Trajectories of In-Person and Telehealth Utilization: A Sequence Analysis of Primary Care Modalities Pre- and Post-COVID-19

Presentation Time: 08:00 AM - 08:18 AM

Abstract Keywords: Telemedicine, Quantitative Methods, Informatics Implementation, Health Equity, Chronic Care Management, Healthcare Quality
Primary Track: Applications
Programmatic Theme: Clinical Informatics

This study examines in-person, telehealth, and between-visit encounter patterns among individuals with diabetes at two San Francisco health systems, using sequence analysis to identify clusters reflecting increased utilization, transition to digital care, and care loss post-pandemic. Increased utilization was associated with higher comorbidity burden, transition to digital care with younger age, and care loss to older age and language barriers. Disparities in digital health adoption underscore the need for equitable strategies to prevent care disengagement.

Speaker:
Namuun Clifford, MSN, RN, FNP-C
The University of Texas at Austin

Authors:
Elaine Khoong, MD, MS - University of California San Francisco; Taylor Rapson, PhDc, MPH - University of Washington; Namuun Clifford, MSN, RN, FNP-C - The University of Texas at Austin; Kathryn Kemper, MPH - University of California San Francisco; Courtney Lyles, PhD - UC Davis Center for Healthcare Policy and Research; Elizabeth Sherwin, MPH - Stanford University;

---

Grouping Stratified Multi- Trajectory Modeling to Identify Sex-specific Differences in Depression Visits during Pre-Dementia Period

Presentation Time: 08:18 AM - 08:36 AM

Abstract Keywords: Data Mining, Knowledge Representation and Information Modeling, Information Extraction, Information Retrieval
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Recent surge in identifying differences among patient groups have emphasized the need of precision medicine. Existing patient grouping methods use a single patient cohort trajectory. We argue that the patient cohort follows multiple trajectories and differences among patients must be identified based on multiple trajectories. We define a stratified strategy to introduce a constrained fitness function for different groups and obtain optimized weights for joint distance matrix. The idea behind introducing our method is to develop a robust approach for identifying differences through stratified analysis of patient cohorts across multiple healthcare trajectories and systematically grouping them for supporting personalized healthcare.

Speaker:
Muskan LNU, Postdoctoral Research Fellow
Mayo Clinic

Authors:
Muskan LNU, Postdoctoral Research Fellow - Mayo Clinic; Xingyi Liu, Ph.D. - Mayo Clinic; Mari Vassilaki, PhD, M.D., - Mayo Clinic; Ronald C Petersen, MD, PhD - Mayo Clinic; Jennifer St.Sauver, PhD, M.P.H. - Mayo Clinic; Sunghwan Sohn, PhD - Mayo Clinic;

---

Mild Cognitive Impairment Patients’ Trajectory Analysis Based on a Clinical Foundation Model

Presentation Time: 08:36 AM - 08:54 AM

Abstract Keywords: Data Mining, Deep Learning, Quantitative Methods
Primary Track: Applications
Programmatic Theme: Clinical Informatics

This study aims to enhance understanding of Mild cognitive impairment (MCI) progression by integrating BiGRU and Med-BERT to generate patient representations from cumulative medical histories. By clustering these representations into stages and applying pseudotime analysis, we reveal MCI patients’ progression patterns and pathways. Results show a more nuanced, quantitative understanding of MCI progression, compared to other binary classification methods, providing valuable insights into individual disease trajectories.

Speaker:
Jianping He, Master's Degree
University of Texas Health Science Center at Houston, McWilliams School of Biomedical Informatics.

Authors:
Jianping He, Master's Degree - University of Texas Health Science Center at Houston, McWilliams School of Biomedical Informatics.; Laila Rasmy, PhD, MSc, MBA, RPh. - UTHealth MSBMI; Yue Yu, Ph.D. - Mayo Clinic; Ziqian Xie, Ph.D. - UTHealth; Anna Motupalli, MBBS - UTHealth; Cui Tao, PhD - Mayo Clinic; Degui Zhi, Ph.D. - The University of Texas Health Science Center at Houston (UTHealth) McWilliams School of Biomedical Informatics;

---

DeepJ: Graph Convolutional Transformers with Differentiable Pooling for Patient Trajectory Modeling

Presentation Time: 08:54 AM - 09:12 AM

Abstract Keywords: Data Mining, Deep Learning, Knowledge Representation and Information Modeling
Primary Track: Applications
Programmatic Theme: Clinical Research Informatics

In recent years, graph learning has gained significant interest for modeling complex interactions among medical events in structured Electronic Health Record (EHR) data. However, existing graph-based approaches often work in a static manner, either restricting interactions within individual encounters or collapsing all historical encounters into a single snapshot. As a result, when it is necessary to identify meaningful groups of medical events spanning longitudinal encounters, existing methods are inadequate in modeling interactions cross encounters while accounting for temporal dependencies. To address this limitation, we introduce Deep Patient Journey (DeepJ), a novel graph convolutional transformer model with differentiable graph pooling to effectively capture intra-encounter and inter-encounter medical event interactions. DeepJ can identify groups of temporally and functionally related medical events, offering valuable insights into key event clusters pertinent to patient outcome prediction. DeepJ significantly outperformed five state-of-the-art baseline models while enhancing interpretability, demonstrating its potential for improved patient risk stratification.

Speaker:
Deyi Li, Master of Science
University of Florida

Authors:
Deyi Li, Master of Science - University of Florida; Zijun Yao, Ph.D. - University of Kansas; Muxuan Liang, Ph.D. - University of Florida; Mei Liu, PhD - University of Florida;

---