Enhancing Patient Medication Safety at Home: A Patient-Facing Technology Architecture Integrating REDCap, Visualization Dashboards, and an AI Driven Chatbot

Presentation Time: 10:30 AM - 10:45 AM

Abstract Keywords: Patient Safety, Tracking and Self-management Systems, Transitions of Care, Chronic Care Management, Patient / Person Generated Health Data (Patient Reported Outcomes), Patient Engagement and Preferences, Human-computer Interaction, Large Language Models (LLMs)
Primary Track: Applications
Programmatic Theme: Clinical Informatics

This work demonstrates a novel architecture for a patient-facing technology (PFT) that supports patients with cancer to self-manage medication concerns and symptoms after care transitions back home. Patient-generated data are collected and stored using the framework of Research Electronic Data Capture (REDCap), which serves as a key component of the architecture. Individual patient and administrator dashboards are integrated with REDCap to visualize medication or symptom-related data and generate customized report summaries. Additionally, a Q&A chatbot, using a retrieval augmented generation (RAG) framework, is integrated into the architecture to further enhance the interactivity of the PFT. The proposed architecture is designed as a strategic prototype for easy maintenance, cost efficiency, readiness for integration, and data security, serving as a guidance for future PFT design and development.

Speaker(s):
Yun Jiang, PhD, MS, RN, FAMIA
University of Michigan

Author(s):
Yue Yu, MD, Ph.D. - UTHealth at Houston SBMI; Yuheng Shi, MS - UTHealth Houston; Eric Yang, BS - UTHealth Houston; Katie Gahn, BS - University of Michigan; Heidi Mason, RN, DNP, ACNP-BC - University of Michigan; Yun Jiang, PhD, MS, RN, FAMIA - University of Michigan; Yang Gong, MD, PhD - UTHealth Houston;

Leveraging Large Language Models for Generating Responses to Patient Messages – A Subjective Analysis

Presentation Time: 10:45 AM - 11:00 AM

Abstract Keywords: Large Language Models (LLMs), Clinical Decision Support, Documentation Burden
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Objective: To investigate the feasibility of using large language models (LLMs) to engage with patients at the time they are drafting a question to healthcare providers and generate pertinent follow-up questions that the patient can answer before sending their messages.
Methods: We collected a dataset of patient messages sent between January 1, 2022 to March 7, 2023 at Vanderbilt University Medical Center. Two physicians identified 7 common scenarios and we located typical messages for those scenarios and removed protected health information. We used 3 LLMs to generate follow-up questions: 1) CLAIR: a locally fine-tuned LLM, 2) GPT4 with a simple prompt, and 3) GPT4 with a complex prompt. Five physicians rated them on clarity, completeness, conciseness, and utility.
Results: For seven of the scenarios, two top-rated question sets were generated by GPT4 with a complex prompt, and the other five top-rated follow-up questions were from CLAIR. The GPT4 model could generate more useful and complete but less clear follow-up questions than the questions from healthcare providers. The CLAIR model could generate follow-up question with similar clarity and conciseness as the questions from healthcare providers, with higher utility than healthcare providers and GPT4, and lower completeness than GPT4, but better than healthcare providers.
Conclusion: Patient messages sent to clinicians often lack important details, resulting in multiple rounds of messages for clinicians to gather the necessary information. Our study demonstrates that LLM can be used to generate follow-up questions when patients writing their messages, showing a great potential for improving patient-provider communication.

Speaker(s):
Siru Liu, PhD
Vanderbilt University Medical Center

Author(s):
Siru Liu, PhD - Vanderbilt University Medical Center; Allison McCoy, PhD, ACHIP, FACMI, FAMIA - Vanderbilt University Medical Center; Aileen Wright, MD - Vanderbilt; Babatunde Carew, MD - Vanderbilt University Medical Center; Julian Genkins - Stanford Health Care; Sean Huang, MD - Vanderbilt University; Josh Peterson, MD, MPH - Vanderbilt University Medical Center; Bryan Steitz, PhD - Vanderbilt University Medical Center; Adam Wright, PhD - Vanderbilt University Medical Center;

Development of Chatbot-Assisted Smoking Cessation Telehealth Intervention: A Case Report

Presentation Time: 11:00 AM - 11:15 AM

Abstract Keywords: Nursing Informatics, Telemedicine, Population Health, Behavioral Change, Informatics Implementation
Primary Track: Applications
Programmatic Theme: Consumer Health Informatics

Tobacco use is still a top preventable public health concern in the US. It causes serious health problems, including the cardiovascular system, lungs, mental issues, and cancers. To help tobacco users who want to quit, we developed a chatbot-assisted smoking cessation telehealth intervention targeting adults. This case report describes the development and evaluation of developed telehealth intervention. Utilizing rule-based chatbots and motivational interviewing principles, the study aimed to assess intervention efficacy. Data from four participants revealed varying impacts on motivation, nicotine dependency, and confidence to quit smoking. Most participants reported engaging in suggested strategies for handling tobacco cravings, indicating the potential advantage of chatbot assistance. However, the small sample size limits the generalizability of findings, necessitating further evaluation with a larger group. Despite limitations, this study demonstrates the feasibility and potential effectiveness of chatbot-assisted telehealth interventions in smoking cessation efforts, underscoring the need for future research.

