Root Cause Analysis Training for Healthcare Professionals With AI-Powered Virtual Simulation: A Proof-of-Concept DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS:

Yuqi Hu, Qiwen Xiong, Zhenzhen Qin, Brandon Watanabe, Yujing Wang, Mirjana Prpa, Ilmi Yoon · 2025

Read Root Cause Analysis Training for Healthcare Professionals With AI-Powered Virtual Simulation: A Proof-of-Concept DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS: by Yuqi Hu, Qiwen Xiong, Zhenzhen Qin, Brandon Watanabe, Yujing Wang, Mirjana Prpa, Ilmi Yoon (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-04904.txt.

Introduction

Yuqi Hu and co-authors built an AI-staffed virtual hospital for teaching patient safety investigation, in a 2025 paper. Trainees learn root cause analysis by probing a bad event, work that usually needs heavy faculty time and standardized actors. Their prototype places an ICU incident inside a Unity 3D scene. Five virtual team members, played by language-model avatars with set roles and moods, submit to interviews.

Emotional synthetic voices and animated faces deepen the realism. Afterward, trainees file a written analysis and get rubric-based feedback, again generated automatically. The design targets communication, investigation, and analytic skills at low recurring cost. The paper lays out the platform and one full case.

Catalog record

CREATORYuqi Hu, Qiwen Xiong, Zhenzhen Qin, Brandon Watanabe, Yujing Wang, Mirjana Prpa, Ilmi Yoon
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-04904

