation within LLMs also differed. Results underscore the importance of precision in probabilistic language to prevent miscommunication in clinical settings.
Speaker(s):
Juan Chaparro, MD, MS
Nationwide Children's Hospital

Author(s):
Juan Chaparro, MD, MS - Nationwide Children's Hospital; Naveed Farrukh, MD, MPH - The Ohio State University and Nationwide Childrens;

Easing the Burden: Comparative Pilot of Ambient Artificial Intelligence Technology at a Large Academic Health System

Poster Number: P45
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Ambient documentation, Clinician Burnout, Change Management, Disruptive and Innovative Technologies
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

Extensive documentation contributes to clinician burnout and reduced patient face-to-face time. This study compares two ambient AI tools, Vendor A and Vendor B, assessing their efficiency, usability, accuracy, and clinician satisfaction across various care settings. Results showed both tools reduced documentation time, with Vendor A offering greater savings in complex cases. Vendor B’s multilingual support adds further benefits. Continued deployment and assessment will guide decisions on enterprise-wide implementation at a large academic health system.
Speaker(s):
Monil Patel, MD
UC Irvine Health

Author(s):
Roderick Eguilos, DO - University of California, Irvine School of Medicine; Danielle Perret, MD - UCI Health; Steven Tam, MD - UC Irvine; Charles Gillman, PhD Student - UCI Health;

Building a Secure Messaging Urgency Classification Model and List of Best Practices with GPT4, Nurses, and Doctors

Poster Number: P46
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Patient Safety, Learning Health System
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

This session addresses the misuse of secure messaging for emergent clinical content, a growing concern in healthcare due to the lack of objective gold-standard labels. GPT-4, doctors, and nurses co-developed a set of themes for messages considered “Emergent,” “Urgent,” or “Non-Urgent”, which were used to develop and refine a GPT-4 prompt for a high-performing classification model. In the future, this model may improve the appropriate use of secure messaging technology through real-time interventions.
Speaker(s):
William Small, MD, MBA
NYU Langone Health

Author(s):
Lisa Groom, PhD, RN - NYU; John Will, MPA - NYU Langone Health; Jonah Feldman, MD, FACP - NYU Langone Health; Paul Testa, MD, JD, MPH - NYU Langone Health; Jonathan Austrian, MD - NYU Langone Medical Center; Paawan Punjabi, MD, MSc - New York University School of Medicine/NYU Langone Health System; Christopher Sonne, MD - NYULMC; Priyanka Solanki, MD - NYU; Jared Silberlust, MD MPH - NYU Langone Health; Lucille Fenelon, MSN, MHA - New York University Health Center; Jacob Martin - NYU Langone Health; Vincenza Coughlin, RN MS - NYU Langone Health; Lauren Moran, RN - NYU Langone Health; Adriana Gutierrez, RN - NYU Langone Health; Nailah Moore, RN - NYU Langone Health; Yindalon Aphinyanaphongs, MD - NYU Langone Health; Clarine Long, MD - NYU Langone Health;

User intent modeling and analysis of secure large language model prompt logs at an academic children’s hospital

Poster Number: P47
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Clinician Burnout, Workflow Efficiency, Usability and Measuring User Experience
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

Secure versions of ChatGPT in healthcare are becoming more prevalent, including at institutions such as the University of Washington, Vanderbilt University, Stanford University, and Harvard University. However, understanding how healthcare personnel interact with a patient health information (PHI)-secure large language model (LLM) tool remains a gap. This study investigates these usage behaviors. The study analyzed the usage logs of a secure ChatGPT API, AskDigi, used by healthcare staff and faculty. The dataset contained 3,154 threads from 04/22/2024 to 09/13/2024. An initial code set based on Bedi et al. was applied to 100 messages by human reviewers, who then performed consensus coding to refine the codebook. This refined codebook was used for both automated and manual labeling. Automatic tagging of free-text prompts was performed using AskDigi, with random query samples also manually classified by independent reviewers. Cohen’s kappa interrater reliability scores were: 0.62 (WH & KB), 0.74 (WH & SM), and 0.58 (KB & SM). Model performance relative to reviewer WH was: accuracy 0.82, precision 0.82, and recall 0.82. The AUPRC was 0.59, and AUC was 0.85. Common use cases included email/document writing, general information questions, and text manipulation. AskDigi performed best in categorizing language translation, patient communication, and technical support/IT issues. It had the highest recall and precision in language translation, email/document writing, and medical questions. This study details real-world use cases of a secure chatbot at an academic medical center. Future research may explore its impact on the IHI quintuple aim.
Speaker(s):
Kameron Black, DO, MPH
Stanford University

