Projects

Principal Investigator: Daniel Mulder

Institution: Queen’s University

Description: Understanding the immune system changes that occur due to SARS-CoV-2 infection could dramatically improve our ability to care for patients with COVID-19. Inflammatory bowel disease (IBD) is a common, lifelong immune-mediated condition that can be thought of as a chronic state of immune hyper-reactivity. Patients with IBD experience increased morbidity due to SARS-CoV-2 infection (Zollner et al. Gastro 2022). Like many chronic inflammatory conditions, IBD is characterized by periods of exacerbated inflammation, also known as "flares". Exacerbations are often unpredictable. Recently, the SARS-CoV-2 virus has been clearly shown to cause increased inflammation and disease exacerbations in some patients with IBD, but not others (Ludvigsson et al. UEGJ 2021; Lukin et al. CGH 2021). The reasons for this susceptibility are unknown. It seems likely that due to the heterogeneity of IBD immunology and immunosuppression from IBD therapy, some patients with IBD are more susceptible to inflammation from SARS-CoV-2 (Zollner et al. Gastro 2022). Another way of conceptualizing this situation is to consider some IBD patients to be protected from SARS-CoV-2 due to their unique immunology. We hypothesize that differences in immune markers in IBD patients can be linked to susceptibility to SARS-CoV-2-related inflammatory response. By examining differences in routinely collected blood tests (white blood cell populations and inflammatory markers) before and after SARS-CoV-2 infection and then comparing patients who experienced inflammatory exacerbations with those who did not, we will learn which patient characteristics are indicative of protection from hyper-inflammatory states and which indicate risk for these states. A combination of health care database evaluation and statistical analysis with supervised machine learning will be used to identify critical distinctions between the inflammatory responses of different IBD patients in response to SARS-CoV-2 infection.

Our 3 main objectives are:

Objective 1: Establish a high-fidelity cohort of IBD patients infected by SARS-CoV-2 using the Ontario Health Data Platform (OHDP). O

Objective 2: Use regression modelling to determine clinical and demographic features that could be protective from COVID-19-related inflammation in IBD, including demographic features (age, sex, socio-economic status, and rural/urban location), IBD features (subtype, location, severity, and current treatment) and COVID-19 features (vaccination status, variant subtype).

Objective 3: Use a supervised machine learning approach to determine the unique clinical features that distinguish type and degree of inflammation in IBD patients when infected with COVID-19.

Datasets: COVaxON, DAD, NACRS, ODB, OHIP, OLIS, OLIS-C19, RPDB

Principal Investigator: Aleksandra Zuk

Institution: Queen’s University

Description: The objectives of this project are as follows:

1. To determine the relative impacts of marginalization, high-risk clinical characteristics, and COVID on patient mortality.
2. To determine the relative impacts of marginalization, high-risk clinical characteristics, and COVID on HCB (utilization).
3. To establish the extent that symptom triage may be adjusted to accommodate SDH to support marginalized patients during the COVID pandemic?

The systemic marginalization of Canadians disproportionately impacts COVID-19 incidence, morbidity, and mortality. Current data has demonstrated this inequality through reported deaths per 100,000 persons within marginalization categories (e.g., gender, age, income, housing). Urban densely populated neighborhoods with a high proportion of immigrants, single-parent families, and low-income families had significantly higher COVID-19 mortality rate (66.7/100,000 population) according to Statistics Canada data. Unfortunately, there is much less information pertaining to the COVID-19 morbidity, system burden, and human suffering that is occurring at the intersectionality of two or more marginalization factors. Using health care-linked databases we will apply both statistical methods and supervised machine learning to examine the impact of marginalization and COVID-19. This study will allow us to follow marginalized Canadians longitudinally to understand the relative influence of intersecting marginalization factors on COVID-19 diagnostics tests, infection rates, treatment, morbidity, and mortality.

