Feasibility of Bluetooth Low Energy wearable tags to quantify healthcare worker proximity networks and patient close contact: A pilot study

Ms Stephanie Curtis¹, Mr Asanka  Rathnayaka³, Dr Fan Wu³, Mr Md Abdulla Al Mamun³, Mr Craig Spiers⁴, Dr Gordon Bingham⁵, Professor Colleen Lau⁶, Prof. Anton Peleg², Associate Professor Mehmet Yuce³, Associate Professor Andrew Stewardson²

¹Research School of Population Health, The Australian National University, Canberra, Australia
²Department of Infectious Diseases, The Alfred Hospital and Monash University, Melbourne, Australia
³Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, Australia
⁴Alfred Brain Program, Alfred Health, Melbourne, Australia
⁵Information Development Division, Alfred Health, Melbourne, Australia
⁶School of Public Health, University of Queensland, Brisbane, Australia

Introduction

The hospital environment is characterised by a dense network of interactions involving healthcare workers (HCWs) and patients. As highlighted by the coronavirus pandemic, this represents a risk for disease transmission and a challenge for contact tracing. We developed and piloted an automated system to address this challenge and describe contacts between HCWs and patients.

Methods

We developed a bespoke Bluetooth Low Energy (BLE) system for the hospital environment with anonymous tags worn by HCWs and fixed receivers at patient room doors. Proximity between wearable tags inferred contact between HCWs. Tag-receiver interactions inferred patient room entry and exit by HCWs. We performed a pilot study in four negative pressure isolation rooms from 13 April to 18 April 2021. Nursing and medical staff who consented to participate, were able to collect one of ten wearable BLE tags during their shift.

Results

Over the four days, when divided by shift times, 27 nursing tags and 3 medical tags were monitored. We recorded 332 nurse-nurse interactions, with a median duration of 58 seconds [Inter Quartile Range (IQR): 39-101]. We recorded 45 nursing patient room entries, with a median of 7 minutes [IQR: 3-21] of patient close contact. Patient close contact was shorter in rooms on airborne precautions, compared to those not on transmission-based precautions.

Conclusion

This pilot study supported the functionality of this approach to quantify HCW proximity networks and patient close contact. With further refinements, the system could be scaled-up to support contact tracing in high-risk environments.

Biography:

Stephanie is a Master of Philosophy (Applied Epidemiology) Scholar at the Australian National University, completing part of her field epidemiology training at Alfred Health. Her research focuses on healthcare-associated infections and antimicrobial resistance, with an interest in incorporating technology and data science methods to advance understanding.