An equity-focused comparison of the COVID-19 vaccine rollout plans across six Canadian provinces

Researcher Characteristics

The research team, consisting of graduate students and research mentors from the UofT, was established through the Dalla Lana School of Public Health (DLSPH) Implementation Science Trainee Cluster Program at the UofT. The student researchers consisted of six doctoral students (four females: KB, MO, MS, VA; two males: CD, DMV) specializing in epidemiology, health services research, or public health sciences. They were guided by three female mentors (AA, SA, SF) with experience in implementation science, population health and immunity, health systems and policies, health system performance, and health equity. Representing diverse ethnicities and lived experiences, this multidisciplinary team brought a wealth of expertise to the study, including in implementation science, health equity research, infectious disease outbreak investigations, and vaccine science and policy. All team members had prior experience in qualitative research, including conducting structured or semi-structured interviews and interview-based questionnaires with various stakeholders such as key informants, community leaders/members, or patients. Additionally, several team members specialized in community-based, participatory action research, particularly among marginalized or vulnerable populations.

Study Population and Context

We analysed prioritization plans for five key populations: Black communities, First Nations, Inuit, and Métis populations (FNIM), non-medical essential workers, individuals experiencing homelessness, and individuals with disabilities across six Canadian provinces (Alberta, British Columbia, Nova Scotia, Manitoba, Ontario, and Quebec). Our five population groups were selected because they were prioritized by NACI as being at substantial risk of transmission and outbreaks, severe illness, and/or barriers to vaccine uptake²¹. We narrowed our study to the above six provinces after a preliminary environmental scan our team conducted of all Canadian jurisdictions. We concluded that these provinces had the most publicly available data on published prioritization plans and implementation timelines. Additionally, they provided a representative snapshot of the diversity of Canadian jurisdictions, in terms of population size and constitution, resources, and geography.

We also investigated how subgroups within our five populations of interest were prioritized, as our pilot scans indicated that there were differences in the extent to which these groups were prioritized across provinces. First, we defined Black communities as including people of African or Caribbean ancestry who live in Canada. Second, we examined prioritization for on- and off-reserve, urban, and remote FNIM populations. For essential workers, we narrowed our focus to only examining workers in non-healthcare settings because our preliminary scans indicated that essential workers in health-related settings (e.g. healthcare workers, staff in congregate settings, etc.) were prioritized first across all provinces. Additionally, we were interested in how provinces defined essential workers in their rollout plans as part of our analysis. This provided further context for how and why prioritization decisions were made for this diverse group. For individuals experiencing homelessness, we considered both those who lived in shelters (or individuals in second-stage housing) as well as those with precarious living conditions (e.g. not living in shelters, living on the streets, experiencing unstable housing, etc.). Finally, among individuals with disabilities, we focused on individuals facing significant barriers to vaccination due to either developmental (e.g., severe developmental delays, individuals in institutionalized care) or physical challenges (e.g., homebound individuals). Overall, we wanted to keep in mind our key study objective of studying populations at high risk of negative outcomes and/or barriers to vaccine access.

Data Collection

We used a two-step, mixed-methods approach to assess provincial efforts to reach priority populations. We first conducted environmental scans of the six provinces between December 1st, 2020, and May 31st, 2021, followed by key informant (KI) interviews between December 1st, 2021, and April 30th, 2022. The study period began when emergency vaccines were approved in Canada (Dec 2020) and ended just before all provinces reached their general population vaccination phase (around June 2021), as it was assumed that all high-risk populations requiring early prioritization would have been reached by that point. This period was selected to capture initial provincial decisions that were made about which populations to prioritize and strategies that were implemented for the primary dose series (first and second doses), in response to significant vaccine shortages and delays at the beginning of the rollout.

Environmental Scans

For environmental scans, we reviewed provincial reports, statements, and scientific or media articles, released between December 2020 and May 2021, related to vaccine prioritization plans for our five key populations. We intentionally focused on provincial government-led efforts to reach priority populations, excluding community or grassroots initiatives that addressed gaps in provincial-led outreach, such as those facilitated by organizations like the Health Association of African Canadians in Nova Scotia^33,34, unless provinces intentionally collaborated with these community groups as part of their pre-determined strategy to prioritize and engage with communities (e.g. Black churches in Nova Scotia³⁵). For the Environmental Scan Data Collection Tool used, please consult S2 Appendix.

The extracted data was used to generate a single summary for each province detailing their COVID-19 vaccine prioritization plans for our priority populations. For each key population, the summaries detailed the extent to which they were prioritized (if it all), along with the timeline and methodologies employed for vaccine implementation. Additionally, the summaries included justifications listed on provincial websites that would indicate how decision-makers rationalized their choices.

Environmental scans were conducted in two stages. Preliminary scans were conducted before KI interviews, and the initial findings were shared with KIs for validation, correction, and additional information. After incorporating feedback from the KIs (see below), a final environmental scan summary was compiled, incorporating any new data released since the initial scan and contextual information or corrections provided by KIs.

