COVID-19 among bartenders and waiters before and after pub lockdown

Introduction

The emergence of the coronavirus disease 2019 (COVID-19) in late 2019 and 2020 has led to an extraordinary social and economic upheaval across the globe. The SARS-Cov-2 virus has been assumed to spread easily in pubs and restaurants, because of lack of social distancing, consumption of alcohol, and load talk, which may increase the spread of respiratory droplets [1-2]. In sum, these factors may have put bartenders and waiters at a much higher risk of contracting the virus than persons in other occupations.

We recently reported that bartenders and waiters are among the occupations with the highest COVID-19 incidence in Norway [3]. In March 2020, a large number of COVID-19 cases across Europe could be traced to après-ski bars in Ischgl in Austria [4]. Clusters of virus outbreaks have also been related to pubs and restaurants in Japan and Thailand [1-2]. Thus, strict limitations on citizens’ behavior have been enforced in most countries to mitigate the spread of the virus, including partial or full bans on the serving of alcohol in pubs and restaurants. Despite its widespread use, the effect of different restrictions imposed on bars, pubs and restaurants in reducing the spread of the virus is unknown. It is also unknown how strict restrictions should be in order to reduce the spread of the virus, i.e. whether full bans on serving alcohol are required to limit the spread of SARS-CoV-2 among bartenders and waiters, or whether partial bans are enough.

Improved knowledge of the effect of social distancing measures, such as limitations on the serving of alcohol, may be important for stopping the spread of the virus in both bartenders and waiters, and the general population. It might also be important knowledge to owners of restaurants and bars, to protect their workers.

Using longitudinal population-wide data from Norway, we aimed to study associations between bans on serving of alcohol in Norwegian bars and restaurants, and the detection of SARS-CoV-2 in bartenders and waiters when compared to workers in other occupations.

Methods

Results

Of an overall 2,562,240 persons aged 20-70 years living in Norway on January 1^st 2020 who were employed by week 34, 2020, we studied the 1,312,246 persons living in the 56 municipalities in populous regions of Norway. Their mean [SD] age was 41.7 [12.8] years and 51.7% were men. Of these, 6,209 were employed as bartenders and 18,067 as waiters.

Between week 39 and 52, 2020, 13,875 cases of COVID-19 were notified among the employees, of whom 463 were bartenders (N=128) or waiters (N=335). In the same period 532,028 negative tests were reported in our study population, with 10,465 being from bartenders (N=3,071) and waiters (7,394).

Overall patterns

Results from the analysis including all waiters, bartenders and other occupations in all the 56 municipalities showed that the rate with COVID-19 was higher among bartenders and waiters than among persons with other occupations, across the time period (Figure 1). The steepest increase for bartenders and waiters was seen in week 45 to 46. Additional measures were imposed both nationally and locally in large parts of Norway in week 45 and 46 [5-7], and we observed a decline in infection, especially among bartenders and waiters, after these weeks (Figure 1).

• Download figure
• Open in new tab

Figure 1:

Estimated rate (95% confidence intervals) with confirmed COVID-19 per 1000 in 56 municipalities in Norway for bartenders and waiters (red line) and all other occupations (blue line) by calendar week, adjusted for age, sex and testing.

Local variations

Of the 56 municipalities, 12 imposed a full ban (covering 667,108 persons) and 16 a partial ban (covering 389,290 persons) on serving of alcohol, at various times. The remaining 28 municipalities did not impose any local restrictions (covering 255,848 persons) (Figure 2). Oslo and surrounding areas imposed the strictest restrictions, i.e. they had a full blanket ban on serving alcohol (Figure 2, E-Table 1).

• Download figure
• Open in new tab

Figure 2:

Map of all the 56 municipalities included in the study, with color codes showing the type of ban on serving alcohol. Red colored municipalities imposed a complete ban on serving alcohol. Yellow colored municipalities imposed partial restrictions (e.g. ban on serving alcohol after 10 pm). Green colored municipalities did not impose local restrictions (i.e. they followed only the compulsory national restrictions).

When weeks were relabeled to relative time from the week the restrictions were implemented, we could compare the rates of COVID-19 across municipalities with full versus partial vs no alcohol serving bans (Figure 3a-c). In municipalities that implemented a full ban (Figure 3a), the rate with COVID-19 for bartenders and waiters (3.4 per 1000, 95% CI=2.5-4.3) was higher than that for persons in other occupations (1.8 per 1000, 95% CI=1.7-1.9) in the week of implementation, declining by 65% (to 1.2 per 1000, 95% CI=0.7-1.7) by three weeks after the implementation to the same levels as for persons in other occupations (1.2 per 1000, 95%=CI 1.1-1.3).

• Download figure
• Open in new tab

Figure 3.

Estimated rates of confirmed COVID-19 per 1000 for bartenders and waiters (colored lines) and all other occupations (black lines) per week, from 5 weeks before to 5 weeks after the week when the municipality of residence implemented partial or full ban on serving of alcohol in bars and restaurants. For municipalities with no restrictions implemented, a hypothetical week of implementation was set by random draw (see Methods). Figure 3(a) shows the rate in municipalities with a full blanket ban on serving alcohol. Figure 3(b) shows the rate in municipalities with partial restrictions (e.g. ban on serving alcohol after 10 pm). Figure 3(c) shows the rate in municipalities with no local restrictions. Estimates based on logistic regression analyses adjusted for age, sex and testing.

