School Closure Decisions in Alberta and Ontario during COVID-19: Discourse and Data

Abstract The COVID-19 pandemic has led to significant disruptions in Canada's public school system as provincial and territorial governments have enacted sudden and prolonged school closures. We compare the different school closure decisions in Alberta and Ontario during spring 2021, using official public briefings and publicly available data about rates of COVID-19 cases. We ask if provincial policy decisions can be explained by different epidemiological contexts and risks. We find that key epidemiological indicators such as the rate of cases were not directly linked to school closure decisions. This is important for policy makers and experts: it problematizes the assumption of a straight line between evidence and decisions and has implications for transparency and public trust in pandemic policy choices. A systematic description of the gap between evidence and policy is an important starting point for asking, What does drive decisions to close schools?


Abstract
The COVID-19 pandemic has led to significant disruptions in Canada's public school system as provincial and territorial governments have enacted sudden and prolonged school closures. We compare the different school closure decisions in Alberta and Ontario during spring 2021, using official public briefings and publicly available data about rates of COVID-19 cases. We ask if provincial policy decisions can be explained by different epidemiological contexts and risks. We find that key epidemiological indicators such as the rate of cases were not directly linked to school closure decisions. This is important for policy makers and experts: it problematizes the assumption of a straight line between evidence and decisions and has implications for transparency and public trust in pandemic policy choices. A systematic description of the gap between evidence and policy is an important starting point for asking, What does drive decisions to close schools? As we enter the sixth wave of the COVID-19 pandemic, schools are open and-in many Canadian jurisdictions-masks are off. The decision to keep schools open is broadly supported by medical and educational experts, but the decision to remove masks in congregate school settings appears to fly in the face of a solidifying scientific consensus around the nature of the virus's airborne transmission (CDC, 2021). This most recent disconnect between evidence and action aligns with a pattern seen in earlier pandemic decisions about school closures and raises important questions about how governments learn and the ways they respond to epidemiological risks and contexts.
In this Currents piece, we explore one such puzzle regarding government decisions to close schools during the third wave (spring 2021) of the COVID-19 pandemic. Our research examines school closures in Alberta and Ontario, provinces with majority Conservative governments for the duration of the pandemic thus far. Despite similar politics and similar pandemics, Alberta and Ontario made different decisions about school closures and reopenings, particularly during the pandemic's third wave in spring 2021.
We ask whether provincial decisions about school closures and reopenings can be explained by different epidemiological contexts and risks. We find they cannot: in fact, key indicators about the severity of the pandemic were worse in Alberta than in Ontario at a time when Alberta was reopening schools in May 2021 and Ontario was keeping schools closed until the end of the school year. This question about the relationship between evidence and policy during a crisis is more relevant than ever, as key epidemiological indicators like the rate of COVID-19 cases are no longer being collected or reported, governments make decisions about mask mandates in schools, and families decide how to respond.
These moments of disconnect between evidence and policy undermine public trust in government, politicize science and evidence, and potentially make collective action to solve similar problems more difficult in future. The pandemic has produced many such moments in a relatively short period of time, and we need to understand them better.

Theorizing the Link between Evidence and Policy
In a critique of evidence-based policy making, Greenhalgh and Russell (2009) note that while "the idea that policy should be based on best research evidence might appear to be self-evident," such a narrow view of evidence-based policy making in health tends to negate moral and ethical issues. Evidence is sometimes ambiguous and value-laden, and political factors such as institutional constraints and pressure from organized groups shape both the definition of policy goals and the options available when making policy decisions (Stone, 2012).
As social scientists, we are familiar with these arguments. We therefore do not begin our study of pandemic-era policy making from a position of "naïve rationalism" (Greenhalgh and Russell, 2009), assuming that there is a direct line from the best available evidence at the time to a given policy decision. Nonetheless, it remains empirically puzzling that policy makers rely on scientific evidence to greater or lesser degrees (Heikkila et al., 2020), interpret similar types of evidence differently (Boothe, 2021) and draw different lessons from early pandemic experiences (Dunlop and Radaelli, 2017). Provincial decisions to close K-12 schools is a particularly interesting case where officials in both provinces used one type of epidemiological evidence to communicate about levels of risk and the severity of the pandemic (the status of COVID-19 cases in the community), but this measure ultimately did not appear to align with school closure and reopening decisions.