Key Informant Interviews

To contextualize and validate our environmental scan findings, we conducted KI interviews with public health experts involved in the vaccine rollout across provinces between December 2021 and April 2022. The research team invited a total of 89 KIs from the six study provinces. KI interviews were carried out until reaching saturation, comprising 31 interviews with 41 participants. Generally, interviews were conducted with individual KIs, unless they specifically requested a group interview. Of the 31 interviews conducted, 25 were individual sessions and six took place in groups of 2 to 4 individuals. To ensure comprehensive representation for each province, we conducted a minimum of 4 interviews per province. Two KIs later decided to withdraw their interviews from the study over concerns of being identified.

KIs chosen for our study held diverse roles in public health, including medical officers of health, vaccine task force leads, physicians, members of pandemic response teams, medical anthropologists, public health experts, and representatives from NACI. Of the 31 interviews, 27 (87%) represented academic, government, or research institutions, while 13% were affiliated with community organizations. Notably, approximately 90% of participants had prior experience with communicable disease control and vaccine delivery programs before the COVID-19 pandemic.

KIs were selected through a comprehensive review of publicly available information on key decision makers and stakeholders involved in COVID-19 vaccine rollout committees, working groups, advisory panels, task forces, as well as through the social networks of the team. Two members of the research team (MO and CD) sent invitation emails to selected individuals in stages, with up to two follow-up emails when necessary. A group of informants received the initial invitation; if there were no responses, additional potential informants from the respective province were contacted. Interested informants received an information sheet detailing the study’s objectives along with a consent form. Participation incentives were not provided. Those who accepted the invitation were scheduled for an online interview using Zoom.

Prior to the interviews, participants were asked to review our environmental scan summary of the COVID-19 vaccine rollout plan for their respective province. During the interview, KIs were asked to verify the accuracy of the findings from the environmental scan. They were also given the opportunity to offer any corrections or supplementary information during the interview or afterward via email. At the end of each interview, KIs were asked to recommend any potential informants that they believed could contribute to the issues discussed in the study. Those referred were then contacted by either the study team or directly by the referring KI, similar to a snowball sampling method of recruiting participants³⁶.

Interviews were semi-structured, following an interview guide but allowing interviewers to ask probing questions for more information if needed (Refer to S3 Appendix for the Interview Guide). The interview guide was pilot tested by research team members. Interviews were scheduled for 1.5 hours each and generally took the full time. KI interviews were primarily conducted by two members of the research team (MS and VA) with support from two team members (KB and CD) on two separate occasions due to scheduling conflicts. Researchers conducted interviews from private rooms with KIs to ensure confidentiality. All interviews were conducted over Zoom and were audio recorded after expressly seeking participant consent.

Preliminary transcription was completed by the Zoom platform. Transcripts were further reviewed by all student research members. Within two weeks of each interview, KIs received the cleaned transcript for review and were given the opportunity to respond via email with any concerns they had regarding the accuracy of captured information.

Data Analysis

Our analysis was guided by the “Reach” component of the RE-AIM implementation science framework³⁷. Using this framework, we were able to implement a content analysis approach of extracting data on the prioritization of populations and the strategies employed to promote COVID-19 vaccine uptake. The “Reach” component of RE-AIM framework assesses the extent to which an intervention or program effectively reaches its intended target population and achieves broad reach among the individuals or groups it aims to impact³⁸. Although the RE-AIM framework has four other components (effectiveness, adoption, implementation, and maintenance), we determined that “Reach” was the only component that would sufficiently address our project goals during our study time period, which was very early on in the rollout. We conceptualized our study around September 2020, prior to the first COVID-19 vaccines being licensed (December 2020). NACI guidelines on reaching priority populations and provincial vaccine rollout plans were just starting to get published. We concluded that it would be premature to utilize the full RE-AIM framework since it would likely take several months, if not a few years, to determine the effectiveness of these policy plans, how well they were adopted and implemented, and whether they were successful enough to incorporate into regular policy practices over time.

Guided by “Reach”, we focused on the first two of four indicator questions (Table 1) to determine the following for each province: (a) if and when each of the five key populations was prioritized and (b) how provinces justified their prioritization decisions to the general public. Obtaining answers to these study objectives drove how we structured our environmental scans as well as the themes analysed from the KI interviews.

To analyse KI interview data, we first employed an interpretive descriptive qualitative approach³⁹. Transcripts were managed using NVivo 12 software and independently analysed by members of the research team (KB, MO, CD and DMV) through inductive thematic analysis. This inductive approach allowed for a deep exploration of the context surrounding vaccination strategies and implementation, facilitating the identification of patterns, categories, or themes not previously considered^39,40, thereby enhancing the study’s rigor by ensuring the findings were grounded in the data³⁹.