Author(s):
William Haberkorn, BA - Stanford University; Stephen Ma, MD, PhD - Stanford University School of Medicine; Keith Morse, MD - Stanford University School of Medicine;

Using Large Language Models to Evaluate Patient-Centered Note Writing in Clinical Documentation

Poster Number: P48
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Communication Strategies, Documentation Burden, 21st Century Cures (including considerations for special populations such as adolescents), Artificial Intelligence/Machine Learning
Primary Track: AI and Care Outcomes
Programmatic Theme: Emerging Technology and Technical Infrastructure

We developed a note quality improvement rubric for evaluating the degree to which a clinical note reflects patient-centered attributes, including avoiding stigmatizing language. We use the rubric to annotate clinical notes from the MIMIC-III dataset and evaluate the accuracy of large language models (LLMs) to assess patient-centered attributes in clinical notes based on the rubric.
Speaker(s):
Raina Langevin, PhD
University of Washington

Author(s):
Raina Langevin, PhD - University of Washington; Livingston Martin, MD - University of Washington; Karime Bolivar, MD - University of Washington; Sophia Keen, BS - NA; Patrick Wedgeworth, MD, MISM - University of Washington; Priscilla Lui, PhD - University of Washington; Andrea Hartzler, PhD - University of Washington;

Evaluating AI-Enhanced Remote Monitoring for Hypertension Management: A Comparison of Standard and Personalized RPM Models

Poster Number: P49
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Clinician Burnout, Documentation Burden, Coordination Across the Continuum of Care, Artificial Intelligence/Machine Learning, Patient-Generated Data / Patient Reported Outcomes (PROs)
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Hypertension management requires continuous monitoring, but standard approaches may not engage all patients effectively. Previous investigation shows that remote patient monitoring (RPM) reduces blood pressure and healthcare costs but is difficult to scale due to staffing costs. This project randomizes 500 patients into standard and AI-driven RPM arms to evaluate if AI-driven personalized feedback improves blood pressure control, patient engagement, and economic outcomes.
Speaker(s):
Zachary Pope, MD, MPH
UC San Diego Health

Author(s):

Integrating Large Language Models with Machine Learning to Enhance Readmission Prediction for Patients with Colorectal Cancer

Poster Number: P50
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Risk Measurement, Data Science
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Oncology patients have a higher risk of hospital readmission than the general patient population. Predictive modeling contributes to identifying high-risk patients and enables proactive interventions. Existing models for predicting cancer patient rehospitalization typically rely on structured data for analysis, such as demographic characteristics and clinical events. However, unstructured data, such as clinical notes, hold valuable information that can significantly enhance the predictive capabilities of models. In this study, we utilize the EHR data from MIMIC-IV to evaluate the enhancement of predictive models for 30-day hospital readmission in older adults with colorectal cancer by comparing baseline machine-learning models based on structured data only and models constructed by integrating large language models (LLMs) with clinical notes.
Speaker(s):
Yun Jiang, PhD, MS, RN, FAMIA
University of Michigan

Author(s):
Xinyue Mao, BA student - University of Michigan; Serena Chen, BA student - University of Michigan; Xiayuan Huang - Yale University;

Acute and Critical Care Nurses’ Perspectives on Systemic Level Contributors of Documentation Burden: A Qualitative Study

Poster Number: P51
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Documentation Burden, EHR Implementation and Optimization, Workflow Efficiency
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Our parent study used semi-structured interviews with acute and critical care nurses to understand their perceptions of nursing documentation burden and how electronic health records system usability influences these perceptions. We applied the Systems Engineering Initiative for Patient Safety 2.0 framework in our analysis to determine if the framework could be used to understand the contributors of documentation burden and to explore the perceived multi-level contributors of nursing documentation burden.
Speaker(s):
Oliver Nguyen, MSHI
University of Wisconsin at Madison

Author(s):
Oliver Nguyen, MSHI - University of Wisconsin at Madison; Taeheon Lee, BS in progress - Ghent University; Elizabeth de Paula Fonseca, BSN - UF Health; Hwayoung Cho, PhD, RN - University of Florida College of Nursing;