Datasets: CCRS, CCM, COVaxON, DAD, eCTAS, NACRS, NMS, ODB, OHIP, OLIS-C19, OMHRS, RPDB

Principal Investigator: Vedat Verter

Institution: Queen’s University

Description: There were over 1.5 million confirmed cases of COVID-19 in Ontario, which claimed over 15,000 lives in the province. In addition to this tragedy, the pandemic caused significant strain on the healthcare system and substantial economic burden, which are well documented. One of the less studied consequences of this pandemic is its impact on mental health of the population. Although there are a few studies observing an increase in anxiety and depression during COVID-19, none of these prevailing works are comprehensive enough to assist with policy design. Considering the prevalence of mental illnesses in the province, a solid understanding of the impact of COVID-19 in this domain is essential to be able to improve our preparedness for the next public health crisis.

Social distancing and lockdowns were aimed at mitigating the spread of COVID-19. Among the undesirable consequences of these measures are social isolation, loss of income, inactivity, limited access to basic services, increased access to food, alcohol, and online gambling, and decreased family and social support. These effects are presumably more pronounced among vulnerable populations, including mental health patients. In addition, the in-person access of these patients to acute care facilities were limited, again as a preventive measure. Our aim is to study the resulting effects on the patient population.

Using the data set provided by OHDP, we intend to document the trajectories of the mental health patients through the system in receiving care. This requires identifying every touch point each patient with mental illness had with the care providers, including outpatient clinics, emergency room and hospitals. Using explainable machine learning approaches, we would be able to identify (i) the changes in the demand (ii) the changes in the care patterns, and (III) the changes in the health outcomes during COVID-19.

Datasets: CCRS, DAD, NACRS, ODB, OHIP, OLIS-C19, OMHRS

Principal Investigator: Mina Tadrous

Institution: Women’s College Hospital

Description: We will lead a broad body of work that will explore the impact of the COVID-19 pandemic and policies and changes on 3 different patient cohorts. The three cohorts will explore different types of costly medications used to treat chronic diseases or mental health diagnoses, both of which have been identified as important areas of focus by citizen groups and policymakers. In particular, we are interested in how these drugs are being used by patients, how effective they are in treating disease, and how much these medications are costing the health care system.

1. Given the high number of individuals with diabetes and because treatments are costly, we will study novel diabetic treatments. Specifically, we will look at how well and in whom these drugs work best, and identify the best protocol for using these medications, such as in what order they should be used when treating patients with diabetes. We will explore how the pandemic affected the treatment of those living with diabetes.

2. Biologic drugs are used to treat many different diseases but are often costly. Therefore, we will study how well similar, lower-cost drugs - called biosimilars - work in treating disease, as well as how safe they are. Ontario also implemented a policy mandating that new users of biologics use biosimilars instead, so we will look at how this new policy changed the use of these drugs and the frequency of switching of medications. Finally, we will explore how the pandemic affected the utilization of these medications.

3. As the ongoing opioid overdose crisis evolves, we will investigate how many new cases of opioid use disorder (OUD) there are. We will also look at how patients are using new or existing medications for OUD, such as how often they are used and how patients transition between treatments. In recent times, many new and innovative, but potentially costly, drugs are being developed to treat disease. The COVID-19 pandemic has also put pressure on the government to fund new drugs, as there is now a higher demand for drugs but also lower supply of and access to drugs. This project will provide needed evidence to guide the government in creating drug policies and knowing what medications to fund for prevalent diseases. This information will be important for current and future planning as we exit the pandemic and future-proof out healthcare system.

Datasets: COVaxON, DAD, NACRS, NMS, ODB, OHIP, OLIS, OLIS-C19, OMHRS, RPDB

Principal Investigator: Peter Tanuseputro

Institution: Bruyère Research Institute

Description: This project is a strategic collaboration between the Ottawa Hospital Research Institute (OHRI) and the Bruyère Research Institute (BRI). We propose the creation of a learning health system to improve palliative and end-of-life care in long-term-care (LTC) homes across Ontario. There are several components of the proposed learning health system to support the LTC sector in the post-pandemic recovery phase:

1) This project will first the OHDP platform to provide 'audit and feedback' of performance of LTC homes throughout the course of each residents' stay (e.g., transfers to acute care, end-of-life prescribing and quality of life indicators). This includes data for patients with and without COVID infection. Data will be analyzed at the care provider (e.g., physician), home, and region (e.g., former LHIN) level.

