New data provide insight into pandemic inequalities

Using the 2022 Statistics Canada dataset “Impacts of COVID-19 on Canadians – Testing and Vaccination,” our analyses found that variable vaccination rates were not the reason behind differing rates of infection. (Pexels Photo)

Existing analyses of COVID-19 in Canada and internationally suggest infection rates were highly variable across populations, with researchers highlighting the disproportionate burden experienced by groups that are intersectionally disadvantaged.

Early in the pandemic, widespread evidence emerged of the unequal rates of infection experienced by residents and care workers in long-term care homes in Canada. Soon after, racialized populations and immigrants were also found to be especially hard-hit. The COVID-19 mortality rate was significantly higher for racialized populations.

Our new research finds that while these individuals had higher rates of COVID-19 infection, they were equally or more likely to get vaccinated than comparison groups. This has important policy implications. Rather than focusing on individual decision-making, the data suggest a need to prioritize protections and pay for workers in essential jobs that are unpredictable, dangerous, physically demanding and/or low wage.

Overwhelming evidence shows that vaccines are an effective protective measure, both in terms of reducing infection rates and severity of illness. Using the 2022 Statistics Canada dataset “Impacts of COVID-19 on Canadians – Testing and Vaccination,” our analyses found that variable vaccination rates were not the reason behind differing rates of infection.

These data were collected as part of Statistics Canada’s crowdsourcing initiative which aimed to gather timely information on Canadians’ experiences of testing for COVID-19 and access to vaccination during the pandemic. Statistics Canada used open advertising to obtain participants who chose to self-select by completing an online questionnaire from February 21 to March 13, 2022.

For racialized minorities, 20 per cent tested positive, compared to 15 per cent for non-racialized minorities (outcomes were similar when comparing PCR to rapid test results). In particular, those who identified as Black and Filipino (populations that also have high rates of employment in care occupations and in service industries fared worst. Black populations experienced a one-third rate of positivity for rapid tests (see Figure 1).

Citizenship was also an important factor. Those without Canadian citizenship status (including temporary foreign workers) fared worse for both types of testing (at an average of 22 per cent positivity), compared to those immigrants with citizenship status and the Canadian-born. The latter fared by far the best, at 14 per cent positivity via PCR tests, and 17 per cent positivity via rapid tests (see Figure 2).

Women and men, interestingly, were found to have had similar rates of infection. However, those who identified as essential workers had higher rates of infection (see Figure 3), with women overrepresented among this population.

Yet the data on rates of vaccination tell a different story. This suggests that variable infection rates were tied to specific jobs or to systemic inequalities, rather than a function of individual choice.

There was only a marginal difference in vaccination rates between essential workers and non-essential workers. Racialized people and immigrants had higher rates of vaccination than non-racialized people and non-immigrants (see Figure 4).

Finally, the data reject explanations tied to access to health facilities. In the five cities in Canada with the highest rates of infection, residents had higher rates of access to at least one health facility and a pharmacy close to them (see Figure 5).

Together, this information suggests that COVID-19 infection rates were not related to personal decision-making or access to health services. Instead, it raises concerns about broader social responsibilities. Populations that are racialized and/or non-citizens and those doing essential jobs were infected at disproportionate rates, even as they took steps to get vaccinated.

Notably, the data do not provide information on time order, so it is plausible that those workers on the front lines reported positive cases prior to having access to the vaccine. This raises further questions about which occupations and which areas of each city were prioritized for personal protective equipment and early access to vaccination, while refuting suggestions of a lack of awareness or interest in vaccination by underserved communities.

As well, the data do not indicate the severity of illness. Thus, the analyses may tell a story of lives saved due to vaccination for populations that were at greater risk of infection due to their workplace conditions.

In terms of policy, this suggests that there is an urgent need to focus on improving working conditions for essential workers. This includes providing paid sick leave and job guarantees for those who take time off work to care for themselves or others. Staffing levels and accommodations to work from home for those who are sick (when possible) should be ensured. Accommodations to reduce workplace injuries and increase mental health supports must be made available. Providing safe transportation to and from work facilities as well as paid time to travel to get vaccinated should be a policy initiative. Finally, emergency housing for self-isolation when needed and access to child and elder care, as well as affordable housing for those with lower socioeconomic status, should also be top priority.

Perhaps most critically, wages must be raised for workers who are systematically at higher risk of infection due to the face-to-face nature of their employment (while also structuring workplaces to avoid dangerous work conditions where possible.) This would at least partially compensate for increased rates of infection within certain jobs and recognize the toll it takes on individuals, families and communities.

All this reinforces a structural rather than individual analysis of disease burdens and public health measures. Clearly, getting vaccinated is not the end of the story. Ongoing social responsibility is needed to protect vulnerable groups – whether this takes the form of masking, work-from-home measures, or other workplace accommodations.

The implications of the pandemic are nowhere near over. There must be a move beyond rhetoric of “gratitude” to essential workers; instead, governments and employers must implement better policies and pay. These policies must be evidence-based so Canadians have an accurate understanding about the pandemic and its effects.

Given the emerging information about the potentially long-term implications of COVID-19, these measures take on added urgency for groups such as children, people with disabilities and intersectionally disadvantaged populations.

A note about methodology

Statistics Canada started the crowdsourcing initiative as part of a data collection series to address the informational needs of Canadians and enhance their understanding related to the impact of COVID-19. According to an evaluation report, Statistics Canada’s crowdsourcing products have proven to be useful for briefing purposes, policy research and analysis, program and service planning and decision-making, as well as knowledge-production and modeling. Statistics Canada uses a number of measures to ensure the quality of data collected through the crowdsourcing initiative. First, the questionnaires are designed based on Statistics Canada’s standard practices and wording used in a computer-assisted interviewing environment. During data collection, a computer application is used to automatically control the flow of questions, depending on participant responses, and to check for logical inconsistencies and errors in participant responses. The computer application used for these purposes is tested extensively. After data have been collected, Statistics Canada maximizes the quality of crowdsourced data through error detection of invalid or missing values for age, gender and postal code at micro levels. Furthermore, Statistics Canada compensates for overrepresentation and underrepresentation by calculating a benchmark factor for every participant based on demographic projections of the number of people by province and territory, sex and age group. As recommended by Statistics Canada, therefore, the authors of this article used the benchmark factor to produce their results in the same way that survey weights are used to produce estimates from non-crowdsourced data.

This article first appeared on Policy Options and is republished here under a Creative Commons license.