Coding followed Creswell’s approach^41,42, wherein each transcript was initially read, and key ideas or responses relevant to the study objectives were independently highlighted by two team members. These ideas were then assigned a “code,” with operational definitions ensuring consistent application across the data. Similar codes were grouped to form broader themes, serving as the main units of analysis⁴¹. Summaries were then developed for each theme by reviewing all coded data, a process completed by two team members (KB and MO), with revisions and consensus on themes achieved through regular meetings and discussions among team members. Any discrepancies in coding, categorization, or definitions were discussed, with consensus sought throughout the analysis process.

Objective 1: Compare justification and timelines used by the provinces for prioritization of key populations

Overall provincial approaches to population prioritization and ethical frameworks used in designing vaccine rollout plans Table 2 summarizes each province’s overall public health approach in determining priority populations and any ethical frameworks identified. KIs confirmed that all provinces consulted NACI guidelines, including NACI’s EEFA framework, to guide their decisions regarding priority populations, vaccine delivery, and storage. According to our environmental scans and KI interviews, provinces utilised a combination of age-, risk-, and health status-based approaches to select priority populations, in accordance with NACI guidelines. How provinces defined risk varied depending on their local contexts, ranging from protecting those at risk of developing severe illness, minimizing transmission in workplaces or in remote regions, or maintaining the function of critical provincial services.

Table 2 also summarizes notable ways in which provinces deviated from NACI guidelines based on their local contexts. First, four provinces (Alberta, British Columbia, Manitoba, and Quebec) did not directly prioritize Black communities in their vaccine rollout policies, despite being specifically mentioned by NACI as a racialized group at higher risk of developing severe outcomes due to systemic barriers to accessing healthcare²⁷. Justifications provided by provinces for this choice are discussed in the next section. Second, multiple provinces (Alberta, British Columbia, Manitoba, and Ontario) employed a “hot spot” strategy for regional prioritization to reach key demographics that could not be easily captured through other strategies, a method that was not mentioned by NACI. Third, Quebec and Manitoba deviated from NACI-recommended dosing intervals due to provincial considerations regarding how to balance prioritizing those most at risk with reaching the most people amidst early vaccine shortages⁴⁸. Quebec’s strategy was later adopted by NACI in March 2021 because of increasing scientific evidence of the usefulness of this strategy, particularly during the early vaccine shortages from January to March 2021^49,50. Fourth, Nova Scotia made an early decision to strictly adhere to a primarily age-based rollout ^51–53, with very few notable exceptions that will be discussed later. While age-based prioritization was consistent with NACI recommendations, Nova Scotia deviated from NACI by intentionally not prioritizing certain populations recommended in NACI’s equity matrix²¹. This included individuals with underlying medical conditions, individuals with disabilities, as well as certain groups of essential workers^25,27,52,54. Provincial justifications for this decision will be discussed later. Finally, many provinces employed a phased vaccine rollout strategy, as was consistent with NACI guidelines²⁷, while Quebec used a similar rank-based system of prioritization (Fig 1). However, Manitoba used what it referred to as a “full immunization fast” policy rather than a phased approach⁴⁵, where eligibility was based on vaccine availability. Eligibility was communicated to the public through regular COVID-19 Vaccine Bulletins and news releases.

• Download figure
• Open in new tab

Fig 1.

Timeline of prioritization of five key populations across six provinces. *, **, ***: Detailed list of essential worker categories is listed in Table 4. Abbreviations: FN = First Nations; FNIM = First Nations, Inuit, and Métis; intellect. = intellectual; dev. = developmental; CLdS = Community Living disABILITY Services

Justifications and timelines provided by the provinces for prioritization of key populations: According to our environmental scans and KIs, prioritization of our study populations varied between provinces. For this section, Fig 1 provides detailed timelines for prioritization by province and study population. Table 3 provides a summary of justifications provided by provinces for their prioritization choices.

Black communities

Black communities were the least prioritized of our five study populations, despite real-time news reports from multiple provinces showing that these communities had been disproportionately impacted by COVID-19^55–58. A common justification provided by KIs for these provinces was that race-based data was not widely collected in Canada⁵⁹, which limited crucial real-time evidence gathering on high-risk communities and regions.

Of the six provinces, only Nova Scotia and Ontario clearly prioritized Black communities in published policy plans. They were one of very few exceptions to Nova Scotia’s strict age-based prioritization strategy. Both the scans and our KIs stated that the provincial government’s generations-long partnership building with African Nova Scotian communities was critical to successful outreach⁶⁰. Early discussions with African Nova Scotian community leaders highlighted the need for immediate and extensive prioritization of Black communities due to a history of racism and the resulting lack of community trust in healthcare systems^35,61.

