Statistical machine learning algorithm for early detection of infection using data from consumer wearables.
Prof. Siobhan Banks & Dr. Linda Grosser

Organization
University of South Australia, Australia.
Team
Prof. Siobhan Banks
Dr. Linda Grosser
Project Description & Objectives
Apply statistical machine learning to validate unified study design and analysis approaches to generate an algorithm that can be applied to data from off-the-shelf, consumer wearables for early detection of a modelled immune response that precedes active infection.
Data Collection Process
Participants collected and wore the Garmin Venu Sq 2 for 14 days. Participants filled out a daily questionnaire about their subjective state, activities, food, health etc.
On day 11 participants received a vaccination. Participants returned device to the lab on day-14.
Fitrockr Utilization
Fitrockr was used to obtain the raw data for the health variables of interest collected by the Garmin device. Additionally, it assisted in monitoring participants to ensure the device was synced daily.
Wearable Used
Garmin Venu Sq 2
Number of Participants
106
Duration
5 months
Metrics Collected
Steps
Heart Rate
BBI
HRV
Skin Temperature
Actigraphy
Sleep
Pulse Oxygen (SpO2)
Respiration
Fitrockr Sync Type
Sync via Fitrockr app on participant smartphone.

Fitrockr is a leading health and fitness data analytics platform that empowers organizations to connect, manage and analyze data from wearables. Designed for research, clinical studies, healthcare, coaching and fitness gamification, Fitrockr enables seamless integration of biometric and activity data, providing deep insights through advanced analytics, dashboards, reports and raw data downloads. Automated alerting, push notifications, surveys and outcome assessments support data collection. With a focus on data privacy, security, and local hosting options, Fitrockr supports global institutions in transforming raw health data into actionable outcomes. Learn more at www.fitrockr.com.
© Digital Rebels GmbH. All Rights Reserved.