• Risk factors for COVID-19 infection which are particularly relevant for public transport settings are the inability to social distance in restricted spaces such as buses and train/tube carriages, and frequent contact with the general public.
• Age, deprivation, ethnicity, and region are non-workplace risk factors for COVID-19 infection and mortality, but some public transport workers are more likely to be in these risk groups. For example, some of the increased risk of death from COVID-19 among London bus drivers at the start of the pandemic can be (at least partly) explained by a high proportion of these workers being from vulnerable groups (older, from ethnic minorities and living in disadvantaged areas).
• These factors can be taken into account in analyses to understand the contribution of working in public transport to the risk of COVID-19 infection and mortality. For mortality, these other factors explain quite a lot of the increased risk in transport workers. However, adjusting for these factors does not substantially affect the estimates for risk of infection among transport workers.
• It is difficult to separate the impact of workplace risk factors from other factors which are strongly associated with occupation or direct consequences of work, such as shared accommodation and travel to work among co-workers and local/community levels of infection.

Evidence gaps

• Further work within the TRACK study is analysing CCTV data to further understand the impact of behaviour around social distancing, mask wearing and face touching on risks of transmission on public transport.

Evidence

• Evidence from studies investigating control measures in the public transport sector (identified in this literature review) indicates the importance of ventilation for reducing risk of SARS-CoV-2 (the virus that causes COVID-19) transmission, although the benefits of ventilation vary depending on the position of passengers in a bus and by the configuration of open windows in a taxi or car. There is evidence from simulation and modelling studies that face coverings also reduce the risk of transmission, although it is difficult to quantify the reduction in risk.
• In response to the high rates of COVID-19 deaths among London bus drivers early in the pandemic, the UCL VIRAL study made recommendations for modifications to the driver cab areas, based on analysis of air flow simulations. The existing perspex ‘assault’ screens were modified to minimise the gaps to reduce airflow between the passenger area and the driver cab area. A separate ventilation system was recommended for the driver cab area to minimise airflow from the passenger area, with the driver cab window open when possible.
• Modelling of transmission in a London underground carriage, as part of the TRACK study, showed that wearing face coverings reduces airborne and close-range droplet transmission (Miller et al. 2022).
• Recent evidence on the likely low risk of transmission from surfaces relative to risk from airborne aerosols and droplets may suggest that cleaning regimes could be modified with efforts focussed on effective ventilation systems. However, visible interventions such as cleanliness are associated with a perception of reduced risk and could increase passenger confidence in using public transport, so may be important to maintain.
• A qualitative study was conducted to understand perceptions of effectiveness of mitigation measures introduced in the public transport sector in the UK (as discussed in this blog post).
  • Distancing, face coverings, additional cleaning, and hand sanitisation were introduced across the sector early on, with the importance of appropriate ventilation identified more recently. Work bubbles and staff testing have been implemented in some settings, although it was not always practical to set up workplace testing.
  • Respondents reported good communication between companies within each sector, and preventive measures were usually introduced consistently across the public transport industry. There were some concerns around messaging for the public when changes to regulations were frequent and often differed between the devolved nations.
  • Some respondents identified that it is difficult to know which measures are most effective, as most were implemented at speed and around the same time. However, in general, the measures were thought to be effective and frequently the low numbers of cases among employees were reported by respondents as evidence for this.

Evidence gaps

• More evidence is required on the relative effectiveness of different mitigation measures, to inform which should be continued or reinstated in future COVID-19 waves or new epidemics.

• Analyses of routine mortality data from 2020 by occupation have shown that drivers and transport workers had some of the highest COVID-19 mortality rates among men.
  • The Office for National Statistics (ONS) data for 9 March – 28 December 2020 on men aged 20 to 64 in England and Wales showed that the death rate among taxi and cab drivers and chauffeurs was 101.4 deaths per 100,000 males (209 deaths) and among bus and coach drivers was 70.3 deaths per 100,000 males (83 deaths). This compared to 31.4 deaths per 100,000 among all men aged 20 to 64. These rates were adjusted for age but not for other factors which may affect the risk of death from COVID-19.
  • An analysis of mortality data for England linked to 2011 Census data (Nafilyan et al. 2021) showed that a large part of the differences in COVID-19 mortality between occupations could be explained by factors such as regional differences in COVID-19 prevalence, ethnicity, education, deprivation or pre-pandemic health.
    • However, after adjustment for these other factors, among men, taxi drivers and van drivers were still at higher risk of death from COVID-19 compared to ‘non-essential’ workers* (see figure below). Taxi drivers were nearly 1.4 times more likely to die from COVID-19 than non-essential workers (fully adjusted hazard ratio 1.39 [95% Confidence Interval (CI) 1.14 – 1.70]). The fully adjusted hazard ratio for bus drivers was 1.11 [95% CI 0.85 – 1.45] and for van drivers was 1.26 [95% CI 1.03 – 1.55]. After full adjustment there was little evidence of an increased risk of mortality for other transport workers (e.g. train/tram drivers, large goods vehicle drivers, rail transport operatives, managers/directors in transport/distribution), who had a hazard ratio of 1.10 [95% CI 0.93 – 1.30]) compared to non-essential workers. Some of the differences in risk between categories of transport workers may depend on the level of contact with the general public.
    • Increased mortality was also observed for women taxi drivers (see figure below). Note that the relatively low numbers of women in these occupations mean that the confidence intervals are wide.

