Canada Spends Logo

Explore our Budget 2026 analysis

Grant: DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)

University of Toronto

$120,000.00

Using machine learning to predict, prevent, and reduce adverse events in psychiatric and mental health inpatients.

FY 2025-26

Summary
Patient safety in mental health care is a critical concern. In mental healthcare, there are unique risks and specific safety incidents including self-harm, suicide attempts, violence, aggression, restraint, and seclusion. These incidents are associated with a high burden of patient suffering, staff injury and trauma, can lead to longer hospital stays and readmissions, and higher healthcare costs. Currently, staff rely on manual risk assessment tools that are time-consuming, inconsistent, and not always accurate. Predictive machine learning (ML) and early warning systems are methods to prevent patient harm. The aim of my PhD research is to develop, deploy, and evaluate a ML-driven alerting system to predict adverse events in an inpatient psychiatric hospital. In collaboration with doctors, nurses, data scientists, and ML experts, we created the Mental health Adverse events Prediction (MAP) tool, a ML model that uses routinely collected clinical data from the electronic medical record. The MAP tool predicts if patients are at-risk of an adverse event in the next 24 hours to prompt additional evaluation and prevent further deterioration. The next step is to use a small-scale deployment strategy for the MAP tool in two acute inpatient programs at Waypoint, a large mental health hospital. Then, I will lead a clinical trial in forensic and non-forensic units at Waypoint to test the feasibility and clinical effectiveness of the MAP tool in reducing adverse events. In addition to evaluating the MAP tool, I will lead a mixed methods study where we will engage with patients and clinicians to co-develop a clinical response plan for when alerts are triggered. This will result in a standardized plan to allow for effective and efficient strategies following alerts indicating a possible risk. If successful, this research will allow clinicians to predict patient safety incidents and prevent further deterioration and incidents.
Program
DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
559656
Name
Tamayo Velasquez Valentina
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Neurosciences, Mental Health and Addiction
Theme2
Clinical
Prcname2
DT2 (Health services and population health research) CGRS Doctoral
Projecttitle
Using machine learning to predict, prevent, and reduce adverse events in psychiatric and mental health inpatients.
Programtype2
Studentship Programs
Orgnameinp2
University of Toronto
Cihramount2
$120,000
Cihrequipment2
$0
Programname2
DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)
Keyworddelim
Adverse Events; Clinical Trial; Early Warning System; Machine Learning; Mental Health; Mental Health Hospital; Psychiatry
Abstract
Patient safety in mental health care is a critical concern. In mental healthcare, there are unique risks and specific safety incidents including self-harm, suicide attempts, violence, aggression, restraint, and seclusion. These incidents are associated with a high burden of patient suffering, staff injury and trauma, can lead to longer hospital stays and readmissions, and higher healthcare costs. Currently, staff rely on manual risk assessment tools that are time-consuming, inconsistent, and not always accurate. Predictive machine learning (ML) and early warning systems are methods to prevent patient harm. The aim of my PhD research is to develop, deploy, and evaluate a ML-driven alerting system to predict adverse events in an inpatient psychiatric hospital. In collaboration with doctors, nurses, data scientists, and ML experts, we created the Mental health Adverse events Prediction (MAP) tool, a ML model that uses routinely collected clinical data from the electronic medical record. The MAP tool predicts if patients are at-risk of an adverse event in the next 24 hours to prompt additional evaluation and prevent further deterioration. The next step is to use a small-scale deployment strategy for the MAP tool in two acute inpatient programs at Waypoint, a large mental health hospital. Then, I will lead a clinical trial in forensic and non-forensic units at Waypoint to test the feasibility and clinical effectiveness of the MAP tool in reducing adverse events. In addition to evaluating the MAP tool, I will lead a mixed methods study where we will engage with patients and clinicians to co-develop a clinical response plan for when alerts are triggered. This will result in a standardized plan to allow for effective and efficient strategies following alerts indicating a possible risk. If successful, this research will allow clinicians to predict patient safety incidents and prevent further deterioration and incidents.
Competitiondate
202511
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Tamayo Velasquez, Valentina
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $120,000~ $0~