Although race-based data was not routinely collected in Ontario during the study period, the province prioritized Black populations living in “hot spot” communities (determined by postal code) for vaccination. Black scientists and medical experts working directly with the provincial government and the City of Toronto strongly supported this decision by sharing real-time data indicating that these “hot spot” communities were experiencing the highest rates transmission, hospitalization, and death due to COVID-19^58,62,63. They further stated that vaccine hesitancy was highest among Black people of African and Caribbean origins, data consistent with NACI findings^27,58.However, KIs for Ontario were unclear whether Black populations living in non-“hot spot” regions were prioritized for vaccination, highlighting this as a potential gap in prioritization. One KI noted the following:“It is not just Black people in low socio-economic circumstances or hotspot communities who are at risk. Black people in middle income suburbs are also at risk because of the systemic racism that weathers the black body and creates patterns of chronic conditions amongst Black people across Canada.” -Key Informant from Ontario (ID 827)

While the scans did not reveal justifications through published policy documents, KIs for the other four study provinces provided insights into why Black communities were not listed in their rollout plans. KIs for Alberta stated that race-based data was not available to justify a clear need for prioritization. KIs from British Colombia stated that the province’s Black populations comprised only approximately 1% of the overall population⁶⁴. KIs for Quebec asserted that not all Black communities required early prioritization and considered them to be too heterogenous in terms of risk to capture by race alone.

According to KIs, these provinces used an intersectional approach to attempt to reach Black communities. Alberta prioritized groups that had higher risks of severe outcomes due to age or underlying medical conditions, as well those experiencing a higher risk of transmission exposure (e.g. essential workers). Quebec targeted refugees, immigrants, and essential workers for vaccination because these groups demonstrated high rates of hospitalizations and deaths.

British Colombia prioritized geographic areas with higher percentages of immigrants and refugees. Finally, KIs for Manitoba did not provide a rationale for why they did not prioritize their Black populations, but they stressed that the provincial focused on racialized communities more generally as an attempt to reach their Black communities.

First Nations, Inuit, and Métis (FNIM) populations

Environmental scans and KI interviews showed that, of our study populations, the FNIM were prioritized earliest in all provinces, which was consistent with NACI guidelines^25–27 (Fig 1). The most common justifications provided for prioritizing FNIM populations are listed in Table 3.

Provinces employed some interesting prioritization strategies to accelerate vaccination of FNIM populations, given their respective local contexts. A few provinces (British Columbia, Manitoba, Ontario) prioritized FNIM age groups that were 15-20 years lower, on average, than that of the general population at any given point in time. KIs cited published data showing that the life expectancy for FNIM was significantly lower (e.g. up to a 17-year difference in some studies) than that of the non-Indigenous population^65,66. KIs for Manitoba shared that the province did not directly prioritize remote and isolated communities in their rollout policy, which deviated from NACI guidelines²⁶, but instead attempted to capture them indirectly through community-based “hot spot” eligibility. In Nova Scotia, the government prioritized the Mi’kmaq First Nation people first in another exception to their age-based prioritization rule. KIs explained that the Mi’kmaq were the predominant Indigenous population in Nova Scotia. By targeting vaccination efforts towards this population, they hoped to reach both remote and urban communities early in the rollout. They also confirmed that any individual who identified as FNIM but who was not Mi’kmaw would have to follow the standard age-based eligibility criteria implemented for the rest of the province.

Gaps in prioritization

KIs for Alberta and Manitoba shared that excellent real-time scientific evidence existed for First Nations communities (and for the Métis peoples, in the case of Alberta), clearly demonstrating that that these communities experienced significant harms related to COVID-19. This data served as critical justification for prioritizing these populations. However, similar data-sharing agreements were not in place for the Inuit in Alberta, and for both the Métis and Inuit in Manitoba. KIs stated that, unfortunately, there was insufficient risk data regarding these populations for decisionmakers to justify early prioritization without it appearing as preferential treatment. As a result, these communities were not prioritized until several months into the rollout, after reviewing real-time data from other provinces, which impeded vaccination progress at a critical stage of the rollout (see “Limitations of Data Availability” section below).

NACI guidelines recommended prioritizing urban FNIM for vaccination, as “systemic barriers to accessing necessary supportive care for COVID-19 also exist in urban settings related to factors such as poverty, systemic racism, and homelessness”²⁶. Yet, many provinces appeared to rely mainly on age-based prioritization of all FNIM adults to capture this community, rather than on directly prioritizing this important subgroup. KIs for these provinces justified their choice by stating that they deemed risk to be higher in more remote territories than in intermediate or urban communities. Other reasons given for prioritizing remote over urban settings were: (1) the higher likelihood of severe health outcomes because of underlying medical conditions and poor housing conditions in remote settings; (2) a stronger risk of uncontrolled transmission or outbreaks because of challenges in enforcing isolation measures; and (3) reduced access to specialized healthcare services if infected. Urban FNIM in Ontario were later added to the rollout plan after strong advocacy from local FNIM community leaders^67,68 (see “Limitations of Data Availability” section).

