Hospital Discharge Preparedness via Robotic Process Automation

Presentation Time: 03:15 PM - 03:27 PM

Abstract Keywords: Informatics Implementation, User-centered Design Methods, Nursing Informatics
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Transportation is cited as a top contributor to discharge delays from the hospital. Communication gaps between ambulance vendors and hospitals exist today. Robotic process automation (RPA) can improve communication without standard data interfaces. RPA transcribed ambulance data from our third-party vendor’s dispatch system into our EHR, successfully transcribing data for 3,491 trips. However, outcome metrics did not show significant efficiency improvements, suggesting additional work is needed to prepare patients for discharge beyond awareness of transportation.

Speaker:
John Will, MPA
NYU Langone Health

Authors:
Paul Testa, MD, JD, MPH - NYU Langone Health; Jonah Feldman, MD, FACP - NYU Langone Health;

Process Analysis of Hospital Bed Flow to Optimize Access

Presentation Time: 03:27 PM - 03:39 PM

Abstract Keywords: Informatics Implementation, Information Visualization, Workflow, User-centered Design Methods, Usability
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Emergency department crowding, recognized as a national epidemic since 2006, continues to worsen due to rising patient volumes and limited resources. Our study aims to enhance leadership understanding of the bed cleaning processes influencing crowding and to support data-driven management. We performed process analytics on bed transactions to identify opportune workflow points for targeted interventions. Preliminary findings highlight staffing and bed prioritization as actionable levers to improve bed flow and optimize access.

Speaker:
Sun Won Min, MSHI
University of Texas Southwestern Medical Center

Authors:
Sun Won Min, MSHI - University of Texas Southwestern Medical Center; Janet Webb, MD - UT Southwestern; Mujeeb Basit, MD, MMSc - UT Southwestern Medical Center;

Deriving a Seriously Deteriorated Patient Identifier (SDPI) for Ward Patients

Presentation Time: 03:39 PM - 03:51 PM

Abstract Keywords: Critical Care, Clinical Decision Support, Artificial Intelligence, Evaluation, Machine Learning, Patient Safety
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Our objective was to develop a replicable, physiologically derived outcome measure for seriously deteriorated patients in hospital wards. Long standing issues with using death and unplanned transfer to the ICU undermine early warning tool evaluation and AI model training because they occur well after the onset of deterioration and omit patients who deteriorate and recover on the ward. The novel outcome derived in this work addresses both of these issues.

Speaker:
Anton van der Vegt, BSc, BE, PhD
The University of Queensland

Authors:
Anton van der Vegt, BSc, BE, PhD - The University of Queensland; Victoria Campbell, MBBS - Sunshine Coast Hospital and Health Service; Ian Scott, MHA - The University of Queensland; Daryl Jones, PhD - Austin Health; Arthas Flabouris, MD - Royal Adelaide Hospital; James Malycha, MBBS FCICM PhD - Central Adelaide Local Health Network; Imogen Mitchell, PhD - Canberra Health Services; Naitik Mehta, MHSMgt - Clinical Excellence Queensland;

Workflow Challenges and Opportunities of a Pediatric Psychiatric Intake Response Center: A Mixed-Method Study

Presentation Time: 03:51 PM - 04:03 PM

Abstract Keywords: Workflow, Pediatrics, Information Visualization
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Psychiatric Intake Response Centers (PIRCs) play a crucial role in evaluating and coordinating mental health services for pediatric patients in emergency departments (PEDs). However, prolonged PED stays often result from care transition complexities and workflow inefficiencies, especially for pediatric patients. This study employed a mixed-method analysis, integrating qualitative interviews and timestamp data from electronic health records (EHR) and real-time locating systems (RTLS), to identify key bottlenecks. Findings inform informatics-driven interventions aimed at optimizing PIRC workflow efficiency and reducing patient length of stay.

Speaker:
Danny Wu, PhD
University of North Carolina at Chapel Hill

Authors:
Zhe Shan, Miami University - Miami University; Lindsey Barrick Groskopf, DO, MPH - Cincinnati Children's Hospital and Medical Center; Advika Sumit, Undergraduate Student - University of Cincinnati; Danny Wu, PhD - University of North Carolina at Chapel Hill;

Time-series Machine Learning Models to Support Emergency Department Operational Planning

Presentation Time: 04:03 PM - 04:15 PM

Abstract Keywords: Machine Learning, Clinical Decision Support, Data transformation/ETL
Primary Track: Applications
Programmatic Theme: Clinical Research Informatics

Predicting emergency department (ED) utilization can assist in resource planning like staff scheduling. Traditional time series methods and newer machine learning methods have been used to forecast ED metrics; however, they have seldom been implemented in operational settings. We leverage a user-centered design approach that engages nursing operations managers across multiple hospital sites in an integrated health system to identify the key metrics to predict, design and select the best models and time horizons, and design a production dashboard for ED operational planning. We tested various models in terms of mean absolute error and mean absolute percentage error and determined that Prophet (a non-linear open-source method) performed the best across multiple sites. We present the implementation and monitoring design for this model, generating daily, 14-day ahead predictions for ED arrival, admission, sitter needs, and ED holds, to be used by operational leaders to guide staffing decisions.

Speaker:
Tamanna Tabassum Munia, PhD
Geisinger

Authors:
Tamanna Tabassum Munia, PhD - Geisinger; Kyle Marshall, MD, MBI, FACEP, FAAEM, FAMIA - Geisinger; Kitae Kim, MS - Geisinger; Debdipto Misra, MS - Geisinger; Grant DeLong, BA - Geisinger; Ahmed Durrani, BS - Geisinger; Jesse Manikowski, MS - Geisinger; David Vawdrey, PhD - Geisinger; Biplab S Bhattacharya, PhD - Geisinger Health System;

Data-Driven Evidence-Based Patient-Centered Optimal Initiation Time for Dialysis Treatment

Presentation Time: 04:15 PM - 04:27 PM

Abstract Keywords: Chronic Care Management, Clinical Decision Support, Machine Learning, Policy
Primary Track: Applications
Programmatic Theme: Clinical Informatics

In this study, we propose a novel decision-making framework based on natural-language-processing, machine learning, and stochastic optimization, for the purpose of providing a data-driven perspective to optimize the initiation time for dialysis treatment. When to start dialysis treatment is an important decision for end-stage chronic kidney disease (CKD) care. In the absence of a national guideline, the best time to start dialysis remains an open challenge The decision support framework includes (a) a comprehensive, efficient “pipeline” for extracting, de-identifying, and standardizing EMR data; (b) an informatics toolkit that couples natural language processing and event mapping, clustering and machine learning to uncover the disease prognosis and treatment effects to deduce the symptoms and utility rewards for each disease-action stage; (c) a first-of-its-kind personalized dialysis timing stochastic model to determine the optimal initiation time; and (d) a clinical practice guideline (CPG) for systematic testing and implementation in the clinical setting. We evaluate the results using utility rewards for each decision process and published financial costs for each CKD disease and treatment stages. Compared to the current clinical policy, the optimal initiation time policies offer potential 20.0% - 54.7% mortality reduction, an increase of 6.4% to 14.74% in overall utility reward, and a reduction of 9.6% to 16.7% in overall cost. Working with nephrologists and CKD/ESRD care team, a clinical practice guideline (CPG) was developed and tested in the clinic. Initial usage shows promising results. We caution that clinical trials must be conducted to gauge the overall effectiveness and impact on patient care of our approach. The new CPG has the potential to become a national standard.

Speaker:
Eva Lee, PhD
Georgia Institute of Technology

Author:
Di Liu, PhD - Georgia Institute of Technology;