2) Secondly, this project will use the OHDP platform to create a prognostic tool of health outcomes such as survival, cognitive/functional decline, and transfers to acute care. COVID infection and outbreak status will be considered as one of the predictors of interest. We will use traditional statistical methods but also include the use of AI/machine learning methods when appropriate. We have created similar tools using ICES data (www.respect.projectbiglife.ca and www.individualizedhealth.ca). We aim to place our new tools on these online platforms for individual use.

3) OHRI and BRI will work with our partners such as the Centers for Learning, Research and Innovation in LTC (CLRI) and the Ontario Palliative Care Network (OPCN) to integrate our prognostic tool across LTC homes and retirement homes in Ontario. We aim to improve automated identification of those who may be approaching the end-of-life and provide information about other health outcomes. We have done so successfully for home care clients in Champlain and hope to expand our reach. Finally, we will work with our partners to provide support to homes/physicians/other health care workers to improve the delivery of palliative care, including: creation of order sets to support end-of-life prescribing, training staff on serious illness conversations, and connecting to regional palliative care teams. This work will be informed by the data generated in the above activities - targeting those (e.g., LTC homes, health care providers) that may benefit from these interventions.

The project has the following components:

1. Long-term care end-of-life prescribing and emergency room transfers: Audit Feedback at the long-term-care home level during pandemic recovery
2. Examining the delivery of end-of-life care for long-term care residents by physicians across Ontario throughout the COVID-19 pandemic
3. Examining variations and predictors of transfers of from long-term care (LTC) homes to emergency rooms for residents with and without COVID infection
4. Evaluating where patients designated as Alternate Level of Care (ALC) end up after discharge from an acute care hospital during the different phases of the COVID-19 pandemic.
5. Post-surgical transitions from acute care for frail long-term care residents and home care recipients during the different phases of the pandemic
6. Long-COVID among long-term care residents: describing the trajectories and predictors of health decline following a COVID-19 infection
7. Quality of life in long-term care homes in the recovery phase of the COVID-19 pandemic
8. Mental health among LTC residents during the COVID-19 pandemic and its impact on health and well-being
9. Prognosticating the survival time of long-term care residents to improve care delivery in the recovery phase of the COVID-19 pandemic
10. Prognosticating health outcomes for long-term care residents in the recovery phase of the COVID-19 pandemic
11. Evaluating the impact of home care services and caregiver burden for those with dementia in the community on transitions into long-term care homes
12. Incidence of Parkinson's disease in Ontario and calibrating the PREDIGT Score for predicting the risk of developing PD

Datasets: CCRS, COVaxON, DAD, NACRS, NMS, ODB, OHIP, OLIS, OLIS-C19, OMHRS, RPDB

Principal Investigator: Shy Amlani

Institution: William Osler Health System

Description: The North American COVID-19 and STEMI (NACMI) registry is a prospective, investigator-initiated, multi-center registry of hospitalized patients with STEMI and confirmed or suspected COVID-19 infection from 64 centers in North America. The objective of the NACMI registry is to provide a resource from which to obtain insights into the demographic, characteristics, management strategies, and outcomes of STEMI patients with COVID-19 to guide clinical practice(1). Studies from this registry showed that, compared to STEMI patients without COVID 19, STEMI patients with COVID-19 (1) have a higher mortality, (2) have a different clinical presentation, (3) and have differences in outcome according to gender and race. Myocardial infarction (MI) continues to be the leading cause of death across North America. While MI epidemiology has been documented, specific demographics of patients with MI in Ontario are still unknown. Very little is known about the effect of COVID-19 in the Non-ST elevation myocardial Infarction population and its impact on patient outcomes, given than the volume of NSTEMI patients is almost triple the volume of STEMI patients. Of the limited research, concomitant COVID-19 with NSTEMI diagnosis is known to significantly increase mortality rate, with COVID-19 also being an independent predictor of all-cause mortality at both 30 days and 6-months. Therefore, there is a need to further understand the implications of having COVID-19 during NSTEMI on clinical outcomes and pathophysiology. The availability of datasets through the Ontario Health Data Platform provides a unique opportunity to merge and create a database that will allow the study to compare the effect of COVID-19 infection in the STEMI, NSTEMI populations. The proposed study will provide greater insight into the distribution, demographics, clinical presentation, disease characteristics, complications, ICU admissions, length of stay and in-hospital mortality, knowledge that is currently lacking. It will further our understanding of how patient populations differ and approaches to care. Finally, it will provide information on resource utilization and will inform future planning for pandemic treatment of patients at hospitals across Ontario. Taken together, this project will allow us to gain a better understanding of the mechanisms in MI and COVID-19 which include an elevated thrombus grade, higher prevalence of micro-thrombi, higher rates of microvascular dysfunction, and decreased TIMI flow. In addition, we will explore the associating of angiographic parameters of patients with