Essential workers in non-healthcare settings

All provinces prioritized essential workers in non-health related occupations in the second phase of the vaccine rollout (around April 2021) after healthcare workers, which was consistent with NACI guidelines²⁶ (Fig 1). KIs for the provinces frequently stated that this timeline was also necessary because of the severe, country-wide vaccine shortages between January and March 2021. The scarcity and delivery delays constrained provinces to prioritize vaccinating only their highest risk groups during those months⁶⁹.

Essential workers, whether referred to as “critical workers” (Manitoba) or “priority front-line workers” (British Columbia), referred to a similar category of workers whose services could not be completed from home. Provinces adapted essential worker categories suggested by NACI to more accurately reflect critical services provided in their local contexts (Table 4). Provinces generally agreed with the NACI assertion that essential workers, who often worked in crowded conditions where public health safety protocols could not be maintained, had an increased risk of COVID-19 exposure, and therefore should be prioritized^27,70. Further justifications for prioritization are summarized in Table 3.

Provinces deviated from NACI guidelines in some notable ways due to local contextual factors. Nova Scotia made another exception to its age-based prioritization strategy by prioritizing essential workers who were not able to maintain public health guidelines effectively because of the nature of their occupations⁵². However, certain NACI-recommended essential worker categories were not included or were removed from prioritization (Table 4). According to KIs, decisionmakers assessed that these subgroups would be vaccinated more efficiently through age-based prioritization because of the province’s small population size. In Quebec, grocery store workers and teachers were not prioritized (Table 4). KIs explained that, due to limited vaccine supply, Quebec prioritized older age groups followed by individuals at higher risk of severe disease if infected. KIs indicated that decisionmakers found it difficult to justify prioritizing occupations with younger average working ages, as they would likely have better outcomes if infected over older individuals or those with underlying medical conditions.

Individuals experiencing homelessness

As mentioned earlier, we were interested in how provinces prioritized shelter residents as well as those with unstable housing situations. All provinces prioritized shelter residents in the second phase of their vaccine rollout plans (Fig 1), consistent with NACI guidelines²⁷, with justifications provided in Table 3. Notably, Quebec did not initially prioritize individuals experiencing homelessness, but later included them in their revised policy plan released in March 2021. KIs reported that night curfews implemented across the province to curb transmission led to the unintended consequence of outbreaks in congregate settings such as shelters. KIs for Nova Scotia confirmed that prioritization was based on age of shelter residents, as was consistent with the general provincial strategy.

Gaps in prioritization

By contrast, individuals experiencing homelessness but not living in shelters were only prioritized in two provinces: Alberta and Ontario (Fig 1). KIs for Alberta confirmed that they intentionally defined individuals experiencing homelessness as ‘transient populations’ in their published policy plan to include both shelter and non-shelter residents. However, it remains unclear how well this was communicated to the public, as this definition was not explicitly stated in their rollout plan. Ontario explicitly mentioned prioritizing individuals living in encampments in Phase 2 (April 2021)⁷⁰. KIs for other provinces did not provide rationales for overlooking this key subgroup, aside from admitting that it was a “hard-to-reach” population.

Individuals with disabilities

As mentioned earlier, we focused on individuals facing significant barriers to vaccination, which included those with developmental disabilities and those who were homebound with a disability. Of all our study populations, individuals with disabilities had the most variability in prioritization across provinces. Based on our findings, four provinces prioritized individuals with disabilities in their original vaccine rollout plans (Alberta, British Columbia, Manitoba, Ontario), adopting a tiered approach based on severity levels within this population (highest risk, high risk, and at-risk)^72–75. British Columbia followed NACI guidelines most closely by vaccinating those in institutionalized care in Phase 2, followed by those who were clinically extremely vulnerable due to their disability in Phase 3⁷⁶ (Fig 1). KIs for these provinces highlighted the importance of real-time risk data in informing their early prioritization strategy, revealing higher rates of hospitalizations, ICU admissions, death, and other severe outcomes among individuals with developmental delays or intellectual disabilities compared to the general population^77,78 (Table 3). In fact, despite being initially excluded from prioritization, Quebec later revised its rollout plan to include individuals with disabilities in April 2021 after reviewing real-time risk data⁷⁹ (see “Data Informed Prioritization Strategies” below).

Notably, Nova Scotia intentionally did not prioritize individuals with disabilities, despite evidence that Nova Scotia had a higher percentage of individuals living with a disability compared to other provinces^54,80. KIs from Nova Scotia acknowledged the heightened risk experienced by those with disabilities but once again cited that the province assessed that its age-based rollout strategy would be more efficient at reaching this population than if they were to prioritize those with underlying medical conditions.

