Leveraging ChatGPT as programming support for PhD-level nursing students: Results from a mixed-methods experimental study

Presentation Time: 08:30 AM - 08:45 AM

Abstract Keywords: Education and Training, Large Language Models (LLMs), Nursing Informatics
Working Group: Nursing Informatics Working Group
Primary Track: Applications
Programmatic Theme: Academic Informatics / LIEAF

We aimed to evaluate the advantages and disadvantages of using ChatGPT for programming support for PhD-level nursing students. We employed an experimental study design within a data science and visualization course for PhD-level nursing students. ChatGPT helped students write and troubleshoot code, reduced frustration, and saved time spent on assignments for novices but did not alter confidence scores. ChatGPT may be an effective tool in conjunction with other resources for teaching data science and visualization.

Speaker(s):
Meghan Reading Turchioe, PhD, MPH, RN
Columbia University School of Nursing

Author(s):
Sergey Kisselev, MA - Columbia University; Suzanne Bakken, RN, PhD - Columbia University School of Nursing;

Program for Artificial Intelligence Readiness (PAIR): Building and Sustaining AI/ML Research Capacity at Low-Resource Institutions

Presentation Time: 08:45 AM - 09:00 AM

Abstract Keywords: Health Equity, Diversity, Equity, Inclusion, Accessibility, and Health Equity, Education and Training, Fairness and Elimination of Bias, Machine Learning, Racial Disparities
Primary Track: Foundations
Programmatic Theme: Academic Informatics / LIEAF

Many communities and institutions have untapped potential to contribute new expertise, data, recruitment strategies, and cutting-edge science to Artificial Intelligence and Machine Learning. NIH funded the AIM-AHEAD program to enhance the participation and representation of underrepresented researchers and communities in developing AI/ML models. AIM-AHEAD's PAIR program fills this gap by building capacity for, and creating multidisciplinary AI + Health Equity Labs at low-resource institutions. This podium presentation will highlight the program's success and lessons learned

Speaker(s):
Toufeeq Syed, PhD, MS
UT Health Houston

Author(s):

Enhancing Causes of Death Prediction from Electronic Health Records through Multi-Modal Integration of Structured and Unstructured EHR Data

Presentation Time: 09:00 AM - 09:15 AM

Abstract Keywords: Natural Language Processing, Machine Learning, Knowledge Representation and Information Modeling, Large Language Models (LLMs), Data Mining, Informatics Implementation, Bioinformatics, Population Health
Primary Track: Applications
Programmatic Theme: Academic Informatics / LIEAF

This study demonstrates the importance of integrating heterogeneous EHR data to enhance mortality prediction accuracy. The proposed framework leverages the complementary predictive strengths of structured data elements and unstructured clinical text. By integrating patient-level embeddings generated from clinical notes, the model performance was improved, achieving a 5% increase in F-measure and a 4% lift in AUC over using structured data alone. These results highlight the significance of multi-modal modeling approaches that integrate insights from both structured and unstructured EHR data to achieve performance gains. In the context of specific CoD, combining unstructured notes with structured data in multi-modal modeling enhanced mortality predictions for 12 out of 15 CoDs, better than the performance achieved with only structured data. The significant gains were for less common conditions like cerebrovascular disease, essential hypertension, intentional self-harm, and Alzheimer's. This suggests unstructured notes contain signals that can enhance performance for classes with fewer samples, as structured data alone may underestimate minority CoD. This approach can potentially enrich epidemiological research and contribute to developing improved healthcare policies and practices. This methodology faces challenges of computational complexity, resource demands from vast unstructured data, and limitations from data quality that might lead to poor generalization, particularly in healthcare with constrained computational resources.