COVID-19 and MI to any specific pathophysiology. As such, we are also proposing a mechanistic study to assess angiograms and electrocardiograms (ECG) of hospitalized patients with COVID-19 and MI in a blinded-approach using core labs. This will be the first study of its kind and will provide the insight to the mechanism of this unique population in humans. We are proposing to use OHDP funding to help answer these remaining questions from this registry that will help inform clinical guidelines:

1. We plan to cross reference patient registries available in Ontario to assess recurrent events (myocardial infarction, stroke, mortality) since their admission to understand the demographics of this population including door to balloon time and reperfusion strategies as compared to non-COVID historical cohorts.
2. We will assess the prevalence of Long COVID/post-acute COVID-19 syndrome in those with a history of MI
3. Stratify the data by sex, gender, and ethnicity to understand how MI impacts the different groups, along with understanding the implications of COVID-19 and MI in minority populations
4. Assess mechanistic pathways to help explain the increased mortality such as analyzing angiograms of Ontario patients in core labs to understand TIMI flow, thrombus burden, and blush score
5. Assess differences in presentation such as Core Lab analysis of initial ECGs and subsequent patterns of ECG resolution.
6. Assess in-hospital and post-discharge outcomes based on history of vaccination.

Datasets: CCRS, COVaxON, DAD, NACRS, ODB, OHIP, OLIS, OLIS-C19, OMHRS, RPDB

Principal Investigator: Karen Tu

Institution: North York General Hospital

Description: We propose a portfolio of projects with a unifying theme of Timely and Appropriate Primary care for Vulnerable Populations During and Post-Pandemic. This would cover an evaluation of public health measures applied to deal with the pandemic. As we head into a new normal which struggles to deal with a shortage of healthcare providers-notably primary care providers--throughout our healthcare system, it is important to understand the resultant impact on our vulnerable populations--specifically to identify gaps in care in order to improve health planning. In our portfolio of projects we will assess the current state of vulnerable populations (mental health conditions such as people with schizophrenia, bipolar disorder, and attention-deficit/hyperactivity disorder (ADHD)), impact of halted services (pulmonary function testing), the efficacy of inhaler prescribing (budesonide - recommended in the treatment of COVID) on keeping people out of hospital or emergency rooms, and last, the impact of virtual care on primary care (a public health measure/strategy designed to reduce the risk of transmission of respiratory illnesses) in Ontario, Canada compared to eight other countries around the world through the International Consortium of Primary Care Big Data Researchers (INTRePID). Dr. Karen Tu Research Scientist at North York General Hospital, Professor in the Department of Family and Community Medicine at the University of Toronto, and Primary Care Physician will be the Portfolio Lead joined by Dr. Andrea Gershon Scientist at ICES, Professor in the Department of Medicine at the University of Toronto and Respirologist as Portfolio Co-Lead. Both Dr. Tu and Dr. Gershon have extensive experience with administrative data analysis and will work to guide a group of new to mid-career family physician researchers and a biostatistician to carry out our portfolio of projects.