Gaps in prioritization

Among provinces that prioritized individuals with disabilities, those who were homebound were often prioritized later, if at all, compared to those in institutionalized settings. Information on prioritization of homebound residents for vaccination was not easily available publicly, but KIs for all provinces, except Alberta, provided some useful insights. For instance, homebound individuals in Manitoba began receiving prioritization in June 2021 through the Community Living Disability Services (CLDS), which provides services for individuals with both mental and physical disabilities in the province. It was, however, unclear if homebound individuals not eligible for CLDS received prioritization or transportation services. Ontario began offering services to assist homebound individuals (both at-home services as well as transporting them to vaccination sites) in Summer 2021, and Quebec included homebound individuals in their revised plan in April 2021. Although Nova Scotia did not specifically prioritize individuals with disabilities, KIs verified that government-initiated home vaccinations for individuals with mobility issues became available in April 2021. Finally, British Columbia offered transportation services for homebound individuals upon request, although no specific timeline was provided.

Objective 2: Identify contextual factors that influenced how provinces selected priority populations for early COVID-19 vaccine access

The prioritization of vaccines varied across provinces due to contextual factors specific to each region. Factors such as population demographics and geography led to the realization that no single approach, including those recommended by NACI, could be uniformly applied across all provinces. Through KI interviews, three factors emerged as significant drivers for the selection of priority populations and the corresponding timelines for early COVID-19 vaccine distribution: Data Availability, Population Size, and Geography.

Data-Informed Prioritization Strategies

The scarcity of vaccines and the severe health risks posed by COVID-19⁸¹ led to public concerns and different groups requesting to be prioritized for vaccination^82–85. KI interviews and our environmental scans revealed that decision makers faced immense pressure when prioritizing certain groups over others^84–90. Data gave decision makers a seemingly objective and scientific tool for answering questions such as identifying individuals most likely to contract COVID-19, those facing barriers to vaccine accessibility, and those at higher risk of severe outcomes.

For instance, Ontario employed a data-driven approach by prioritizing individuals residing in geographical “hot spot” communities, identified through historic and ongoing infectious disease trends, as previously described⁹⁸. The province aimed to lower the disproportionate impacts of COVID-19 on racialized urban populations, individuals living in multigenerational households, and essential workers in non-healthcare settings through this approach ^91,92.

Data-driven approaches also influenced the prioritization of First Nations groups across several provinces. In Manitoba, for example, a KI highlighted that previous publications on the disproportionate negative health impacts of the 2009 H1N1 Swine Flu pandemic on First Nations^93,94 were leveraged by the First Nations Health and Social Secretariat of Manitoba to successfully highlight the disproportionate risk this community faced for negative COVID-19 outcomes. This knowledge informed the province’s decision to prioritize First Nations groups in early February, one of the earliest prioritized groups in Manitoba (Fig 1).

In Quebec, the decision to prioritize vaccination of individuals with disabilities evolved during the rollout as a result of real-time scientific evidence. This group was not listed in the province’s initial rollout plan released in January 2021 because Quebec’s Ministry of Health and Social Services concluded that there was insufficient evidence of an increased risk of COVID-19 infection for this population^95–97. KIs from Quebec verified this sentiment, explaining that the Ministry initially found no evidence that vaccinating early would provide a noticeable benefit.

However, real-time scientific data that severe outcomes among individuals with developmental delays or intellectual disabilities were higher compared to the general population proved instrumental in changing the province’s stance on prioritization^77,78. Quebec later revised its plan to prioritize both individuals with disabilities living in group homes in March 2021 and those living at home with families in April 2021⁹⁸.

Limitations of Data Availability

Data availability was, however, not equitable across all population groups in Canada. The quality and quantity of data varied across population groups, leading to inequities in identifying who required prioritization. Inconsistent data availability for Inuit and Métis populations in Canada meant that these groups were not always prioritized when First Nations were, despite knowledge that they experienced similarly high risk for barriers to vaccine access and negative health outcomes⁹⁹. In Manitoba, while First Nations were prioritized as early as February 2021, Inuit and Métis populations were not, despite experiencing similar disproportionate health consequences during the pandemic⁹⁹. Key informants from other provinces reported similar variations in prioritization for First Nation, Métis, and Inuit populations.We had wonderful data demonstrating that First Nations people in Manitoba were experiencing more significant harms related to COVID and at a much younger age. We knew, for example, that First Nations people who died of COVID were dying, on average, 17 years earlier than non-First Nations people. That allowed us very early on to advocate to government to have a 20-year age differential when it came to eligibility for the vaccine. -Key Informant from Manitoba (ID 698)

Key informants revealed that First Nations organizations in Manitoba had established data-sharing agreements and structures before the COVID-19 pandemic with the Manitoba government. These agreements facilitated data access for the province, through the lens of First Nations Principles of ownership, control, access, and possession (OCAP). OCAP are a set of standards that ensure First Nations have authority over how their data is collected, used, and shared, emphasizing self-determination and data sovereignty. These pre-established agreements and structures allowed for data availability that facilitated vaccine prioritization for First Nation groups. However, similar processes were not in place for Métis data in Manitoba. Provincial representatives presented the lack of provincial data on Métis populations as a reason why Métis and Inuit populations were not prioritized at the same time as First Nations groups. After much advocacy from Métis groups, Métis and Inuit populations were eventually prioritized in May of 2021¹⁰⁰.… the province said there wasn’t sufficient data to suggest that Métis and Inuit were experiencing significantly worse outcomes in terms of COVID infection and outcomes associated with COVID for them to be included in the prioritization. So, for Inuit, it was, “Well there’s just so few.” And then, Métis, it was, “Well we don’t have data to really tell us that their outcomes are significantly worse than the general population.” That was the nature of the discussions a year ago when all this started. -Key Informant from Manitoba (ID 613)

