This session will highlight recent and ongoing work in disease modelling at Statistics Canada. Building on a strong foundation of collecting high-quality population health data, Statistics Canada has developed cancer and chronic disease projection models to aid prevention efforts and policy making. During the COVID-19 pandemic, the appetite to use disease modelling to aid in decision making grew considerably. Disease modelling and modellers were suddenly thrust into the limelight! The pandemic has highlighted that population health modelling, which considers the interplay between sociodemographic factors, risk factors and health/disease states in complex open populations, is perhaps more relevant now than at any other time in history. National Statistical Agencies can play several different roles in developing population health models, including data collection, methodology development, and fostering collaboration with policy departments and external academics. Increasingly, these multidisciplinary and multi-institute teams work within a modern “data-science” and Open Science framework to support complex modelling projects. Presenting this work at the ISI World Statistics Congress will allow Statistics Canada researchers to gain a diverse and rich perspective on their work from the international statistics community.

The format of the session will include three speakers and one discussant.

1. Epidemiological modelling of COVID-19 to inform PPE demand and supply in Canada 
The global SARS-CoV2 pandemic initiated a surge in demand for personal protective equipment (PPE) in Canada. PPE continues to have a prominent role in combating the transmission of infection, especially in health and personal care settings. As such, the Government of Canada urgently needed to provide timely, accurate and relevant information on PPE procurement and deployment to provinces and territories, in the face of global supply shortages. In response, Health Canada partnered with Statistics Canada to create the Pan-Canadian Demand and Supply Model. This session will describe the collaboration between the two organizations, including how the model was used to support decision-making and how modern data-driven methods were brought to bear on this difficult question.

2. Using Statistics Canada data to measure the population burden of diabetes, cardiovascular disease, dementia, and mortality 
Statistics Canada administers a suite of national population health surveys that span more than 20 years and include 1.5 million respondents. The core health survey (Canadian Community Health Survey) covers a range of sociodemographic, behavioural, health and health care questions. Common questions are replicated in smaller surveys that include biophysical measurements or periodic special topics such as detailed food and nutrition. Disease development, transition to different care settings, and death for over 10 million person-years of follow-up are examined through linkage to an increasing range of health administrative data, disease registries, and vital statistics. During the pandemic, these health surveys included additional questions regarding Covid-19, and these can also be linked to administrative data for Covid-19 testing, vaccination and health outcomes. 
Combining large samples and longitudinal follow-up provides the foundation for observational studies that assess health risks. Cross-sectional population health surveys provide the starting population for models, and these are used to assess and recalibrate models. Predictive studies (algorithms) developed using health surveys are used in the clinical (individual) and population setting. In the population setting, risk algorithms are used for various modelling approaches, including closed cohort, stationary, and microsimulation models.

3. Microsimulation modelling of chronic disease and cancer to inform prevention and policy 
Statistics Canada has a long history of developing microsimulation models on various topics, including population health. The Population Health Model (POHEM) was developed to project the presence and impact of health-related risk factors on the burden of certain chronic conditions and mortality in the Canadian population. Additionally, OncoSim was developed in collaboration with the Canadian Partnership Against Cancer to model four cancer sites (breast, colorectal, lung, and cervical) and related screening programs in detail, and to provide high-level projections for 25 cancer sites for use by decision-makers to better understand the impact of cancer control interventions.
During the COVID-19 pandemic, POHEM and OncoSim were used to evaluate consequences of pandemic-related lockdowns/restrictions, including changes in health behaviours and paused cancer screening programs. This session will describe how these models were used to evaluate pandemic-related impacts, and how they can be used in a post-pandemic reality to evaluate longer-term health outcomes for the Canadian population.

Organiser: Deirdre Hennessy 

Chair: Deirdre Hennessy 

Speaker: Dr Rochelle Garner 

Speaker: Deirdre Hennessy 

Speaker: Douglas Manuel