Covid-19: how can we prevent people from ethnic minorities being disproportionately affected in a second wave?

The first wave of covid-19 in the UK has passed. We now know that there is over-representation of people from ethnic minorities among those infected; they are also at higher risk of severe outcomes.1-4 Despite a number of broadscale data analyses highlighting individual risk factors, there has been a dearth of analyses to identify which public health interventions should be prioritised to prevent infection of these diverse social groups. As lockdown eases and the risk of transmission potentially increases, we must learn from the first wave and not allow people from ethnic minorities to suffer disproportionately for a second time.5

It is clear that for a number of reasons, different ethnic minority groups have increased health risks from covid-19. Implementing interventions for all these aspects will be difficult. Although trials provide a gold-standard evidence base of the effectiveness of interventions, they are unlikely to be feasible or informative in the short term. Designing conceptual frameworks that describe mechanisms and pathways for an increased risk of covid-19 can generate testable hypotheses. From these we can consider broad reasons for the increased risk observed and start to conceive interventions. We must pull together all the relevant data to evaluate these hypotheses through mathematical modelling. It will play an important role in providing insights about which interventions targeting ethnic minority groups may be most impactful.

What kind of interventions show promise? Household determinants of covid-19 risk, such as intergenerational household mixing and multi-occupancy housing, which likely contributed to the Leicester outbreak, can be mitigated using appropriate public health messaging.6,7 This can help to improve individuals’ perception of their risk of being infected and infecting others, and aim to reduce transmission through increased household hygiene, hand washing, and social distancing. 

Employment risk is not the same for all ethnic minority groups, but interventions to reduce occupational exposure must take into account individuals’ age, ethnicity and underlying health conditions. The NHS and British Medical Association recommend risk assessment to identify healthcare workers at greatest risk and changes to their working conditions to mitigate this, for example, by redeployment.8,9 We endorse a broader scope using standardised risk categories (high, medium, low) to inform appropriate interventions across all occupations, for example reducing frequency of time on covid-19 wards and prioritising personal protective equipment use for healthcare workers, and reducing time on the shop floor for retail workers. 

Covid-19 has meant increased food and housing insecurity for vulnerable migrant populations, where unregistered employees have lost their income.10 Increased reliance on food banks and use of temporary, often crowded and substandard accommodation leads to a higher risk of exposure to covid-19. Financial and housing support for undocumented migrants would alleviate these pressures.  

How can we evaluate interventions to address risks such as these when time is not on our side? Mathematical modelling has already been a major influence on decisions regarding covid-19 in the UK, and has been used to generate qualitative and quantitative insights.11,12 It can rank the estimated effectiveness of proposed interventions or provide targets to gauge how successful implementation of an intervention would have to be to make a meaningful impact.

However, some model structures are more appropriate to model certain interventions than others. The centrality of the household in exposing various ethnic minority groups to risk of infection—multigenerational households, overcrowding, dormitory-style accommodation for low-paid, often migrant workers, housing insecurity and temporary accommodation—requires explicit modelling of household structure to simulate intra- and inter-household transmission. Simpler models, averaging out exposures and transmission risks into fewer groups, will miss the complexity of mixing patterns that make some groups more vulnerable than others. Greater complexity requires greater numbers of parameters and allows for a greater variety of structural approaches, which increases model diversity. Robust predictions therefore depend on the comparison of outputs from several models, which must be transparent about their data sources and key assumptions.13

Modelling such interventions requires more and better-quality data to inform model parameters. Data collection by ethnic group has been poor—few of the early covid-19 studies presented outcomes stratified by ethnicity.14 Few contact surveys, a core data source for modelling transmission patterns, and emerging contact tracing data, are stratified by ethnicity, but they do show signals of increased numbers of contacts in key worker occupations.15

From now on, minimum standards of data collection must include appropriately detailed information about ethnicity. Lessons must be learned from Leicester and other local outbreaks; ethnicity data availability are still lacking.7,16,17 We have moved from the initial, acute phase of the UK epidemic into outbreak response, but there must be continued data gathering to inform our understanding of the natural history and epidemiology of covid-19 and prioritise control measures. Which groups are disproportionately harbouring infection? Do these groups change over time? And to what extent is there transmission from these groups to others? While “models are not crystal balls” and cannot provide precise, quantitative predictions of the course of the pandemic, their estimates of the relative effect of various interventions in reducing disease burden should inform efforts to reduce the gap between ethnic minority groups and the rest of the population.13,18