Methods and Data
In this article, we focus on public briefings and epidemiological data from the third wave of the pandemic in Alberta and Ontario in spring 2021. It is during this third wave that we see the most significant divergence between provincial decisions to close or reopen schools and the level of risk suggested by COVID-19 case data.
We identified the following anchor dates of interest based on key policy decisions relating to school closures and reopenings. For Alberta: the January 11, 2021 school reopening following the winter break, the May 4, 2021 school closure and May 25, 2021 reopening. For Ontario: the January 5, 2021 remote-schooling start to the semester, the February 16 return to in-person schooling, the April 12, 2021 school closure, and June 2, 2021 announcement that schools would remain closed. Public briefings released 14 days before and after the anchor dates were collected in August 2021 from relevant political and public health actors. When full transcripts were not available from the provinces' news briefings, transcripts were created from the official Government of Alberta, Ontario Government and Premier of Ontario YouTube channels. For wave three, a total of 111 briefings were collected across both provinces.
We also collected data regarding the rate of cases within each province to assess their use in decision making about school closures. The rates of cases per 100,000 population (last seven days) of COVID-19 in Alberta and Ontario were collected for every anchor date in each province. The rate of cases was selected because it was the metric commonly referenced by officials in both provinces during public briefings (although as noted above, rarely in direct reference to school closure decisions). Other measures of pandemic severity-such as numbers of COVID patients in hospital and intensive care units (ICUs), hospital and ICU bed availability, or case fatality rates-were not used because they were referenced less often in official briefings at this time. The intent in choosing the rate of cases as the measure for this study was not to track the data most relevant to the decision to close schools but rather to measure possible gaps between policy decisions and the most frequently cited, publicly available measure that citizens might have been using to judge pandemic severity. The data were collected in September 2021 using the Government of Canada's COVID-19 Daily Epidemiology Update and the Canadian Institute for Health Information's COVID-19 Intervention Timeline in Canada.
School Closures during Wave Three: Context and Case Rates School closures have become something of a regular occurrence in provincial pandemic policy making, with some fairly notable differences in provincial responses.
When the pandemic began in March 2020, all provinces and territories in Canada closed their schools to in-person learning. Uncertainty about both the severity and transmission patterns of the disease suggested a cautious approach (Silverman et al., 2020). While Quebec reopened schools in May of 2020, no other provincial government brought students back to school in the 2019/2020 school year. Provinces also varied in their ability to manage the pandemic, but in Alberta and Ontario-two provinces with broadly similar strategies-we see important differences in the provincial decisions to close and reopen schools in spring 2021.

Alberta
After the December 2020/January 2021 winter break, Alberta's public schools returned to in-person learning on January 11, 2021, when the province was reporting 148 new cases of COVID-19 a week per 100,000 individuals. At the time of the Alberta school closures, on May 4, 2021, Alberta had 308 new cases of COVID-19 a week per 100,000 individuals ( Figure 1).
On May 4, Alberta announced that all schools would be shifting to at-home learning for two weeks, until May 25, 2021. Although the government acknowledged that the increasing rate of cases and the number of school outbreaks were factors in this decision, the premier stated that the decision to close schools was a "reset" to address "operational issues" (Government of Alberta, 2021b). On May 18, the chief medical officer of health confirmed that the decision was based on operational considerations and was not based on public health reasons (Government of Alberta, 2021c).
This official distinction between public health and operational considerations will be important in future analyses. Operational issues were described as "independent from public health measures" and within the responsibility of the Ministry of Education. They include staffing issues associated with high community transmission and the difficulty of finding substitute teachers (Government of Alberta, 2021c). During the May 2021 school closure, Alberta's chief medical officer described schools as safe, saying "I don't believe there is a public health risk in bringing children back to school" (Government of Alberta, 2021c) but noted the lack of staff required the closure.
Alberta schools returned to in-person learning, as planned, on May 25. At this time, Alberta was reporting 112 new cases of COVID-19 a week per 100,000 individuals. On May 28, the premier reiterated that "when it comes to students attending school in person, we've always based our plans on the best medical advice and on operational considerations" (Government of Alberta, 2021a).