Speaker(s):
Mohammed Al-Garadi, PhD
VUMC

Author(s):
Mohammed Al-Garadi, PhD - VUMC; Ruth Reeves - Tennessee Valley Health Care System, US Veterans' Affairs; Rishi Desai - Brigham and Women's Hospital; Michele LeNoue-Newton; Daniel Park, M.S - VUMC; Shirley Wang - Harvard Medical School/Brigham & Women's; Judith Maro, PhD - Harvard Pilgrim Health Care Institute and Department of Population Medicine, Harvard Medical School, Boston, MA; Candace Fuller, PhD - Harvard Pilgrim Health Care Institute and Department of Population Medicine, Harvard Medical School, Boston, MA, USA; Joshua Lin, PhD - Division of Pharmacoepidemiology and Pharmacoeconomics at the Brigham and Women’s Hospital; José Hernández-Muñoz, PhD - Food and Drug Administration, Silver Spring, MD; Aida Kuzucan, PhD - Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD; Xi Wang, PhD - Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD; Haritha Pillai, M.S - Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA; Kerry Ngan, MCS - Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA; Melissa McPheeters, PhD, MPH - RTI International; Jill Whitaker, MSN, RN-BC - VUMC; Michael Matheny, MD, MS, MPH, FACMI, FAMIA - Vanderbilt University Medical Center;

Multimodal Ensemble Learning for Accurate Detection of Glaucoma

Presentation Time: 09:15 AM - 09:30 AM

Abstract Keywords: Machine Learning, Natural Language Processing, Personal Health Informatics
Primary Track: Applications
Programmatic Theme: Academic Informatics / LIEAF

In the present study, we developed a multimodal machine learning (ML) model to accurately identify glaucoma patients using electronic health records. The combined approach uses both structured data and the embeddings from clinical notes, employing data from the Sight Outcomes Research Collaborative Ophthalmology Big Data consortium. We used ensemble learning taking multiple multimodal ML models, each trained with different resampling techniques to improve balance between precision and recall. The ensemble method outperformed individual models in overall performance, proving more effective in identifying glaucoma across various disease stages. It can be further developed as a tool to support research and quality improvement in eye healthcare. It has a potential use in genotypic-phenotypic association studies and precision medicine initiatives.

Speaker(s):
Tushar Mungle, PhD
Stanford University

Author(s):
Behzad Naderalvojoud, PhD; Hong Su An, PhD - Department of Ophthalmology and Visual Sciences; University of Michigan; Chris Andrews, PhD - Department of Ophthalmology and Visual Sciences; University of Michigan; Amanda Bicket, MD, MSE - Department of Ophthalmology and Visual Sciences; University of Michigan; Amy Zhang, MD - Department of Ophthalmology and Visual Sciences; University of Michigan; Julie Rosenthal, MD - Department of Ophthalmology and Visual Sciences; University of Michigan; Suzann Pershing, MD - Department of Ophthalmology and Visual Sciences, Byers Eye Institute, Stanford University; Joshua Stein, MD, MS - Department of Ophthalmology and Visual Sciences; University of Michigan; Department of Health Management and Policy, School of Public Health, University of Michigan; Tina Hernandez-Boussard, PhD - Department of Medicine, Stanford University;

PathSAM: Enhancing Oral Cancer Detection with Advanced Segmentation and Explainability

Presentation Time: 09:30 AM - 09:45 AM

Abstract Keywords: Machine Learning, Cancer Prevention, Diagnostic Systems, Large Language Models (LLMs)
Primary Track: Foundations
Programmatic Theme: Academic Informatics / LIEAF

Leveraging the Segment Anything Model's (SAM) success in general image segmentation, "PathSAM: SAM for Pathological Images for Oral Cancer Detection" specifically targets the nuanced challenges of oral cancer diagnosis. Although SAM is celebrated for its adaptability, its application to pathological images is hindered by their complexity and variability. PathSAM outperforms conventional deep-learning approaches, demonstrating superior accuracy and detail on critical datasets like ORCA and OCDC through both quantitative and qualitative measures. By integrating Large Language Models (LLMs), PathSAM significantly enhances the explainability of its segmentation results, a vital feature for accurately identifying tumors and improving patient-provider communication. This capability to navigate the intricacies of pathological images cements PathSAM's role as an innovative solution in the field of medical diagnostics.

Speaker(s):
Suraj Sood, Ph.D
University of Missouri-Kansas City

Author(s):
Sunny Sood, BS - University of Missouri - Kansas City; Syed Jawad H. Shah, MS - University of Missouri - Kansas City; Saeed Alqarni, MS - University of Missouri - Kansas City; Yugyung Lee, PhD - University of Missouri - Kansas City;