AI-Assisted Nurse Dashboard for Intelligent Bed Placement
Student: Rebecca Mbadiwe Awofisayo, Lawrence Kumah, Nitika Bhandari Instructor: Hossein Mohit, Ph.D. Industry Sponsor: WellStar
Abstract: This capstone proposes an AI-assisted nurse dashboard to optimize emergency department bed placement. By integrating a real time bed board, clinical scoring engine, and AI chatbot, the system reduces placement delays, cognitive burden, and assignment inconsistency while preserving nurse override authority. The framework supports patient safety, operational efficiency, and Wellstar’s digital transformation goals.
The Inpatient Perspective: Evaluating the Effectiveness of the Virtual Nursing Department
Workflow
Abstract: Virtual nursing (VN) is a subsection of telehealth that provides remote care to patients.
Within the hospital setting, VN models support the improvement of the nursing care
framework by strengthening inpatient support and advancing patient care standards.
The objective is to obtain quantitative data from inpatient leadership to increase
the perception of data driven opportunities for communicating between virtual nursing
and inpatient leaders. Methods include data collection through a survey; survey feedback
assesses the perception and opinions of the current workflow, cultivating opportunities
to evaluate effectiveness and identify potential workflow gaps. Results indicated
the quality of communication between inpatient and virtual nurses at 3.75 out of 5.
Nonetheless, there is a 100% satisfaction rating of the overall relationship between
inpatient units and the VN department. This project provides valuable insights on
the inpatient perspective of the current VN workflow and identifies an area for improvement.
Advancing Patient Portal Adoption at Vitruvian Health: A Strategic MyChart Activation
Initiative
Student: Tabatha Clifton Instructor: Hossein Mohit, Ph.D. Industry Sponsor: Vitruvian Health
Abstract: This project evaluated and improved MyChart patient portal activation at Vitruvian Health using a Continuous Quality Improvement (CQI) approach guided by the Plan-Do-Study-Act (PDSA) framework. Key interventions included workflow standardization, staff education, point-of-care enrollment, and targeted patient outreach. Epic SlicerDicer and MyChart dashboards supported ongoing performance monitoring and data-driven decision-making. The initiative aims to increase activation from 44% to 63%. Key findings highlight the role of health informatics in improving patient engagement, optimizing workflows, and supporting sustainable digital transformation.
AI-Powered Bed Assignment for Optimized Patient Management
Abstract: ED boarding the delay between an admission order and inpatient bed placement threatens patient safety and hospital efficiency. Manual chart reviews and fragmented communication prolong these delays even in EHR- equipped hospitals. This paper proposes extending Wellstar’s KATE AI triage tool into a real-time, AI-assisted bed assignment system within Epic, connecting clinical data with bed availability, unit capabilities, and OR schedules to generate transparent, nurse-controlled placement recommendations. Scenario-based testing across four clinical presentations validates the decision logic. Though conceptual, the framework offers a viable path to reducing ED boarding and advancing AI-driven patient flow management.
An Integrated Mobile Application for Diabetes Self-Management
Student: Cassi Proctor Instructor: Hossein Mohit, Ph.D. Industry Sponsor: WellStar Health Equity
Abstract: Diabetes mellitus remains one of the most prevalent and costly chronic diseases in the United States, requiring continuous self-management to prevent long-term complications. Although mobile health (mHealth) applications have emerged as supportive tools, many existing diabetes-focused applications address only isolated components of care. This capstone proposes and designs Sweet Spot — a comprehensive mobile application prototype integrating appointment tracking, glucose monitoring, meal planning, and medication management into a single user-friendly platform. Literature review highlights persistent gaps in preventive service utilization, medication adherence, glycemic control, and patient health literacy (CDC, 2026; ADA, 2024). These shortcomings support the need for a patient-centered digital solution that improves engagement and care coordination.
Genomic Surveillance and Informatics Framework for C. auris Outbreak Management: A
New York Case Study
Student: Kadi Doumbia Instructor: Hossein Mohit, Ph.D. Advisor: Christopher Cornelison
Abstract: Candida auris (C. auris) is a multidrug resistant fungal pathogen with high mortality
rates, capable of colonizing patients and healthcare equipment. First Identified in
2009 in Japan, has spread to over 61 countries. The New York metropolitan region has
the highest outbreak in the U.S. with 45th fatality within 90 days of initial report.
