Analysis of COVID-19 Risk Following a Ring Vaccination Intervention to Address SARS-CoV-2 Alpha Variant Transmission in Montreal, Canada

IMPORTANCE Given limited COVID-19 vaccine availability early in the pandemic, optimizing immunization strategies was of paramount importance. Ring vaccination has been used successfully to control transmission of other airborne respiratory viruses. OBJECTIVE To assess the association of a ring vaccination intervention on COVID-19 spread in the initial epicenter of SARS-CoV-2 Alpha variant transmission in Montreal, Canada. DESIGN, SETTING, AND PARTICIPANTS This cohort study compared COVID-19 daily disease risk in 3 population-based groups of neighborhoods in Montreal, Canada, defined by their intervention-specific vaccine coverage at the neighborhood level: the primary intervention group (500 or more vaccinated persons per 10000 persons), secondary intervention group (95 to 499), and control group (0 to 50). The groups were compared within each of 3 time periods: before intervention (December 1, 2020, to March 16, 2021), during and immediately after intervention (March 17 to April 17,2021),and3weeksaftertheinterventionmidpoint(April18toJuly18,2021).Datawereanalyzed between June 2021 and November 2021. invited 1


Introduction
Early randomized controlled trials of vaccines against SARS-CoV-2 were designed to assess efficacy against COVID-19 disease, hospitalization, and death [1][2][3][4] ; evidence regarding a reduced risk of asymptomatic infection emerged later. [5][6][7] Available evidence and limited vaccine supply led many countries to adopt a vaccine prioritization scheme similar to influenza, which aims to reduce severe morbidity and mortality, and mitigate the burden of COVID-19 spread on health care systems [8][9][10][11][12][13][14][15] rather than to control transmission. This prioritization scheme led some jurisdictions, like the province of Quebec in Canada, to allocate vaccine doses based primarily on the size of the populations prioritized for vaccination in each administrative region, and to not formally take into account the number of COVID-19 cases in each region. This prioritization scheme missed the opportunity to use immunization as a means to control transmission in areas with high COVID-19 case counts.
Indeed, there is evidence from other communicable diseases that immunization can be a tool not only to prevent disease but also to stop transmission. Ring vaccination, as opposed to mass vaccination, targets contacts of confirmed cases, as well as persons who are in close contact with these contacts, with the ultimate aim of stopping transmission. It was first developed against smallpox 16 and has been used successfully to contain measles outbreaks in underimmunized communities. [17][18][19] Its potential effectiveness against Ebola virus disease was suggested by epidemic modeling 20,21 and confirmed by a randomized controlled trial comparing immediate with delayed vaccination. 22 Modeling also supports the effectiveness of ring vaccination against COVID-19, 23 and it has been recommended as a supplementary strategy against this infection in medium-or high-risk areas within low-incidence countries. 24 With mounting evidence of vaccine effectiveness against COVID-19 and asymptomatic SARS-CoV-2 infection, [5][6][7] as well as early evidence of efficacy against the Alpha variant, [25][26][27][28][29] when a surge in COVID-19 transmission among emergency shelter users and staff was detected in December 2020, Montreal Public Health, with the collaboration of its health care system and community partners, undertook a ring vaccination intervention of contacts and the contacts of contacts of cases among its shelter users and staff in January and February 2021. Whereas this intervention seemed to contribute to rapidly bringing COVID-19 transmission under control in the target population, the opportunity to evaluate ring vaccination and its potential limitations was not available because of the absence of a control group. Based on this experience, Montreal Public Health then sought to assess COVID-19 risk following a ring vaccination intervention around childcare center and school attendees, which have been observed to be important drivers of transmission, in the initial epicenter of SARS-CoV-2 Alpha variant transmission in Montreal in comparison with other areas neither targeted nor reached by the intervention.

