Grant: CIHR Fellowship
University of Toronto
$100,000.00
Reducing patient boarding practices and impact on patient mortality by using optimization and machine learning methodologies that account for the heterogeneity of individual patients
FY 2018-19
Summary
Critically ill patients, requiring timely and specialized treatment, are subject to substantial delays before admission into intensive care units (ICUs) due to patient overcrowding and the lack of staffed ICU beds. Patient "boarding" has become a common strategy for hospitals to provide care for patients despite bed allocation problems in the ICU and involves caring for patients in alternative locations, such as emergency departments, until an ICU bed is available. However, these alternative locations are generally neither trained for, nor designed to treat critically ill patients, compromising the quality of care provided. As such, patient boarding and delays in ICU admission significantly increases increase patient mortality
The development of optimization and simulation tools have been promising towards reducing the number of boarded patients and delays to admission but rarely account for differences between each individual patient. By assessing patients at an individual level, hospitals may more accurately measure the impact of delays to ICU admission on patient survival.
The goal of my proposed research is to develop a novel optimization model, integrated with machine learning algorithms, that accounts for the heterogeneity of individual patients to reduce boarding practices and patient mortality. To do this, I will estimate the difference in survival between boarded and nonboarded patients based on individual patient demographics and disease characteristics, and operating information from hospitals. I will then develop an optimization model to optimize decisions regarding patient boarding, bed capacity, and staffing to maximize the predicted survival of each patient. The novel combination of predictive and optimization methodologies will allow hospitals to minimize the effects of patient boarding, reduce and eliminate boarding practices, and ultimately increase patient survival.
Department
Program
CIHR Fellowship
Recipient
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database
As published
Id
397505
Name
Sun Christopher L
Region
Ontario; Massachusetts
Country
Canada; United States of America
Orgtype
Institution Paid; Research Institution
Instname2
Population and Public Health
Theme2
Health systems / services
Prcname2
Health Research Training A - Post-PhD (HTA)
Projecttitle
Reducing patient boarding practices and impact on patient mortality by using optimization and machine learning methodologies that account for the heterogeneity of individual patients
Programtype2
Fellowship Programs
Orgnameinp2
University of Toronto
Cihramount2
$100,000
Cihrequipment2
$0
Programname2
CIHR Fellowship
Keyworddelim
Causal Inference; Clinical Informatics; Engineering; Global Health Informatics; Health Data; Icu; Optimziation; Patient And Bed Management; Patient Boarding
Abstract
Critically ill patients, requiring timely and specialized treatment, are subject to substantial delays before admission into intensive care units (ICUs) due to patient overcrowding and the lack of staffed ICU beds. Patient "boarding" has become a common strategy for hospitals to provide care for patients despite bed allocation problems in the ICU and involves caring for patients in alternative locations, such as emergency departments, until an ICU bed is available. However, these alternative locations are generally neither trained for, nor designed to treat critically ill patients, compromising the quality of care provided. As such, patient boarding and delays in ICU admission significantly increases increase patient mortality
The development of optimization and simulation tools have been promising towards reducing the number of boarded patients and delays to admission but rarely account for differences between each individual patient. By assessing patients at an individual level, hospitals may more accurately measure the impact of delays to ICU admission on patient survival.
The goal of my proposed research is to develop a novel optimization model, integrated with machine learning algorithms, that accounts for the heterogeneity of individual patients to reduce boarding practices and patient mortality. To do this, I will estimate the difference in survival between boarded and nonboarded patients based on individual patient demographics and disease characteristics, and operating information from hospitals. I will then develop an optimization model to optimize decisions regarding patient boarding, bed capacity, and staffing to maximize the predicted survival of each patient. The novel combination of predictive and optimization methodologies will allow hospitals to minimize the effects of patient boarding, reduce and eliminate boarding practices, and ultimately increase patient survival.
Competitiondate
201811
Approvedterm2
2 yrs 0 mth
Pinamesdelim
Sun, Christopher L
Approvedterm
2 ans 0 mois
Cihrcontribution2
~ $100,000~ $0~