This project has the following components:

1. Examining the impact of COVID-19 on patients with severe mental illness
2. Prescription and health care utilization trends for ADD/ADHD in Ontario/Canada during the COVID pandemic
3. Evaluating and predicting COVID related hospitalization outcomes among individuals with severe mental illness in Ontario
4. Pulmonary Function Testing in Ontario during the COVID-19 pandemic
5. Inhaled Corticosteroid (ICS) Puffer prescribing in Ontario during the COVID-19 pandemic
6. How has the switch to virtual care impacted the delivery of primary care in Canada and internationally?

Datasets: CCRS, COVaxON, DAD, eCTAS, NACRS, NMS, ODB, OHIP, OLIS, OLIS-C19, OMHRS, RPDB

Principal Investigator: Timothy Hanna

Institution: Queen’s University

Description: Timely intelligence on cancer diagnosis and treatment has been unavailable in Ontario and most other jurisdictions during the pandemic. Linked health administrative data provide detailed information on these aspects, though the pace of data access is unacceptably slow, and deteriorated during the COVID-19 pandemic. Decision making has often relied on anecdotes, out of date data or limited institutional reports. The OHDP provides a unique opportunity to address this. In this Ontario-wide study, we propose developing algorithms to identify cancer patients in real time, including their surgical stage and biomarker information required for risk adjustment analysis and detailed analytics. We will undertake this proof of principle work within the context of studying the impact of pandemic-associated cancer treatment delay on mortality. This is a critical subject to provincial and global cancer control, as timely diagnosis and treatment are important for optimal outcomes amongst people with cancer.

Datasets: DAD, NACRS, OHIP, OLIS-C19, RPDB

Principal Investigator: Morgan Slater

Institution: Queen’s University

Description: Long-term care homes have felt the brunt of the COVID-19 pandemic. Not only did resident health and wellness suffer during the crisis, but the sector experienced issues with staffing due to changes in governmental mandates and outbreaks among staff. The pandemic highlighted the importance of interprofessional teams to help address the profound clinical, social, and mental health care demands emerging during the pandemic, particularly amongst older adults.

We are interested in understanding the impact of the care provided by interprofessional health care providers on long-term care resident health and wellness during the COVID-19 pandemic.

The objectives of this study are:

1. What professional services (e.g., occupational therapy, physical therapy, etc) were deployed in LTC before and during the COVID-19 pandemic?
a. How is access to interprofessional services associated with resident health and wellness?

2. What are the patterns of outbreaks amongst LTC health human resources?
a. How are these staffing patterns associated with resident health and wellness?