Study Design and Participants
The objective of the ring vaccination intervention was to suppress transmission of the Alpha variant in Montreal by targeting its initial epicenter. The neighborhoods in Montreal where sustained local transmission of Alpha variant with high case counts was first detected experienced multiple introductions of Alpha variant related to travel to and from New York State, where Alpha variant transmission was higher. These neighborhoods also had large households with multiple children who attended childcare centers and schools, and had multiple contacts related to participation in social and religious events.
The proof-of-concept ring vaccination intervention was initially designed to offer, over a 2-week period, 1 dose of a COVID-19 vaccine to all adults residing in the 2 neighborhoods with the highest number of new cases of presumptive SARS-CoV-2 Alpha variant in the previous 2 weeks, with the aim of reaching 70% first-dose immunization coverage. The intervention was conceived in early March 2021, when only older persons and health care workers were eligible for immunization in the province of Quebec, and immunization coverage targets (first dose) of 75% had just been reached in those populations.
Because of limited vaccine supply and provincial government restrictions, the size of the population targeted by the intervention had to be reduced. At the time, over a third of all new COVID-19 cases in the targeted neighborhoods were aged less than 18 years, and in-person childcare center and school attendance were observed to be a major driver of overall transmission, in addition to transmission within households. However, none of the vaccines were yet authorized for use in children. Based on the patterns of transmission observed, it was decided to target exclusively parents of children attending childcare centers or schools in an area comprised of 2 adjacent neighborhoods with the highest presumptive Alpha variant case count in Montreal. In response to concerns raised by personnel, schoolteachers and childcare center educators were later added to the target population although they were not thought to be major drivers of transmission. Given the high case counts in these neighborhoods, many parents and teachers would have been considered contacts or contacts of contacts of cases of Alpha variant. We used a quasi-experimental design with a nonrandomized untreated control group, dependent pretest and posttest samples, and a double pretest 30 to evaluate this ring vaccination intervention. The compared groups were defined post hoc using Montreal neighborhoods, ie, 111 geographic areas that aim to depict a more accurate representation of the city's local communities than other administrative geographical units. 31 Ethical review of public health program evaluations is not required in Canada as per article 2.5 of the federal government's interagency advisory panel on research ethics. 32 One hundred and six neighborhoods (95% of Montreal neighborhoods, representing a total of 1 843 290 residents) were retrospectively assigned to 1 of 3 groups according to the proportion of their population vaccinated as part of the ring vaccination intervention. The primary intervention group comprised 4 adjacent neighborhoods: 2 with the targeted childcare centers and schools and 2 adjacent neighborhoods where many families and staff resided. All had an intervention-specific immunization rate of 500 or more per 10 000 persons. The secondary intervention group comprised 11 neighborhoods surrounding the primary intervention group, with an intervention-specific immunization rate between 95 and 499 per 10 000 persons. The control group represented neighborhoods that had very low (0 to 50 per 10 000 persons) intervention-specific vaccine coverage and included many neighborhoods distributed across the city. Five neighborhoods were excluded from the evaluation as they achieved moderate vaccine coverage as part of the intervention (between 51 and 94 per 10 000).
The 2 intervention groups were compared with a nonrandomized control group within each of 3 time periods: (1) before the ring vaccination intervention (December 1, 2020, to March 16, 2021), (2) in a stabilization period overlapping with the intervention period and before a potential effect could be observed because of the delay between vaccine administration, development of immunity, and eventual case reporting (March 17 to April 17, 2021), and (3) 3 weeks after the immunization efforts' midpoint until the end of the Quebec-wide third wave of the pandemic (April 18 to July 18, 2021). The 3-week delay after the midpoint was determined a priori and was informed by dose distribution data (not shown).

Data Sources
An intervention-specific extraction of the provincial immunization registry was provided by the health care team in charge of logistical aspects. These data were used for descriptive analysis of the population reached by the intervention. Data on COVID-19 cases were extracted from the provincial COVID-19 surveillance database on September 1, 2021; a variable entered at the time of case investigation indicated whether cases were confirmed (ie, by whole genome sequencing) or

Statistical Analysis
We performed a descriptive time-series analysis of disease risk smoothed using a 7-day moving average. We compared risk across groups using unadjusted risk ratios (RR) and their 95% confidence intervals for all cases and for cases presumptively infected with a SARS-CoV-2 variant of concern (mostly Alpha) to confirm whether the observations were statistically significant at a P < .05 significance level. The primary intervention group and the secondary intervention group were each compared with the control group, within 3 time periods. COVID-19 risk and risk ratios (RR) were calculated for all age groups, as well as for persons aged 30 to 59 years, the age group most reached by the intervention. Data analyses were performed using the SAS base software version 9.4 (SAS Institute) and Microsoft Excel 2016.  Table 1). Estimated vaccine uptake among eligible individuals was very high, ranging from 94% in parents to 98% in all eligible persons (data not shown). records, a slightly smaller proportion (83.4%; 95% CI, 83.1%-83.6%) of the primary intervention group was susceptible to SARS-CoV-2 at baseline compared with the secondary intervention and control groups (87.6%; 95% CI, 87.5%-87.6%), which is consistent with a higher 6-month cumulative COVID-19 risk in this group.