Mortality Prediction

Poster Number: P52
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Risk Measurement, Change Management, Cross Setting Collaboration, Coordination Across the Continuum of Care, Big Data, Data Governance
Primary Track: AI and Care Outcomes
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Early palliative care referral could potentially reduce unnecessary tests, cost of care, and improve patient quality of life. Clinicians frequently underestimate patient illness severity and risk of death, which can result in underutilization of palliative care or hospice care services.
We identified the need for a tool providing prognostic information for patients in the ACO facing serious illness to improve appropriate and timely referral to palliative care consultative services or hospice services.
The goal of this project was to use a machine learning model to identify patients with a high risk of mortality at 360 days that would benefit from care management referral pathways to palliative care and hospice.
We developed an algorithm with high accuracy and AUC, with stable performance in the external validation set and at future time points.
This successful machine learning algorithm helped predict patients at high risk for mortality within 360 days within our clinical networks with stable performance over time and created care management workflows to review patient eligibility for palliative care interventions or hospice care transition.
Close collaboration between data analytics and clinical teams is important for translating predictive analytics into clinically impactful workflows. The information gathered in this project can help clinicians identify patients with high mortality risk within the following 360 days by using data that is already in the medical health record. Appropriate identification of high-risk patients, with effective care management pathways, can improve the ability of health systems to connect eligible patients with palliative and hospice care services.
Speaker(s):
Neil Patel, MD, MBA
Stony Brook University

Author(s):
Neil Patel, MD, MBA - Stony Brook University; April Feld, DNP, RN - Stony Brook; Rachel Wong, MD, MPH, MBA, MS - Stony Brook University, School of Medicine; Ahmad Aljobeh, M.D. - Stony Brook University Hospital; Jonathan Ambut, MD - Stony Brook University Hospital; Lara DeSanti-Siska, MD - Meeting House Lane Medical Practice/Stony Brook; Grace La Torre, DO - Stony Brook; Ned Micelli - Stony Brook Medicine; Caitlin O'Dea, LCSW - Stony Brook; Aadarsh Patel, MS Student - Stony Brook; Mary Saltz, MD - Stony Brook Medicine; Mathew Tharakan, MD - Stony Brook Medicine; Lyncean Ung, D.O. - Stony Brook Medicine; Janos Hajagos, Ph.D. - Stony Brook University;

Achieving value-based care success with an AI driven SDOH Actionability Score

Poster Number: P53
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Social Determinants of Health, Population Health, Artificial Intelligence/Machine Learning
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Community has a significant impact on individual health, with health outcomes often worsening as social vulnerabilities grow. Several publicly available data sources support social risk stratification, but they often make it difficult to create individualized profiles and lack clinical usability. The Radial SDOH Actionability Score addresses this by taking a holistic approach to social risk stratification, integrating both individual (social) and environmental risk factors to generate a personalized score with actionable insights at the level of each individual.
Speaker(s):
Kerry Son, FNP-BC, MSN, RN
Radial

Author(s):
Anant Vasudevan, MD, MBA - Radial; Rivka Atadja, DNP - Radial; Matt Wilson, DPT - Radial; Derek Olsson, BS - Radial;

Enhancing Pediatric Lead Follow-Up Testing with Clinical Decision Support and Non-Interruptive Nudges

Poster Number: P54
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Documentation Burden, Workflow Efficiency, Population Health, EHR Implementation and Optimization, Adaptive Clinical Decision Support
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Childhood lead exposure is harmful to development. Through a QI initiative, our institution increased lead testing rates and identified a need for EHR-based tools to support providers in managing elevated blood lead levels. To address this, we developed a standardized pathway, incorporating regional guidelines and implemented CDS interventions. These tools automated orders and documentation using background logic and reminded providers trough non-interruptive nudges. Real-time dashboard views were created to monitor the effectiveness of these tools.
Speaker(s):
Peter Zhang, MD, MS
Children's Hospital of Philadelphia, University of Pennsylvannia

Author(s):
Peter Zhang, MD, MS - Children's Hospital of Philadelphia, University of Pennsylvannia; Lauren Coogle, MD - Children's Hospital of Philadelphia; Elena Huang, MD - Children's Hospital of Philadelphia; Jeremy Michel, MD, MHS - The Children's Hospital of Philadelphia, Center for Biomedical Informatics;

Personalized Machine Learning Models for the Identification of Congestive Heart Failure Severity

Poster Number: P55
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Adaptive Clinical Decision Support, Data Science, Precision Health and Genomics, Care Delivery Models, Learning Health System
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

This project explores the use of incremental learning models, specifically passive-aggressive (PA) algorithms, for personalized prediction of congestive heart failure (CHF) severity using data from the NIH "All of Us" Research Program. Integrating patient data from each clinician visit, the PA model rapidly adapts to individual profiles, achieving higher accuracy and faster adaptation than traditional models. This approach empowers clinicians to improve early detection, support timely interventions, and enhance CHF management, enabling more precise medicine.
Speaker(s):
Trevor Winger
University of Minnesota

Author(s):

OPMED Informatics: Future Fight Faster!