Datasets: CCRS, CCM, RPDB

Principal Investigator: Daniel Mulder

Institution: Queen’s University

Description: Recently, there has been an alarming increase in cases of pediatric acute hepatitis of unknown origin (PA-HUO) around the world (1). More than 600 cases that meet the WHO working criteria have been identified (2,3). This phenomenon was first identified in April 2022, but it appears from surveillance data that the increase began in approximately October 2021 (1). This increase in PA-HUO includes cases in Canada, including Ontario (4). Effective treatment is lacking, as the cause is unclear. The condition can cause major morbidity (including requiring liver transplantation) and in some cases is fatal (2). Despite detailed epidemiologic and laboratory analysis, the recent increase in PA-HUO remains unexplained. No clear epidemiologic link between patients has been found - there has been no identified direct contact between patients, or even clear secondary contacts, and the patients do not have any clear common risk factor for hepatitis (1,3). PA-HUO is not a new phenomenon, although previously it seemed to occur at a much lower rate in the population (5). Many publications have hypothesized that this rare condition is caused by exaggerated response to a viral infection, given the similar clinical and histologic features to viral hepatitis (6). The attribution to viral infection is difficult to prove since all these pediatric acute hepatitis case are, by definition, of unknown origin. The possibility that PA-HUO is due to an unidentified virus is also supported by the limited testing available for most viruses, and that antigenic shift and drift can render molecular testing inaccurate. The World Health Organization has formalized a case definition (1) and postulated multiple possible causes of the recent global clusters of PA-HUO. One early theory has been adenovirus F41 infection after the blood of multiple children with PA-HUO tested positive for this particular adenovirus (3), however this virus is a common pathogen, previously known to cause limited viral gastroenteritis in the pediatric population and commonly undergoes host genome integration (7). The PA-HUO cluster patients did not have gastroenteritis (3). Thus, it seems unlikely for this virus to have gained liver tropism and simultaneously lost its ability to cause gastroenteritis. None of these potential causes has been evaluated in detail yet. A leading theory of the cause of the PA-HUO clusters, is a delayed reaction to SARS-CoV-2 infection. This theory is supported by the well described similar condition: multisystem inflammatory syndrome in children (MIS-C), which is linked to SARS-CoV-2 infection in children (8). MIS-C is another rare complication of SARS-CoV-2 infection that appears to occur mostly in young children and involves acute hepatitis as a component of the syndrome (8). Thus, PA-HUO may be a variation of MIS-C that predominantly features liver injury. Epidemiologically, SARS-CoV-2 infection also has occurred recently in large clusters worldwide and in a large proportion of the population, especially children. If SARS-CoV-2 generally does not cause major liver injury, but occasionally causes a spectrum of liver injury ranging from mild-to-severe liver injury, it is plausible that the rare cases where SARS-CoV-2 causes severe liver injury are now being identified as PA-HUO (9). Several other viral infections, such as hepatitis A and Epstein-Barr virus demonstrate a very similar pattern of liver injury (10). Thus, it would follow that the PA-HUO clusters may be a rare complication of SARS-CoV-2 infection. Another challenge in identifying a link between COVID-19 illness and PA-HUO would be temporal delay between SARS-CoV-2 exposure and PA-HUO clusters, which is notably not a usual feature of severe acute viral hepatitis. It is possible that the PA-HUO cases are a delayed autoimmune response to SARS-CoV-2 infection rather than a direct viral hepatotoxic reaction. In fact, this phenomenon of SARS-CoV-2 infection, then severe acute autoimmune hepatitis has recently been demonstrated, for the first time, in a fatal case in a child (11). Establishing a link between PA-HUO and SARS-CoV-2 is critical. Identifying new, unreported severe cases is unlikely given the severe obvious clinical effects of PA-HUO. Instead, this hypothesized link between SARS-CoV-2 and PA-HUO would be strongly supported by population level data linking liver injury with regional SARS-CoV-2 rates. Given that SARS-CoV-2 is known to inconsistently cause liver injury along a spectrum, there are likely many cases of mild-to-moderate hepatitis due to SARS-CoV-2 that could be detected in health administrative data. This is especially likely since there has been robust, broad, high-quality testing for the novel SARS-CoV-2 viral strains that have been responsible for the recent worldwide COVID-19 pandemic. The Ontario Health Data Platform (OHDP) is ideal for examining potential increases in mild-to-moderate liver injury due to SARS-CoV-2. There is robust exposure data (SARS-CoV-2), and the potential to discover unrecognized event data (hepatitis) using laboratory testing. It would follow that recent historical health administrative data would show increased rates of liver injury in regions where there had recently been increased circulating SARS-CoV-2. We hypothesize that there is an epidemiologic link between regional increases in hepatitis in relation to increased regional incidence of SARS-CoV-2 infections. We will examine the rate of acute hepatitis in Ontario health data, through bloodwork results, and compare the regional and temporal relationship of these increases with regional SARS-CoV-2 rates in the preceding months. We will use unsupervised machine learning techniques to identify patterns in this dataset that would not be discernable using classic data analysis.

Our 3 main objectives are:

Objective 1: Build a high-fidelity dataset linking clinical markers of both pediatric hepatitis and COVD-19 with regional resolution, to be extracted using the Ontario Health Data Platform (OHDP).

Objective 2: Use a spatial unsupervised machine learning approach to determine the regional variation in hepatitis and the clustered relationship to COVID cases in the corresponding region.

Objective 3: Determine clinical and demographic features that could be predictive of COVID-19-related hepatitis at a regional level, including demographic features (age, sex, socio-economic status, and rural/urban location), hepatitis features (bloodwork patterns, hospitalization), and COVID-19 illness features (symptoms, vaccination status, variant subtype).

Of note, even if we can show there to be no clear connection between hepatitis and increased regional SARS-CoV-2 infection rates, this information would be highly valuable data in assisting in the worldwide effort to better understand the cause of PA-HUO.