Results
At baseline, COVID-19 daily risks were substantially higher in the primary intervention group compared with the control group and secondary intervention group (Figure). The primary intervention group had an approximately 60% greater risk of COVID-19 and a 9 times greater risk of infection with a variant of concern compared with the control group (RR, 9.43; 95% CI, 8.43-10.55) ( Table 3). During the intervention period, the COVID-19 daily risk in the primary intervention group remained stable and statistically significantly higher than the control group and secondary intervention group (RR, 1.63; 95% CI, 1.52-1.76) (Figure and Table 3). Over the course of the postintervention period, disease risk showed a steep decline in the primary intervention group only, and the observed differences in disease risk between groups were no longer statistically significant (RR, 1.03; 95% CI, 0.94-1.12) (Figure and Table 3). The RRs before and during the intervention in the 30-to 59-year-old age group were in general slightly higher than for the overall

Discussion
Using a controlled quasi-experimental design, this cohort study found a statistically significant   infections, such as smallpox 16 and measles. 17,19 Limitations The main limitation of our study stems from the nonrandom allocation of the intervention; as a result, other factors may have contributed to the decreased incidence observed in the primary intervention group. One such factor was the mass immunization campaign that was underway concurrently in Montreal; however, data suggests that the proportion of persons vaccinated as part of the mass immunization campaign was similar in our primary (21.7%) and secondary (22.3%) intervention groups compared with the control group (22.0%) as of the end date of the intervention.  The highlighted area indicates the ring vaccination intervention period. The variant of concern over the intervention period was almost exclusively the Alpha variant (2974 of 3087 [96.3%] variant of concern cases between December 2020 and April 2021).
Differential exhaustion of susceptible individuals between the intervention and control groups could in theory have accounted for the observed effect. However, given that over 80% of the population in the primary intervention group was considered susceptible to COVID-19 at baseline, it is unlikely that the exhaustion of susceptible people in the primary intervention group played a role.
Furthermore, if exhaustion of the susceptible population had occurred in the primary intervention group, the latter's COVID-19 incidence rate should have decreased to less than that of the control group, but this was not observed.
Another potential factor that may have led to a differential decrease in disease risk after the intervention is reduced participation in COVID-19 testing in the intervention groups compared with the control group; however, daily testing rates remained similar across the 3 groups. Before the intervention, the rate was 2 persons tested daily per 1000 persons in all 3 groups, and 3 weeks after the intervention midpoint, the daily testing rates were 14 and 13 individuals per 1000 for the primary and secondary intervention groups, respectively, and 14 per 1000 for the control group. Unadjusted measures of association are presented; however, given that socioeconomic level, age, and population susceptibility to COVID-19 must have remained largely unchanged in the short evaluation period, it is unlikely that these variables confounded our results.
Before the intervention, the risk ratio for all COVID cases (as opposed to Alpha variant cases) in the secondary intervention group was lower than in the control group; this trend was observed as far back as 6 months prior to the start of the intervention (and possibly earlier) and may have been partially related to differences in socioeconomic or other factors. However, when looking at Alpha variant cases, the risk in the primary and secondary intervention groups were both higher than in the control group, which is consistent with the epidemiological situation at the time (ie, sustained local transmission of Alpha variant first occurring in those neighborhoods).
Whereas behavior change other than vaccination was not targeted by our ring vaccination intervention, the intervention groups may have modified other behaviors that affect COVID-19 risk (eg, mask-wearing, physical distancing) as a result of the intervention, which may have contributed to the reduction in disease risk observed postintervention. Therefore, the observed effect should be considered as associated with the overall ring vaccination intervention rather than the vaccine alone.
Finally, our analyses, including our investigations of potential confounding, were performed at the population or group level after aggregating neighborhood-level data, including those on prior e Variant of concern cases in the intervention period were almost exclusively Alpha variant cases (2974 of 3087 [96.3%] variant of concern cases between December 2020 and April 2021).