In 2024, 6,304 new cases were reported nationally, creating financial strain on facilities
lacking preventative tools. New York cases are primarily Clade I (South Asian lineage)
showing nearly 100% resistance to fluconazole and 50% to Amphotericin B.
Proactive Pediatric Workflow Optimization: Operationalizing Pediatric Readiness Through
Standardized Workflow and Decision Support
Student: Kayla Hunley Instructor: Michael Taylor, Ph.D. Industry Sponsor: Wellstar
Abstract: This project addresses the systemic “competency gap” in pediatric emergency care through the engineering of structural guardrails within the pediatric emergency room. In a high-turnover environment where staffing and management are variable, the project operates on the principle that the system must remain the constant. By transitioning from person-dependent workflows to a system-dependent informatics model, the project aims to mitigate the increase mortality risk associated with low pediatric readiness. Key interventions include a comprehensive audit of supplies, easy to access credentialing frameworks, and the alignment of nurse protocols with new employee training materials. Despite significant variation across shifts throughout the workday, the project has worked to achieve “structural success” by establishing a SharePoint Clinical Knowledge Hub a single source of resource. This digital infrastructure integrates internal standard workflows with external clinical guidelines, ensuring that rotating staff have 24/7 access to the validated decision support tools required to provide safe, evidence-based care.
Wellstar ED Inpatient Throughput Data Collection Dashboard
Students: Elisabeth Federici, Trupti Kallurwar, Zoe Luter Instructor: Michael Taylor, Ph.D. Industry Sponsor: Wellstar
Abstract: Within the current Wellstar ecosystem, critical unit data often remains fragmented across various parts of the EMRs. While tools like Epic’s SlicerDicer can generate large amounts of data, it is rarely able to synthesize it into actionable insights. This can create various blind spots within the metric tracking, making it difficult to identify whether ED room turnover bottlenecks are related to staffing, poor commun
ication across the disciplines, or resource constraints. Methodology: This project used Figma’s AI features to generate and establish its initial wireframe prototype. These designs were then manually refined to meet specific stakeholder requirements and identified clinical needs. Results: The resulting dashboard centralized import metrics for EVS, room availability, and provider orders as they relate to ED Inpatient throughput timelines. A key tool implemented was an automated refresh system that provides end users with continuous access to data reports, rather than a static number-tracking system. Conclusion: By transitioning from isolated reporting systems to a centralized, automated dashboard, this tool provides the necessary data transparency missing within Wellstar’s ecosystem. This demonstrates the power of AI-driven prototyping and the ways it can bridge the gap between high volumes of collected data and operational intelligence.
CRM Software Implementation Final Report
Student: Daniel Simpson Instructor: Michael Taylor, Ph.D. Industry Sponsor: Trelia Health
Abstract: The proposed project is to improve Trella Health LLC’s CRM software implementation process. I will be working with the director of implementations, Christopher Titone, to assess the current CRM implementation and integration process for defects and points of weakness. We will also develop and execute an initiative to improve the process internally with a goal of significantly and measurably improving the outcomes of the average implementation. Some variables that we are interested in are customer adoption and satisfaction, internal process documentation and streamlining, time to live metrics, and integration accuracy. This will be an interdepartmental effort including representation from Trella’s sales, customer success, implementation, support, and integration teams. Limited, but beneficial input will be considered from current clients as well.
Evaluating the Potential Impact of PACE Implementation in Georgia: A Healthcare Management
and Health Informatics Approach
Students: Purity Varist, Lisa Whisby Instructor: Michael Taylor, Ph.D. Industry Sponsor: Traktion
Abstract: Georgia’s aging population and increasing healthcare demands present an urgent need for integrated, cost-effective models of elder care. Although the state passed legislation in 2024 authorizing the Program of All-Inclusive Care for the Elderly (PACE), no active sites currently operate in Georgia. This capstone project evaluates the potential impact of implementing PACE statewide and explores how health informatics innovations can strengthen their launch, scalability, and effectiveness. PACE is a nationally recognized program that coordinates comprehensive medical and social services for adults aged 55 and older who qualify for nursing home care but prefer to remain in their communities. Evidence from established progr
ams in states such as Pennsylvania, California, Michigan, and North Carolina demonstrates
measurable benefits which includes fewer hospitalizations and emergency visits, lower
institutionalization rates, improved chronic disease management, and overall cost
savings for Medicare and Medicaid.
