Effectiveness of an Edge AI-based Augmented Reality System for Hand Hygiene Training: a Multi-centre, Mixed-methods Cluster Randomised Controlled Trial

Summary

This multi-centre, mixed-methods cluster randomised controlled trial (cRCT) aims to evaluate the effectiveness of an edge artificial intelligence (AI)-based augmented reality (AR) training and assessment system for improving hand hygiene knowledge and practice across healthcare and educational settings. Hand hygiene remains one of the most effective measures to prevent healthcare-associated infections (HAIs), which cause substantial morbidity, mortality, and economic burden worldwide. Despite ongoing education efforts, conventional training methods-such as lectures and standard videos-often lack individualised feedback and interactivity, limiting behavioural improvement. The proposed AI-empowered AR system provides real-time assessment and personalised guidance, potentially transforming hand hygiene education into an adaptive, scalable, and evidence-driven learning experience.

The trial will be conducted at three centres: the Hong Kong Polytechnic University and the Chinese University of Hong Kong in Hong Kong SAR, and Tongji Hospital in Wuhan, China. Approximately 480 participants will be enrolled, including 240 undergraduate students in health-related disciplines and 240 healthcare professionals or supporting staff from hospitals. Cluster randomisation will be applied at the session level (6-25 participants per cluster). Randomisation will be performed by a statistician independent from recruitment, and allocation concealment will be maintained. The study will adopt a non-inferiority design, comparing the AI-based AR system against an existing validated educational program combining hand scanner feedback and instructional videos. Participants and assessors of handwashing quality will remain blinded to group allocation.

At baseline, participants will provide written informed consent and complete a standardized questionnaire assessing hand hygiene knowledge and practice. They will apply fluorescent powder to both hands and perform an initial handwashing attempt, recorded by an overhead camera. The percentage of fluorescent residue (measured by the Semmelweis Scanner™) will serve as an objective indicator of decontamination effectiveness, while trained infection prevention and control (IPC) experts will independently assess technique quality from anonymized video recordings. These evaluations will also be used as reference data for AI algorithm development.

In the intervention arm, participants will perform AI-based AR training sessions providing individualized feedback on handwashing performance. In the control arm, participants will receive feedback from the fluorescent scanner and watch a standard "seven-step" handwashing educational video. Following training, participants will repeat the handwashing procedure and post-intervention assessments, including fluorescent residue measurement and the same knowledge-practice questionnaire. The primary outcome is the improvement in decontamination effectiveness (reduction in fluorescent residue percentage between pre- and post-intervention). Secondary outcomes include correct performance of all seven handwashing steps, improvement in knowledge scores, and user satisfaction.

A qualitative component will complement the RCT to explore participants' experiences, attitudes, and perceived barriers toward AI-based AR hand hygiene training. A purposive subsample from both arms will be invited to participate in semi-structured interviews based on the Theoretical Domains Framework. Interviews will examine usability, motivation, learning experience, and perceived behavioral changes. Data will be audio-recorded, transcribed verbatim, and analyzed using thematic analysis with iterative coding to ensure credibility and consistency.

Ethical approval will be obtained from relevant institutional review boards. All participants will be informed of their rights, including voluntary participation and the ability to withdraw at any time without penalty. Potential risks are minimal, limited mainly to rare allergic reactions to fluorescent powder. All data, including video and image recordings, will be anonymised, stored securely, and deleted one year after project completion. No personally identifiable information will be collected.

This study will generate high-quality evidence on the feasibility and educational impact of integrating AI and AR technologies into hand hygiene training across academic and clinical settings. If proven effective, the system can be incorporated into health professional curricula and used at the point of care for real-time performance monitoring and auditing, contributing to sustainable infection prevention and control capacity building.

Conditions

• Hand Hygiene Training
• Infection Prevention and Control
• Healthcare-Associated Infections (HAIs)
• Health Education

Interventions

BEHAVIORAL

Instructional Video + Scanner Feedback

Viewing quantified fluorescent residue from the first handwash plus a standard 7-step instructional video prior to the second handwash.

DEVICE

Edge AI-based AR Hand Hygiene Training System

An edge AI + AR system using computer vision to assess handwashing performance and deliver real-time, individualized guidance before the second handwash.

Sponsors & Collaborators

• Chinese University of Hong Kong

  collaborator OTHER

• Tongji Hospital

  collaborator OTHER

• The Hong Kong Polytechnic University

  lead OTHER

Study Design

Allocation

RANDOMIZED

Purpose

HEALTH_SERVICES_RESEARCH

Masking

SINGLE

Model

PARALLEL

Eligibility

Min Age

18 Years

Sex

ALL

Healthy Volunteers

Yes

Timeline & Regulatory

Start

2025-12-15

Primary Completion

2027-08-01

Completion

2027-09-01

Countries

• China
• Hong Kong

Study Locations