Poster Number: P56
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Clinical informatics organizational models, Clinical Process Automation, Coordination Across the Continuum of Care, Documentation Burden, EHR Implementation and Optimization, Innovation in Digital Care, Learning Health System
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Description of the Problem
During large-scale combat operations, military medicine will face air space superiority challenges unseen for a century. This means the Military Health System (MHS) will have to contend with more patients, sicker patients and longer hold times. In order close the Kill Chain and maximize survival, several initiatives ranging from electronic health records to common data models to generative AI solutions are being leveraged in order to maximize combat casualty care outcomes.
Methods:
An environmental scan of the Military Health Systems past, present and future electronic system, data flow and modernization efforts was performed. Key stakeholders, material solutions, software acquisitions program offices, clinical communities and end-users were identified with interviews conducted to identify pain points, needs, and barriers.

Results: Gap analysis revealed several topics and themes from the key stakeholder discussion. Key domains for the MHS to focus on are: Discrete data capture with modernized electronic healthcare records, clinical decision support based on the Joint Trauma System’s Clinical Practice Guidelines, interoperability between deployed and stateside systems, dashboard generation from near real time data, synthetic data generation for training, and passive data collection resulting in real time data collection. Several projects levering emerging technology are currently under investigation to maximize outcomes in Operational Medicine environment.

Conclusion: This presentation will give an overview of the new electronic healthcare records, integrated clinical decision support, passive data mapping, near real-time dashboards, synthetic data generation and neural network creation for operational medicine.
Speaker(s):
Darshan Thota, MD
United States Navy

Author(s):
Jennifer Gurney, MD - Defense Health Agency;

Leveraging an EHR Vendor Released Predictive Risk Model to improve referral rates to a home visit program using an interruptive alert

Poster Number: P57
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Adaptive Clinical Decision Support, Care Delivery Models, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Artificial Intelligence/Machine Learning, Remote Patient Monitoring
Primary Track: Innovation in Driving Digital Equity and Social Care
Programmatic Theme: Informatics-Driven Value-Based Healthcare

Attendees will learn how to leverage predictive risk modeling to enhance clinical decision support in the EHR. This session will demonstrate how targeting high-risk subpopulations with tailored alerts can optimize utilization of high-value services, such as clinician home visit programs. By refining interruptive alerts to align with predictive models, participants will explore strategies to improve referral rates, ensure appropriate patient selection, and minimize workflow disruption, ultimately driving better patient outcomes and resource efficiency.
Speaker(s):
Timothy Lee, MD, MS
Altamed Health Sevices

Author(s):
Michael Eaton - Altamed Health Services; Timothy Lee, MD, MS - Altamed Health Sevices;

SKI: Self-Supervised Method for Pediatric CXR Classification

Poster Number: P58
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Artificial Intelligence/Machine Learning, Adaptive Clinical Decision Support, Algorithmic bias and impacts on Health Equity
Primary Track: AI and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Recent advancements in deep learning for medical AI have shown that models can match clinical experts in adult chest X-ray (CXR)diagnostics. However, their application in pediatric cases is hindered by the scarcity of annotated pediatric datasets and significant variability in CXR images across hospitals and patient age groups. To address these challenges, we propose SKI, a self-supervised approach that adapts pretrained adult models to pediatric domains by incorporating prior knowledge of intensity ranges of lung lesions. Our evaluation demonstrates that SKI (1) learns robust features for pediatric CXRs in data-limited scenarios, and (2) significantly improves model generalization and performance on unseen datasets, thereby reducing annotation costs across diverse medical image sources.
Speaker(s):
Sheng Cheng, Ph.D. student
Rice university

Author(s):
Sheng Cheng, Ph.D. student - Rice university; Devika Subramanian - Rice University;

Understanding Dynamic Communication Patterns of Inpatient Medicine Teams Using Electronic Health Record-Based Secure Messaging Metadata and Social Network Analysis