An interviewee from Alberta explained that these disparities in health data availability for Métis populations stemmed from the Federal government’s prioritization of data-sharing agreements with First Nations, as they fall under Federal jurisdiction through the Indian Act. Métis people do not fall under federal jurisdiction, despite being section 35 recognized (one of the three indigenous peoples of Canada)¹⁰¹, resulting in Métis and Inuit people being left to navigate agreements with their provincial health authorities without support from the Federal government. In contrast, the Métis Nation of Alberta negotiated an information-sharing agreement with Alberta Health in 2010, so all FNIM populations were simultaneously prioritized for the COVID-19 vaccine.

Limited data availability on important FNIM subgroups also hindered the efficiency of the rollout and impacted the trust between FNIM populations and provincial governments. As mentioned earlier, urban FNIM were not explicitly prioritized in most provinces. KIs for Ontario shared that the urban FNIM population was a critical subgroup that was missed due to a lack of sufficient data on urban population size, risk, and the health inequities faced by this group. KIs emphasized that the majority of FNIM peoples in Canada now live in cities yet were significantly undercounted in the 2016 census (e.g. by a factor of 2 to 4 in a Toronto-based study)^68,102. They explained that, since historic partnerships of trust with FNIM communities had not been established prior to the pandemic, urban FNIM were considered “hard-to-reach” populations for COVID-19 vaccination and were, therefore, not efficiently prioritized early on in the rollout. Once again, strong advocacy from FNIM community leaders was required to include them in the prioritization plan.

Population Size

Population size influenced vaccine rollout and prioritization strategies at both the interprovincial and intraprovincial levels. Interprovincial population sizes refer to variations in population across Canadian provinces, while intraprovincial population size refers to the differences in size of population subgroups within a given province. These population sizes played a crucial role in determining which groups were deemed feasible to prioritize and shaped overall vaccine distribution strategies on a large scale.

Interprovincial Population Dynamics and Vaccine Rollout

Provincial population sizes influenced the perceived logistical benefits of implementing prioritization strategies beyond age-based categories. All provinces prioritized their populations based on age, the most significant risk factor for adverse COVID-19 outcomes and aligned with NACI recommendations²¹. However, the differentiation among provinces was the additional layers of prioritizations for certain groups apart from age. Some provinces considered their population size too small to further specify and prioritize specific population groups. As discussed earlier, KIs from Nova Scotia, for instance, revealed that the province determined that focusing on age-based prioritization would be more efficient because of the narrow time-spacing between age groups caused by the province’s relatively small population. Decision makers believed that introducing additional prioritized groups would require more time and resources, whereas prioritizing solely based on age could be implemented more swiftly and effectively.

Intraprovincial Population Variability and Prioritization Strategies

In the case of specific ethno-racial communities, such as Black populations, some provinces found these subgroups too small in their province to justify targeting for vaccination. For Black Canadians, this sentiment is reflected in the data. More than half (52.4%) of Canada’s Black population live in Ontario, followed by Quebec housing the second-largest population⁶⁴. While the prairie provinces have witnessed a rapid growth rate among the Black population, provinces like British Columbia experienced slower growth⁶⁴. Therefore, it is unsurprising that KIs from British Columbia referenced small population size as a reason why Black populations were not explicitly targeted, instead opting for community-based vaccine campaigns that targeted racialized groups in areas known to have a higher proportion of racialized individuals, recent immigrants, and refugees, referred to as Inclusion Groups. Additionally, the province focused on healthcare frontline worker groups, recognizing the overrepresentation of racialized workers in privatized healthcare settings. Conversely, in Toronto, a city in Ontario where 70.4% of Ontario’s Black population resides, there were Black-Led Vaccine Clinics, known as the Black Health Vaccine Initiative, organized by the Black Physicians’ Association of Ontario (BPAO) and several other local Black healthcare organizations^88,103, with the goal of reducing barriers to vaccine uptake. This was in addition to Black communities being targeted through “hot spot” communities in Ontario.

Even when provinces did prioritize small population groups, health officials acknowledged that it may have led to suboptimal outreach efforts. In British Columbia, for example, Inuit populations were targeted for prioritization, despite making up less than 0.01% of the population (0.6% of the FNIM population in British Columbia) in 2016¹⁰⁴. However, a KI stated that outreach efforts were not as organized and effective as they could have been because of the low representation.