Speaker(s):
Jeeyae Choi, PhD
University of North Carolina Wilmington

Author(s):
Jeeyae Choi, PhD - University of North Carolina Wilmington; Anastasiya Ferrell, PhD - University of North Carolina Wilmington; Katelyn Reiland, BS - University of North Carolina Wilmington; Julia Bertoldi, BS - University of North Carolina Wilmington; Alyssa Robb, BS - University of North Carolina Wilmington; Hanjoo Lee, MS - University of North Carolina Chapel Hill;

ARTful Replies to Patient Portal Messages: Operational Experience with Augmented Response Technology

Presentation Time: 11:15 AM - 11:30 AM

Abstract Keywords: Change Management, Documentation Burden, Workflow, Patient Engagement and Preferences, Large Language Models (LLMs)
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Our institutions implemented electronic health record (EHR) embedded functionality to help clinicians draft patient portal replies using generative AI (called augmented response technology). We compare our experiences and will share initial results and lessons learned in large academic health care systems, involving 174 primary care and subspecialty users across 20 sites.

Speaker(s):
Jacqueline You, MD
Mass General Brigham

Author(s):
Jacqueline You, MD - Mass General Brigham; Vikram Narayan, MD - Emory University; Emily Alsentzer, MS, PhD - Brigham and Women's Hospital; Lydia Siegel - None; Amanda Centi, PhD - Mass General Brigham; Elaine Goodman, MD, MBA - Mass General Hospital/Brigham & Women's Hospital/ Harvard Med School; Anne Stanislaus, BS, MS - Mass General Brigham; Christine Iannaccone, MPH - Brigham and Women's Hospital; Michelle Frits, BS - Brigham and Women's Hospital; Sayon Dutta - Mass General Hospital; Colleen Kraft, MD; Christopher Holland, MBA - Emory Healthcare; Bryan Blanchette, BA - Emory Healthcare; Reema Dbouk, MD - Emory University School of Medicine; Rebecca Mishuris, MD, MS, MPH - Mass General Brigham; Lisa Rotenstein, MD, MBA, MSc - UCSF;

VaxBot-HPV: A GPT-based Chatbot for Answering HPV Vaccine-related Questions

Presentation Time: 11:30 AM - 11:45 AM

Abstract Keywords: Delivering Health Information and Knowledge to the Public, Large Language Models (LLMs), Natural Language Processing
Primary Track: Applications

Human Papillomavirus (HPV) infection is a major global health concern, with HPV vaccines playing a crucial role in prevention. In this study, we introduced VaxBot-HPV, a chatbot powered by the Generative Pre-trained Transformer (GPT), designed to provide reliable information about HPV vaccine. We evaluated VaxBot-HPV 's performance in multiple domains and found VaxBot-HPV achieved a balance between efficiency and efficacy. This highlights the potential of large language models in enhancing HPV vaccine education and communication.

Speaker(s):
Yiming Li, Master
UTHealth Science Center Houston

Author(s):
Yiming Li - UTHealth Science Center Houston; Jianfu Li, PhD; Cui Tao, PhD - UTHealth Health Science Center at Houston; Cui Tao, PhD - Mayo Clinic;

Identifying Opportunities for Informatics-Supported Cultural Adaptation of Mental Health Chatbot: Chinese American Caregivers’ Perspectives on Self-Care and Chatbots

Presentation Time: 11:45 AM - 12:00 PM

Abstract Keywords: Diversity, Equity, Inclusion, Accessibility, and Health Equity, Patient Engagement and Preferences, Health Equity, Qualitative Methods, Mobile Health, Nursing Informatics, Natural Language Processing, Large Language Models (LLMs)
Primary Track: Applications
Programmatic Theme: Consumer Health Informatics

Providing care to a loved one can be physically, financially, and emotionally stressful and chatbots have the potential to provide affordable and accessible mental health care to family caregivers. In accordance with the 2022 National Strategy to Support Family Caregivers, which called for the development of culturally and linguistically appropriate services, this study explores the perspectives of Chinese American caregivers on self-care and using chatbots to support their mental health and well-being, with consideration of specific cultural challenges. We conducted interviews with 12 family caregivers and 21 community organization staff members, uncovering four themes regarding self-care perspectives and three themes about chatbot utilization for mental health support. Key findings suggest caregivers’ self-care practices are influenced by cultural values of self-restraint, with a general undervaluation of psychological aspect of self-care. While chatbots are recognized for being non-judgmental and potential to reduce stigma, their effectiveness may be compromised by their limitations in empathy and cultural sensitivity. Informed by the findings, this research identifies opportunities for informatics-supported strategies to culturally adapt mental health chatbots, combining Natural Language Processing (NLP) with human expertise to improve scalability and cultural relevance. For example, integrating community-derived psychoeducation materials and cultural sayings into chatbot interactions. The study indicates both potential and challenges of employing chatbots in providing culturally sensitive mental health support for Chinese American caregivers, and advocates for informatics-supported approaches to incorporate cultural nuances effectively and improve the accessibility and relevance of mental health support for diverse communities.

Speaker(s):
Serena Jinchen Xie, Masters
Biomedical Informatics and Medical Education, University of Washington

Author(s):
Serena Jinchen Xie, Masters - Biomedical Informatics and Medical Education, University of Washington; Yanjing Liang, MS - University of Washington; Jingyi Li, PhD, RN - University of Washington, Tacoma; Trevor Cohen, MBChB, PhD - Biomedical Informatics and Medical Education, University of Washington; Andrea Hartzler, PhD - University of Washington; Weichao Yuwen, PhD, RN - University of Washington Tacoma;