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Root Cause Analysis Training for Healthcare Professionals With AI-Powered Virtual Simulation: A Proof-of-Concept DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Human-centered computing Human computer interaction (HCI) CCS: Human-centered computing Collaborative and social computing CCS: Human-centered computing Interactive systems and tools CCS: Applied computing Life and medical sciences Yuqi Hu, Qiwen Xiong, Zhenzhen Qin, Brandon Watanabe, Yujing Wang, Mirjana Prpa, Ilmi Yoon Abstract . Root Cause Analysis (RCA) is a critical tool for investigating adverse events in healthcare and improving patient safety. However, existing RCA training programs are often limited by high resource demands, leading to insufficient training and inconsistent implementation. To address this challenge, we present an AI-powered 3D simulation game that helps healthcare professionals develop RCA skills through interactive, immersive simulations. This approach offers a cost-effective, scalable, and accessible alternative to traditional training. The prototype simulates an RCA investigation following a death in the ICU, where learners interview five virtual avatars representing ICU team members to investigate the incident and complete a written report. The system enables natural, life-like interactions with avatars via large language models (LLMs), emotional text-to-speech, and AI-powered animations. An additional LLM component provides formative and summative feedback to support continual improvement. We conclude by outlining plans to empirically evaluate the system’s efficacy. 1. Introduction Each year more than 250,000 patients in the US experience adverse events ( Anderson and Abrahamson 2017 ) , where medical treatment causes harm to them ( Jason Boulanger and Keohane 2019 ) . Root cause analysis (RCA) is a crucial tool to reduce adverse events and improve quality of care, which have been widely adopted by healthcare providers ( Percarpio et al. 2008 ; Flink et al. 2005 ; Kellogg et al. 2017 ; Behrhorst et al. 2025 ) . For example, Veteran Affairs medical centers file an average of 4.86 RCAs per year, with some reporting up to 58. ( Percarpio et al. 2008 ) . RCAs focus on system-level issues instead of individual errors ( Rooney and Heuvel 2004 ; Tjia et al. 2014 ) . While there is no universal approach for conducting RCAs, most follow a common structure: describing the event, identifying contributing factors, determining root causes, and proposing corrective actions ( Rooney and Heuvel 2004 ; Williams 2001 ; Gangidi 2019 ) . Despite its recognized value, RCA is often inconsistently implemented in practice due to limited training opportunities, high resource demands, and absence of scalable educational tools ( Wu et al. 2008 ; Latino 2015 ) . RCAs consist of complicated processes of interviews, analyses and reports and typically require collaborative efforts from an interdisciplinary team ( Charles et al. 2016 ; Lee et al. 2014 ; Abreu 2021 ; Shea et al. 2019 ; Huang et al. 2025 ) . However, many healthcare professionals do not receive adequate training on how to conduct these investigations effectively ( Charles et al. 2016 ; Rooney and Heuvel 2004 ) . While RCA has been included in the curricula of medical schools ( Hall et al. 2010 ) , nursing programs ( Dolansky et al. 2013 ; Lambton and Mahlmeister 2010 ) , and continuing education initiatives ( Bowie et al. 2013 ; Murphy et al. 2017 ) , the depth and quality of this training vary widely. Current training programs, mostly in-person workshops, tend to be brief, costly, or lack interactive, hands-on components ( Aboumrad et al. 2019 ; ECRI Institute 2025 ; Tonex 2025 ) . Meanwhile, recent advancements in artificial intelligence (AI) technologies such as large language models (LLMs) offered opportunities to innovate RCA training. LLM-powered simulations have been used in medical education to improve clinical reasoning ( Mohapatra et al. 2023 ; Levin et al. 2024 ; Behrhorst et al. 2025 ) and communication skills ( Gutiérrez Maquilón et al. 2024 ; Kapadia et al. 2024 ) . For RCA training, LLMs can simulate conversations with virtual agents and allow learners to practice RCA skills through interactive and immersive simulations. In addition, LLMs can provide personalized feedback on learner performance to facilitate skill development ( Ruwe and Mayweg-Paus 2024 ; Schneider et al. 2023 ; Xie et al. 2024 ) , especially when provided with well-designed rubrics ( Wu et al. 2025 ) . In this paper, we introduce a novel system designed to facilitate RCA training for healthcare professionals - nurses, clinicians and administrators, regardless of their career stage. The system is built on the foundation of Nurse Town , a Unity-based 3D simulation game for nursing education ( Hu et al. 2025 ) . Adapted from a textbook case scenario ( Campbell and Daley 2017 ) , the simulation allows users to develop RCA skills through an investigation of a patient’s death in the ICU. Users will interview LLM-powered virtual avatars, each representing an ICU team member, to learn from their perspectives and complete a written RCA report. This system offers several advantages compared to traditional training programs: (1) Reduced resource demands : It drastically lowers financial and logistical costs for standardized patients, physical simulation centers, or complex workshop coordination; (2) Comprehensive skill development : It places a focus on fostering investigative and communication skills, which are largely ignored in conventional training methods; (3) Automated assessments : LLM-based assessments provide constructive feedback on both interviewing strategies and RCA quality, further reducing needs for human resources. 2. Related Work 2.1. Root Cause Analysis in Healthcare Root Cause Analysis (RCA) emerged in the 1990s as part of the broader quality improvement movement in healthcare, when professionals began recognizing that hospitals were becoming increasingly hazardous environments for patients ( Emanuel et al. 2009 ) . RCA was identified by the Institute of Medicine as a key analytic tool for investigating adverse events and preventing error recurrence ( Havens and Boroughs 2000 ) . The Joint Commission began requiring its accredited hospitals to conduct RCAs following sentinel events in the 1990s ( Sherwin 2011 ) . RCA has since become widely adopted in both clinical and administrative domains, supported by oversight organizations such as the Institute for Healthcare Improvement ( for Healthcare Improvement 2015 ) and the Veterans Health Administration ( Bagian et al. 2001 ; U.S. Department of Veterans Affairs 2025 ) . RCA’s value lies in its ability to uncover root causes that may not be apparent through surface-level investigation. It promotes honest communication by avoiding individual blame and focusing instead on structural and process improvements ( The Joint Commission 2025 ; Iedema et al. 2006 ) . The approach has proven effective both in U.S. and internationally. Within the U.S. Veterans Health Administration, RCA was central to a successful patient safety program ( Bagian et al. 2001 ) . In Japan, for example, RCA was also seen as a foundational tool for promoting safer care ( Wocher 2015 ) . Scholars and professional associations also support RCA’s value, noting its power to drive culture change and enhance organizational communication ( Wachter 2010 ; Vincent 2003 ) . Despite its widespread adoption, the effectiveness of RCA in practice has often been limited by inconsistent execution. Studies have found that many RCA teams lack sufficient training,

Read the complete text