Datasets: COVaxON, DAD, OHIP, OLIS, OLIS-C19, RPDB

Principal Investigator: Kerstin de Wit

Institution: Queen’s University

Description: The COVID-19 pandemic has affected how doctors test for blood clots. COVID-19 increases the risk of pulmonary embolism (blood clots in the lung) by 9-fold. The type of blood clots in the lung seen with COVID-19 are different from those diagnosed before the COVID-19 pandemic. The symptoms of COVID and the symptoms of blood clots are also very similar, so frequently doctors test for both conditions together. Before the pandemic, we identified that emergency doctors misdiagnosed blood clots in the lungs in around 2 out of every 5 cases. To address this, we used Ontario administrative health data (2015 to 2019) to develop a trigger which alerts emergency doctors that their patient may be experiencing blood clots in the lung. We do not know how well this alert works during the pandemic years.

The objectives of this study are:

1. To assess the predictive ability of our emergency department blood clot alert during the pandemic years (2020 to 2022)
2. To review how well the alert distinguishes between COVID-19 and blood clots
3. To determine how well the alert performs in patients with COVID-19 infection. If required, we will improve the alert so that it functions maximally to identify cases of blood clots in the lung in the era of COVID-19.

This protocol addresses the Rapid Learning Health Systems priority. We used machine learning to create this blood clot alert. We will use all available data on Ontario emergency department visits since the pandemic started to learn how to tell the difference between COVID-19 infection and blood clots, and to identify which patients with COVID-19 infection also have blood clots. Our research will produce an emergency physician alert which will accurately identify patients with blood clots during the COVID-19 pandemic, with the potential for ongoing real-time data collection post-implementation, in order to achieve continued improvement.

Datasets: COVaxON, DAD, eCTAS, NACRS, ODB, OHIP, OLIS, OLIS-C19, RPDB

Principal Investigator: Zihang Lu

Institution: Queen’s University

Description: A fundamental tenet of population health is that there is an association between socioeconomic status (SES) and the health of individuals mediated through material deprivation as well as through social deprivation. This health-deprivation gradient provides an important lens for assessing the impact of major health risks and public health policy. This project will use that lens to assess the impact of the COVID epidemic and the public health policies to contain COVID on SES-related health disparities. The resulting assessment will be provided to policymakers in a timely and relevant manner in order to inform an equitable policy response to this public health crisis.

The project team includes experts in health services research, epidemiology, public health, health disparities, economics, statistics and education who will use sophisticated modelling and analytical techniques and timely access to data on health outcomes, SES and employment to investigate the cross-SES health impacts of the COVID-19 pandemic, and of the containment measures used to mitigate that pandemic in Ontario.

The objectives are to:

1. Measure the set of defined health outcomes across SES strata before and after March 2020 using best available data;
2. Model observed impacts of COVID and containment on a defined set of health outcomes across SES strata using sophisticated statistical techniques.

Datasets: CCRS, COVaxON, DAD, NACRS, ODB, OHIP, OMHRS, PCCF, RPDB

Principal Investigator: Steven Hawken

Institution: Queen’s University

Description: Compared to uptake of initial COVID-19 vaccine doses, uptake of subsequent bivalent booster doses has been relatively low. Although efforts to increase bivalent booster vaccine uptake are ongoing, little research has been conducted into the factors that drive booster vaccine uptake. This study aims to understand the coverage rates for bivalent COVID-19 vaccines and what factors impact vaccine uptake among individuals in Ontario. To answer these questions, we will draw on data holdings from the Ontario Health Data Platform at Queen's University (OHDP-Q). Vaccine coverage will be determined overall and by key subgroups. This information can be used to support targeted, efficient vaccination campaigns aimed at improving coverage in areas and among populations who are less likely to have been vaccinated. These campaigns can help to reduce the risk of outbreaks in under-vaccinated populations, reducing the risk of morbidity and mortality among these groups. In addition, this research will help to identify gaps in vaccination data that may limit our ability and the ability of other stakeholders to analyse vaccination trends in Ontario.

Datasets: COVaxON, OHIP, RPDB