Ontario
Ontario returned to in-person learning after the December 2020/January 2021 winter break on February 16, 2021. At the time of reopening, the province was reporting 49 new cases of COVID-19 a week per 100,000 individuals. Ontario's education minister, Stephen Lecce, stated that the decision to return to school in February was based on the advice of the province's chief medical officer of health, the local medical officers of health and the unanimous recommendation of the council of medical officers of health (Government of Ontario, 2021b). Schools remained open during February and March 2021. Throughout this time, the Ford government maintained its messaging that Ontario schools were safe for in-person learning.
During a press conference on April 1, 2021, Minister Lecce announced that Ontario schools would remain open following the upcoming spring break that was scheduled for the week of April 12. The Sunday before the 2021 spring break (April 11), Minister Lecce sent an open letter to Ontario parents from the Ford government stating, "Our priority remains keeping schools safe and open for in-class instruction because learning is critical to the continued development and mental health of our children" (Government of Ontario, 2021c). However, the next day, Premier Doug Ford made the announcement that Ontario schools would not return to in-person learning following the spring break due to the concerning COVID-19 metrics (Government of Ontario, 2021a). When questioned about what had changed during this short time, Minister Lecce said the Ford government came to the decision "over the last hours" after discussing the matter with Ontario's top medical experts (Premier of Ontario, 2021). At the time of the school closures on April 12, 2021, Ontario was reporting 180 new cases of COVID-19 a week per 100,000 individuals.
Schools remained closed until an announcement was made on June 2, 2021, making it official that students would not return to classrooms for the 2020-2021 academic year. Days before the June 2 announcement, the Ontario COVID-19 Science

Discussion
The COVID-19 pandemic has highlighted pressing questions about the ways that risk is communicated by governments in the context of an ongoing crisis and the ways that different types of evidence are prioritized and interpreted when there is significant uncertainty. This article introduces a puzzling case of policy variation: two Canadian provinces-both with conservative political leadership and presumably similar access to developing evidence about COVID-19 transmission and risk-made different decisions about school closures and reopenings.
Before we can study the use of evidence by policy makers, however, we need to consider the possibility that governments in Alberta and Ontario made different decisions about school closures and reopenings because they were facing different epidemiological contexts. Ontario closed schools earlier in spring 2021 and kept them closed longer. Was this because the pandemic was simply worse in Ontario, compared to Alberta, and therefore the risk of in-person learning was higher?
Determining the risk of in-person learning is complex. As Stuart and Dowdy (2021) point out, "Rarely are blanket conclusions-for example, that reopening schools does not fuel SARS-CoV-2 transmission-appropriate." However, the rate of COVID-19 cases has been used as a shorthand for the severity of the pandemic in many jurisdictions during the first 20 months of the pandemic. It was therefore surprising to find that when decisions about school closures and reopenings diverged in Alberta and Ontario, they did so in the opposite direction we might have expected by observing the rate of cases.
This research allows us to reject the hypothesis that school decisions varied in response to a simple and accessible measure of the severity of the pandemic. It provides the groundwork for ongoing research that will analyze how governments communicated about evidence and risk and justified these decisions over time and will investigate how the timing of different policy choices affects the epidemiology of COVID-19. More immediately, it provides a framework for addressing pressing questions about how to understand current pandemic policy choices: to remove mask mandates, shorten isolation periods, and reduce the types of publicly available data and frequency of public briefings about the pandemic. Pandemic decisions may be driven by a range of factors, but there is a certain expectation that evidence will be a key driver. That has not been the case in the instances of wave-three school closures and reopenings that we studied.
We hope this finding will guide future research and will also help policy makers and the experts who advise them to consider or reconsider the challenges of transparency when reporting on rapidly changing evidence and policy choices. The public briefings we collected are replete with references to "the evidence," but if policy makers do not engage with the complexity of the relationship between evidence and policy, there is a risk of harming public confidence and trust in those decisions. This may be particularly the case as pandemic-weary citizens make their own comparisons and are forced to draw their own conclusions about the evidence for difficult policy choices.