Poster Number: P60
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Big Data, Clinician Burnout, Interprofessional Collaboration
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Applying SNA to temporally-segmented electronic health records-based secure messaging (EHRSM) metadata facilitates understanding of inpatient medicine team communication dynamics and messaging volumes. This analysis provides insights about communication volumes and interconnectedness of individuals within resident and APP teams caring for internal medicine inpatients using EHRSM. These insights can inform strategies to streamline EHRSM communication, reducing unnecessary messaging, and optimize team function across differing compositions.
Speaker(s):
William Small, MD, MBA
NYU Langone Health

Author(s):
Stefanie Sebok-Syer; Nicholas Genes, MD, PhD - NYU Grossman School of Medicine; Jonathan Austrian, MD - NYU Langone Medical Center; Jesse Burk-Rafel, MD MRes - NYU Langone Health; Helen Finkelstein, MSci - NYU Langone Health;

Implementation, adoption, and impact of integrated care pathways at a freestanding children’s hospital

Poster Number: P61
Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Adaptive Clinical Decision Support, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Change Management
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Clinical Decision Support and Analytics

Integrative care pathways (ICPs) are evidence-based, structured care plans used to improve the quality of care and patient outcomes in individuals presenting with a specified clinical problem. ICP adoption historically faces barriers due to a lack of integration into clinician workflows. AgileMD is an ICP application that integrates into the electronic health record (EHR) of health systems and can be utilized to create ICPs for a variety of conditions. In June of 2023, Children’s Nebraska began the rollout of ICPs using AgileMD. This project aims to describe the implementation, adoption, and impact of ICPs utilizing AgileMD in the EHR at this institution.
The Clinical Effectiveness (CE) team implemented change management strategies to support team members in adopting changes to their preexisting workflow. To assess the adoption of the AgileMD pathways, the proportion of condition-specific qualifying encounters utilizing the AgileMD pathways overtime was added as a run chart to existing dashboards displaying that pathway’s patient outcome, process, and balancing metric performance over time.
Several ICPs were rolled out at Children’s Nebraska, though these results focus on the Heated High Flow (HHF) pathway. The average monthly adoption rate of the HHF pathway was 22% overall and 41% in encounters where HHF was initiated. The AgileMD pathway was not as well utilized among patients where HHF was not initiated indicating a need for better promotion of the pathway before initiation.
This study describes the implementation process and provides an example of how adoption and impact of AgileMD pathways are monitored over time.
Speaker(s):
Ellen Kerns, PhD, MPH
UNMC

Author(s):
Ricky Flores - University of Nebraska Omaha; Kelsey Zindel, DNP, APRN-NP, CPNP-AC/PC - Children's Nebraska; Taelyr Weekly, PhD, MPH, BSN, RN - Children's Nebraska; Chris Maloney, MD PhD - Children's Hospital and Medical Center; Russell McCulloh, MD - University of Nebraska Medical Center; Ellen Kerns, PhD, MPH - UNMC;

Rethinking the Representational / Design space of EHRs towards Modeling / Design tools. A pathway for health care to join the “Design Disciplines”

Poster Number: P62
2025 Clinical Informatics Conference 25x5 Presentation

Presentation Time: 04:00 PM - 05:30 PM

Abstract Keywords: Workflow Efficiency, Learning Health System, Documentation Burden, Disruptive and Innovative Technologies, Bridging Analytics, Bedside Care, Clinical Documentation, and Education, Cross Setting Collaboration, Interprofessional Collaboration, Clinical Process Automation
Primary Track: Documentation Burden, Clinician Well-Being and Care Outcomes
Programmatic Theme: Usability, Efficiency, and User Experience

Transforming EHRs into the Modeling/Design paradigm merges healthcare into well known "Design Disciplines," creating streamlined, adaptable systems that reduce documentation and cognitive load, enhance automation, improve data quality and neurosymbolic AI. This approach lays the groundwork for dynamic Learning Health Systems, while automating and deprecating time honoured health care processes through design driven workflow.
Speaker(s):
Arnold Kim, MD
Thunder Bay Regional Health Sciences Center

Author(s):
Arnold Kim, MD - Thunder Bay Regional Health Sciences Center; Sabah Mohammed, PhD in Computer Science - Lakehead University; Jinan Fiaidhi, PhD in Computer Science - Lakehead University; Laila Ikki, Master's Computer Science - Aurora Constellations;