By examining demographic, economic, and healthcare data specific to Georgia, this
project estimates the eligible population, regional service needs, and potential system-wide
economic im
pact. The informatics component focuses on developing strategies to enhance enrollment
identification, care coordination, and performance monitoring. Proposed solutions
include a statewide PACE analytics dashboard, health information exchange integration,
and predictive modeling tools to identify high risk seniors. Together, these informatics-driven
approaches can enable data transparency, outcome monitoring, and efficient program
scaling. Overall, this capstone underscores the significant opportunity for Georgia
to leverage health informatics in launching a sustainable PACE network that improves
senior health outcomes, reduces healthcare expenditures and supports community-based
aging.
Patient Intake and Enrollment Process, Exploring Informatics Solutions
Student: Aleisha Meriweather Instructor: Michael Taylor, Ph.D.
Abstract: The intake and enrollment process of the Program of All-Inclusive Care for the Elderly
(PACE) explains how a digital intake solution can improve efficiency, accessibility,
and patient experience. While PACE programs have demonstrated success in improving
patient outcomes and reducing institutional care, the intake process remains complex
and toilsome. This study maps the current intake workflow to identify inefficiencies
and barriers that delay enrollment and limit access to care. A proposed digital intake
model is developed using healthcare informatics tools such as automated eligibility
verification, mobile-friendly intake platforms, and integrated health information
systems. These tools aim to streamline data collection, improve coordination across
interdisciplinary teams, and reduce manual administrative processes.
The research highlights both the strengths and limitations of the current intake process and demonstrates how digital transformation can enhance operational efficiency and pre-patient intake experience. Expected outcomes include reduced processing time intake improved data accuracy, increased access to services, and a more patient-centered enrollment process. This project emphasizes the role of healthcare informatics in modernizing intake workflows and supporting the expansion
of PACE services.
Students: Lawrence Kumah, Mbadiwe Awofisayo Advisor: Hossein Mohit, Ph.D.
Abstract: DocuSense AI is an AI-driven clinical documentation integrity platform for U.S. healthcare
organizations. It applies NLP text analysis, rules-based code validation, and DRG
impact modeling to detect documentation gaps, flag compliance risks, and estimate
revenue leakage before claims are finalized.
AI Patient Passport & Clinical Decision Support Platform
Abstract: Healthcare data fragmentation across U.S. systems causes diagnostic delays, inefficient
care coordination, and critical gaps in emergency response. This work presents an
AI-enabled Patient Passport and Clinical Decision Support Platform that centralizes
patient health information, delivers real-time risk stratification, and supports frontline
clinical decision-making. The system integrates patient profiles, lab interpretation,
risk prediction, and Al-assisted recommendations within a secure, FHIR-aligned architecture
- demonstrating measurable potential to improve triage efficiency, increase patient
engagement, and reduce operational bottlenecks at scale.
Non-Pharmacological & Technology-Based Interventions for Managing Agitation in Dementia
Care
Student: Nitika Bhandari Advisor: Modupe Atkintomide Industry Sponsor: Smart Care Team
Abstract: Agitation is one of the most prevalent and distressing behavioral symptoms in dementia,
yet existing pharmacological treatments carry significant safety risks. This systematic
review examines both technology-based and non-technology-based (in-person) non-pharmacological
interventions for managing agitation in individuals living with dementia. Technology-based
approaches, including virtual reality, robotic companions, personalized digital music
systems, tablet applications, and sensor-based monitoring tools are compared against
traditional in-person methods such as music therapy, reminiscence activities, animal-assisted
programs, and structured social interaction. Nine studies were included following
PRISMA 2020 and JBI guidelines. Findings consistently demonstrate that technology-based
interventions offer effective, scalable, and safer alternatives, with personalization
identified as the key driver of sustained agitation reduction across all care settings.
