Grant: Project Grant
University of British Columbia
$348,074.00
Real-World Validation of Machine Learning Algorithms for a Wearable Cardiac Arrest Detection System
FY 2024-25
Summary
Cardiac arrest is when the heart suddenly stops beating, and affects 60,000 individuals annually. Cardiac arrest requires immediate recognition and rapid response by paramedics in order for there to be a chance of survival. Unfortunately, ¾ of cardiac arrests do not occur in the presence of a bystander who can call 9-1-1, and among these cases almost none survive. We estimate that if all unwitnessed cardiac arrest patients were detected by technology, overall survival would double, translating to an additional 4800 survivors per year in Canada.
The ultimate goal of this project is to design a cardiac arrest detection system using wearable sensors that can "witness" cardiac arrests that occur in private locations, with the capacity to call 9-1-1. To do so, we need to collect data from humans at the time that the heart stops. In this project, using several sensors within smart watches and rings, we will monitor participants in palliative care, medically assisted dying, and intensive care unit settings at the time that their heart stops (i.e., cardiac arrest). Using artificial intelligence methods, we will use these data to validate highly-accurate artificial intelligence algorithms that can be run on a smart phone to immediately identify when a cardiac arrest occurs. The sensors will communicate with our smart phone app, which will call 9-1-1 on the individual's behalf, with location coordinates.
New approaches are needed to rapidly detect cardiac arrest and facilitate an emergent professional response. In our study we will determine the best types of sensors to use to detect cardiac arrest, and will connect them to a smart phone app that will notify paramedics to arrive at the scene and provide lifesaving interventions before it is too late. With wide-spread use, this system has potential to significantly improve outcomes for cardiac arrest in the general population, especially within high-risk subgroups.
Department
Program
Project Grant
Recipient
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database
As published
Id
523243
Name
Grunau Brian E
Region
British Columbia
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Circulatory and Respiratory Health
Theme2
Clinical
Prcname2
Tri-Agency Interdisciplinary - CIHR TIR
Projecttitle
Real-World Validation of Machine Learning Algorithms for a Wearable Cardiac Arrest Detection System
Programtype2
Operating Grants
Orgnameinp2
University of British Columbia
Cihramount2
$348,074
Cihrequipment2
$0
Programname2
Project Grant
Keyworddelim
Cardiac Arrest; Machine Learning; Out-Of-Hospital Cardiac Arrest; Photoplethysmography; Resuscitation; Sudden Cardiac Death; Wearable Sensors
Abstract
Cardiac arrest is when the heart suddenly stops beating, and affects 60,000 individuals annually. Cardiac arrest requires immediate recognition and rapid response by paramedics in order for there to be a chance of survival. Unfortunately, ¾ of cardiac arrests do not occur in the presence of a bystander who can call 9-1-1, and among these cases almost none survive. We estimate that if all unwitnessed cardiac arrest patients were detected by technology, overall survival would double, translating to an additional 4800 survivors per year in Canada.
The ultimate goal of this project is to design a cardiac arrest detection system using wearable sensors that can "witness" cardiac arrests that occur in private locations, with the capacity to call 9-1-1. To do so, we need to collect data from humans at the time that the heart stops. In this project, using several sensors within smart watches and rings, we will monitor participants in palliative care, medically assisted dying, and intensive care unit settings at the time that their heart stops (i.e., cardiac arrest). Using artificial intelligence methods, we will use these data to validate highly-accurate artificial intelligence algorithms that can be run on a smart phone to immediately identify when a cardiac arrest occurs. The sensors will communicate with our smart phone app, which will call 9-1-1 on the individual's behalf, with location coordinates.
New approaches are needed to rapidly detect cardiac arrest and facilitate an emergent professional response. In our study we will determine the best types of sensors to use to detect cardiac arrest, and will connect them to a smart phone app that will notify paramedics to arrive at the scene and provide lifesaving interventions before it is too late. With wide-spread use, this system has potential to significantly improve outcomes for cardiac arrest in the general population, especially within high-risk subgroups.
Competitiondate
202503
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Grunau, Brian E; Dainty, Katie N; Kuo, Calvin; Sekhon, Mypinder S; Shadgan, Babak
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $348,074~ $0~