* ‘Non-essential’ workers are those not in the following groups: healthcare workers, social and educational workers, other essential workers (including food, transport, police and protective services).

The figure shows an increased risk of COVID-19 death among some categories of transport workers but not others. The details are provided in the text above.

ES: effect size. Greater than 1 indicates an increased risk, less than 1 indicates a decreased risk, e.g. a hazard ratio of 2.45 (1.01, 5.92) indicates that women taxi drivers had nearly 2.5 times the risk of COVID-19 death than non-essential workers (and we can be 95% certain that the true effect size is between 1% more likely and nearly 500% more likely).

• To avoid issues related to inaccurate recording of cause of death, the impact of COVID-19 can be estimated by comparing death rates during the pandemic with rates before the pandemic, known as excess mortality. A study in California, USA, identified the highest per capita excess mortality (compared to pre-pandemic mortality Jan 2016-Feb 2020 among those aged 18-65) to be among transportation and logistics workers (91 per 100,000 [95% Prediction Interval (PI) 81-102]), a relative excess mortality of 31% (95% PI 26-36). This varied by ethnicity and was highest for Latino workers in this sector, at 47% (35-60), and lowest for white workers at 9% (4-15).

Evidence gaps

• Occupation data for the ONS analyses reported above was from the 2011 Census, and may therefore be out of date for some individuals. The 2021 Census data will soon be available for updated analyses. The analyses can also be updated to include COVID-19 deaths beyond the end of 2020, to investigate how the risk of death for transport workers has changed over time.
• There is evidence of differences in vaccine uptake by occupation (Nafilyan et al. 2021). Now that the vaccine programme is well-established in the UK, future analyses could assess the impact of these differences in uptake on mortality differences by occupation.

Evidence

• There was no evidence of an increased risk of death from COVID-19 among food production workers compared to non-essential workers in an analysis of mortality data for England linked to 2011 Census data (Nafilyan et al. 2021) (see figure above).

Evidence gaps

• Occupation data for the ONS analyses reported above was from the 2011 Census and may therefore be out of date for some individuals. The 2021 Census data will soon be available for updated analyses.
• It is important to note that occupation may not be accurately recorded for migrant and temporary workers. This could affect the COVID-19 mortality estimates for the food industry, where a large proportion of the workforce in some sub-sectors are seasonal, temporary or migrant workers.
• Uptake of COVID-19 vaccination varies by occupation (Nafiylan et al. 2021). These differences may impact on rates of infection, COVID-19-related illness and risk of death in different occupations.

Evidence

• The literature review of studies in the food processing sector (Hosseini et al. 2022) identified several factors associated with an increased risk of COVID-19 infection:
  • Ethnicity: non-white ethnic groups were disproportionately affected in several US studies
  • Environmental factors: there were increased risks in workplaces with colder and more humid environments and poor ventilation.
• The COVID-19 At Work baseline survey (Mueller et al. 2022) included responses from 33 food and drink processing companies representing 66 sites across the UK.
  • Higher infection rates were associated with higher deprivation (based on the Index of Multiple Deprivation score of the local area), a lower proportion of remote workers, a lower proportion of workers in close proximity, and higher numbers of workers overall.
• A qualitative study (Stage 3 of the Covid at Work Study) was conducted to understand perceptions of risk of COVID-19 infection and challenges and enablers to successful mitigation measures across the food processing sector. 21 semi-structured interviews were conducted with 32 individuals representing the following stakeholder groups: government agency/departments, academics, federations/associations and unions.
  • In general, respondents did not consider the risk in the food processing sector to be higher than other similar industries.
  • Respondents identified the following risk factors for infection which are specific to the sector: reliance on manual labour rather than automation, which involves large numbers of workers in close proximity on production lines; cold ambient temperatures and humidity – difficult to maintain adequate ventilation with chilled air.
  • It was acknowledged that the risk of COVID-19 infection varies between different sub-sectors of the food and drink processing industry. Sub-sectors considered to be at particular risk of outbreaks are meat processing, agricultural food picking/production and sandwich making.

Evidence gaps

• A better understanding of the role of temperature and ventilation in transmission of SARS-CoV-2 would help to mitigate risks in the food processing sector.

Evidence

• Several of the studies summarised in the literature review (Hosseini et al. 2022) discussed preventive measures in food production facilities, including social distancing, enhanced cleaning and disinfecting of high-touch areas, enhancement of ventilation and providing more community and work-based testing.
• The COVID-19 At Work Stage 2 survey (Covid at Work Summary Report) of 24 sites followed up 6-9 months after the baseline survey showed that most sites had implemented most of the mitigation measures that were asked about. These included: physical barriers, floor markings, one-way systems, increased workplace cleaning, non-touch options, CO₂ monitoring, temperature checks (visitors, staff), provision of PPE and mental health support.
• The most commonly cited mitigations by respondents in the qualitative study (Stage 3 of the Covid at Work Study) included cleaning and hygiene practices, social distancing, measures to limit contact (e.g., screens, one-way systems, cohorts, staggering shifts), ventilation, testing and monitoring (surfaces and workers), PPE/face masks. Most respondents reported good compliance with preventive measures although some commented that social distancing was difficult to maintain within and outside operational areas.

Evidence gaps

• More evidence is required on the relative effectiveness of different mitigation measures in different sub-sectors of the food and drink processing industry. A tailored approach in each setting is likely to be most cost-effective and sustainable for future COVID-19 waves.