This webinar is part of the Advanced Methods Webinar Series
The outbreaks of COVID-19 have posed a significant pressure on the health care system. There is therefore a great demand to predict which cases are likely to require hospitalization, to in turn identify hospitals most at-risk of being overwhelmed should such an outbreak occur. Accurate predictions will inform planning and resource allocation to best prevent failures in healthcare systems.
In response to this demand, the British Columbia (B.C.) Ministry of Health has investigated factors that might lead to a person being hospitalized and/or die from COVID-19. The ministry has developed several individual-level models (oft-revised, as new information on the virus and pandemic arise) to study risk factors specific to different segments of the population and provide community risk profiles. Using a daily-updated feed of newly identified cases of COVID-19, these models have been used to predict hospitalizations in local health authorities province-wide, and has informed decision-making among various senior public-management groups across B.C. to identify potential resource shortages before they become critical.
The predictors include at-risk health conditions from CIHI’s Population Grouping Methodology (CPOP), a case-mix data product generated using administrative health records. Here before talking about COVID-19 daily hospitalization forecast, we will give an introduction on two ministry owned case-mix data sources, CPOP and Heath System Matrix (HSM). These data have been widely used in various projects such as disease prediction and management, performance measure, workload or utilization distribution, and funding model assistance.
This presentation will cover the following parts.
- Introduction to ministry owned datasets CPOP and HSM
- Overview on applications using CPOP and HSM
- COVID-19 daily hospitalization forecast based on a logistic model
- COVID-19 ICU admission forecast based on a negative binominal model
Yongcai Liu, MSc, MEng, PhD is a senior economist in Health Sector Information, Analysis and Reporting Division, B.C. Ministry of Health where he has been working on statistical and economic modeling and analysis for nearly 15 years. He holds a Master’s in Statistics and a PHD in engineering. He has plenty of experience in various machine learning techniques and statistical/economic algorithms, such as logistic regression, neural network, survival analysis, time series analysis, statistical simulation, and cost benefit/effectiveness analysis. He also has many years of academic research experience in numerical modeling and analysis in the fields of engineering and science. Currently he is a member of External Modelling Experts Working Group of Canada.
Samantha Magnus, MPH, MBA is Director of the Methodologies and Cross-Sector Analysis team in the Health Sector Information, Analysis and Reporting Division, B.C. Ministry of Health. She holds a Master’s in Public Health from the University of Victoria as well as an MBA from Quantic School of Business and Technology. She sits on the Canadian Institute for Health Information’s (CIHI) case-mix advisory group and has several years’ experience in both qualitative and quantitative research.