In municipalities that implemented partial bans (Figure 3b), the rate with COVID-19 for bartenders and waiters (2.5 per 1000, 95% CI=1.4-3.6) was higher than that for persons in other occupations (1.7 per 1000, 95% CI=1.5-1.8) in the week of implementation. COVID-19 among bartenders and waiters had been reduced by 68% (to 0.8 per 1000, 95% CI=0.0-1.5) by three weeks, to the same levels as for persons in other occupations (1.0 per 1000, 95% CI=0.8-1.1). In municipalities with no restrictions and a randomly assigned hypothetical implementation week (Figure 3c), the rate for bartenders and waiters (0.8 per 1000, 95% CI=0.0-1.6) was variable yet similar to that of persons in other occupations (0.6 per 1000, 95% CI=0.5-0.7) both in the implementation week and three weeks after the hypothetical implementation (0.0 per 1000, 95% CI=0.0-0.0 for bartenders and waiters, and 0.4 per 1000 95% CI=0.3-0.5 for other occupations). Similar results were observed in separate analyses of bartenders and waiters, however, there was more uncertainty to these estimates (E-Table 2).

Discussion

In this study of 1,312,246 persons including all bartenders, waiters and other workers in 56 municipalities in the most populous regions of Norway, we report that partial and full bans on serving alcohol are associated with a lower incidence of COVID-19 among bartenders and waiters. We also find evidence that partial bans, such as closing pubs, bars and restaurants at 10 pm, are similarly associated with declining rates of detected COVID-19 among bartenders and waiters as full bans. These findings may be of importance for informing local and national authorities in their implementation of restrictions, as well as for companies in the catering and night life business to help protect their workers and customers.

To our knowledge, the current study is the first to assess effects of full or partial bans on serving alcohol during the COVID-19 pandemic. By using routinely collected population-based data and applying longitudinal panel data methods, we may have limited bias due to attrition, selection, confounding and misclassification - strengthening the transparency of causal inference from our study. As an example, data on occupation was available for everyone, and we could adjust for potentially different testing and infection patterns for COVID-19 among bartenders and waiters vs persons in other occupations. We found no previous studies for comparison of the potential effects of bans on serving alcohol. However, our findings confirm our previous reports showing that COVID-19 transmission is disproportionally higher among bartenders and waiters [3]. Our findings also shed new light on previous studies, as we show that transmission is reduced once hourly or full restrictions on serving alcohol are implemented, adding to the evidence that physical distancing interventions are associated with a reduced incidence of COVID-19 [8].

Our findings may be important for pandemic policy. We found indications that hourly restrictions, or partial blanket bans on serving alcohol from 10 pm may reduce the incidence of COVID-19 among bartenders and waiters (Figure 3b). However, there was more uncertainty to the estimates from these analyses of partial bans than there was for analyses of full ban, in which we observed a more certain reduction (Figure 3a). Along the reduction among bartenders and waiters after the two types of alcohol bans, there was also a reduction in COVID-19 among persons in other occupations (Figure 3a, 3b). This reduction in rates of COVID-19 could not be observed in municipalities with no implemented ban, neither among bartenders or waiters, nor among persons in other occupation (Figure 3c). Altogether, these findings may have two different explanations: Either, that full and partial bans on serving alcohol are effective in reducing transmission among persons with any occupation (i.e. both bartenders and waiters and other occupations), or that measures only are implemented at the top of a transmission curve, which would decline regardless of the implemented measures.

Along this line, some important limitations should be mentioned. First, we could not rule out that infection rates might have declined regardless of the imposed restrictions. While this decline might in theory be related to the restrictions in serving of alcohol, it may also be a sign that the restriction was implemented when the local infection rate was peaking. As an example, although the analysis shows that a partial ban may be similarly associated with declining infection rates as a full ban, infection rates were higher in areas imposing a full ban on serving alcohol, and we cannot say whether implementing a partial ban in these areas would have had the same effect. Fully randomized designs would be required to exclude this potential bias.

Second, restrictions on alcohol serving were rarely or never imposed alone. As an example, in the same week as the nationwide restrictions on alcohol were imposed (week 45), the government also encouraged people to limit social contact and avoid unnecessary domestic travels. In the following week, a constrain of maximum 50 persons in indoor events without fixed seating and 200 with fixed seating, was effectuated [6]. It is likely that also the municipalities with partial vs. full bans implemented such co-measures, whereas the municipalities with a hypothetical implementation week did not. Although the larger decline for bartenders and waiters than for other occupations links our findings to bars and restaurants, there is no way to ensure that our results do not reflect changes in the infection rates that are related to all these measures combined. Relatedly, we capture the municipality of residence of the employees, and some may not be working in the same municipality as they live, for whom the combination of restrictions they have faced are additionally involved.

In conclusion, we report that partial bans on serving alcohol in bars and restaurants may be similarly associated with declines in confirmed COVID-19 among bartenders and waiters as full bans. Considering the burden of full bans to owners and workers in bars and restaurants, hourly restrictions of serving alcohol, i.e. from 10 pm, may be explored further in reducing the spread of the SARS-CoV-2.

Conflict of interest disclosures

All authors have completed the ICMJE uniform disclosure form and declare: no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

Author contribution

Fredrik Methi had access to all of the data in the study and takes full responsibility for the integrity of the data and the accuracy of the data analysis. Fredrik Methi and Kjetil Telle performed the statistical analyses and Karin Magnusson drafted the manuscript. All authors contributed with acquisition of data, conceptual design, analyses and interpretation of results. All authors contributed in drafting the article or critically revising it for important intellectual content. All authors gave final approval for the version to be submitted.

Funding/support

The study was funded by the Norwegian Institute of Public Health. No external funding was received.

Role of the funder

The funding sources had no influence on the design or conduct of the study, the collection, management, analysis, or interpretation of the data, the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.