Lastly, KIs noted that population size influenced the representation of certain population groups in terms of their ability to advocate effectively to government. In the early stages of the pandemic, several groups were prioritized because of their ability to gather, advocate, and be heard. However, smaller subpopulations lacked the same numbers and resources to advocate for prioritization or even recognition. For instance, when the Métis and Inuit were initially not prioritized alongside First Nations in Manitoba, the Manitoba Métis Federation had sufficient representation and resources to gather and advocate for their inclusion^99,100. However, numerous other groups did not have had the same representation or resources.I sat three days a week with provincial chief medical officers of health in this discussion and listened to other populations … emergency workers, police, fire, ambulance healthcare workers, teachers, others who came together and lobbied very strongly to be identified as priority populations. The advantage we had was the [First Nations] surveillance data. We could bring that to those discussions and those tables on an almost daily basis. I presented it nationally as well to say [that] this is what made our population different and that the planned response needed to be different. -Key Informant from British Columbia (ID 301)

Geography

The geography of certain provinces also presented a logistical challenge in vaccinating and providing care to the many remote communities within Canada. Canada is the second largest country in the world by land mass, yet its land is largely uninhabited¹⁰⁵. Most people live in the southern part of the country, near the American border¹⁰⁵. This vast geography, coupled with a relatively small population, has led to the formation of many remote and isolated communities that are hours, if not days, away from urban centres or resources. Some communities are so remote that they require dog sleds, snowmobiles, and air services to transport medical supplies from airports to the communities that need them¹⁰⁶.

Geographical isolation had divergent effects on provincial strategies for prioritizing remote and rural communities during the early stages of vaccine rollouts. It either presented obstacles to effective prioritization or provided a rationale for early prioritization.

In the early stages of vaccination, the COVID-19 vaccines available were delicate and required ultra-cold storage conditions. In provinces with rural and remote communities lacking the necessary infrastructure to transport and store vaccines under the required conditions, providing vaccines to these communities became a challenge. Instead, these provinces focused their attention on establishing vaccination centres within facilities that could appropriately store and disseminate early versions of the COVID-19 vaccines. As a result, provinces like Alberta and Manitoba faced limitations in effectively prioritizing remote and rural populations, including on-reserve FNIM communities.The transportation issue had to do with the fact that these vaccines are fairly fragile for transportation. They don’t like to be agitated too much and they like to be super cold. Those two things also meant that in the early days, until we had the infrastructure in place to do good transportation, we had to limit the number of places where you could provide the vaccine. – Key Informant from Alberta (ID 672)

On the other hand, KIs from other provinces referenced geographical isolation as a reason why remote communities were prioritized. Due to their closed and isolated nature, a COVID-19 outbreak could spread quickly and limited access to specialized services and health resources could result in severe COVID-19 related health outcomes. Thus, provinces like British Columbia regarded geographic isolation as a driving factor for prioritizing vaccination efforts in remote communities.

Strengths and Limitations

This paper offers a comprehensive examination of provincial prioritization plans and justifications provided, aimed at ensuring equitable access to and uptake of COVID-19 vaccines five key populations in Canada at high risk of negative outcomes and/or barriers to vaccine accessing the vaccine. A major strength of our study is our direct engagement with key decision-makers who played pivotal roles in shaping provincial vaccination policies and strategies. By interviewing these individuals, we gained unique insights into the thought processes, considerations, and challenges that informed the development and implementation of vaccination plans at the provincial level.

While our research aimed to be comprehensive and representative of all Canadian jurisdictions, there were limitations. First, our study focused on six provinces due to availability of published documents during the study period, potentially overlooking valuable insights from excluded provinces and territories. Nevertheless, we sought to encompass provinces with diverse geographical, demographic, and linguistic characteristics to ensure a broader representation of Canadian contexts. Second, we only interviewed a few KIs per province; therefore, the perspectives gathered may not represent those of the entire province. However, we conducted interviews until saturation was reached and aimed to capture a range of perspectives from key decision-makers involved in vaccine rollout efforts. Third, while our qualitative interviews and environmental scans aimed to capture all relevant strategies and interventions, it is possible that some initiatives may have been overlooked if they were not publicly accessible when environmental scans were completed nor mentioned by our KIs, despite our efforts to keep our environmental scans as up-to-date as possible. Additionally, NACI guidelines and provincial policy plan documents were regularly updated throughout the course of the rollout. This proved to be challenging when asking KIs to recall prioritization decisions that were made at the planning phase. To mitigate these challenges, our team regularly verified and recorded policy updates by date during the study period.

Moving forward, future research endeavors should expand beyond the scope of this study to evaluate the other four dimensions of the RE-AIM implementation framework for assessing vaccine rollout initiatives, including effectiveness, adoption, implementation, and maintenance. Moreover, the active involvement of priority populations in research design and interpretation will be crucial for ensuring that future vaccination efforts align with the needs and perspectives of diverse communities.