Model estimates of interventions’ effectiveness need to be considered alongside their cost, acceptability and feasibility. Designing effective public engagement and information provision will be key to implementing shortlisted interventions. While substantial efforts have been made to translate public health messages into 60 languages, message penetration additionally requires delivery through the channels their target audiences use.19,20 Engagement with religious and community leaders is one way to facilitate this, as is learning lessons from previous public health campaigns.20,21 

Any intervention targeting specific ethnic groups may be controversial, with considerable barriers to acceptance. Prioritising some groups for benefits, protection or perceived discrimination based on ethnicity requires careful message delivery to explain its reasoning both to ethnic minority groups and the general population. However, we must differentiate between ethnic groups in our data gathering to understand the heterogeneity in covid-19 transmission between groups, strengthen the effectiveness of our interventions and prevent further outbreaks. We must understand and address the inequalities in infection risk, produce fairer health outcomes for all and prevent a second generalised wave of infection.

Rebecca F Baggaley, Department of Respiratory Sciences, University of Leicester, UK.

Déirdre Hollingsworth, Big Data Institute, University of Oxford, UK.

Manish Pareek, Department of Respiratory Sciences, University of Leicester, UK; Department of Infection and HIV Medicine, University Hospitals of Leicester NHS Trust, UK.

Competing interests: RFB is supported by the Wellcome Trust through a LISCB Wellcome Trust ISSF award [204801/Z/16/Z]. MP is supported by the NIHR. MP received grants and personal fees from Gilead Sciences outside of this Personal View. All other authors have nothing to declare.

The views expressed in this Personal View are those of the authors and not necessarily those of the National Health Service, the NIHR, the UK Department of Health or the Wellcome Trust. MP acknowledges the NIHR ARC—East Midlands, the NIHR Leicester Biomedical Research Centre, and the Centre for Black Minority Ethnic Health. RFB thanks Joy Onikoyi and Julia Chirnside for comments on earlier drafts of the manuscript.


  1. Aldridge RW, Lewer D, Katikireddi SV, et al. Black, Asian and Minority Ethnic groups in England are at increased risk of death from COVID-19: indirect standardisation of NHS mortality data [version 1; peer review: 3 approved with reservations]. Accessed 30th May 2020. Wellcome Open Research 2020
  2. Pan D, Sze S, Minhas J, et al. The impact of ethnicity on clinical outcomes in COVID-19: A systematic review (in press). Available at: Accessed 4th June 2020. EClinicalMedicine 2020
  3. PHE. PHE report. COVID-19: review of disparities in risks and outcomes. Available at: Accessed 4th June 2020. 2020
  4. PHE. Public Health England Report. Beyond the data: Understanding the impact of COVID-19 on BAME groups. June 2020. Available at: Accessed 16th June 2020. 2020
  5. The Academy of Medical Sciences. Preparing for a challenging winter 2020/21. Published 14th July 2020. Available at: Accessed 14th July 2020. 2020
  6. Coronavirus: ‘Big problem’ at Leicester factories, say workers. Available at: Accessed 13th July 2020. BBC News 2020.
  7. Nazareth J, Minhas JS, Jenkins DR, et al. Early lessons from a second COVID-19 lockdown in Leicester, UK. Lancet 2020 doi: 10.1016/S0140-6736(20)31490-2 [published Online First: 2020/07/06]
  8. NHS Employers. Risk assessments for staff. 28th May 2020. Available at: Accessed 14th June 2020 2020 [
  9. BMA. British Medical Association. COVID-19: risk assessment. Updated 5th June 2020. Available at: Accessed 13th June 2020 2020 [
  10. ‘Without food banks, I don’t know how I’d survive’: A hungry migrant family’s story. Available at: Accessed 13th July 2020. The Independent 2020.
  11. Ferguson NM, Laydon D, Nedjati-Gilani G, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. Published 16th March 2020. Available at: Accessed 14th July 2020. 2020
  12. Panovska-Griffiths J. Can mathematical modelling solve the current Covid-19 crisis? BMC Public Health 2020;20(1):551. doi: 10.1186/s12889-020-08671-z [published Online First: 2020/04/26]
  13. Jewell NP, Lewnard JA, Jewell BL. Predictive Mathematical Models of the COVID-19 Pandemic: Underlying Principles and Value of Projections. JAMA 2020 doi: 10.1001/jama.2020.6585 [published Online First: 2020/04/17]
  14. Pareek M, Bangash MN, Pareek N, et al. Ethnicity and COVID-19: an urgent public health research priority. Lancet 2020;395(10234):1421-22. doi: 10.1016/S0140-6736(20)30922-3 [published Online First: 2020/04/25]
  15. Danon L, Read JM, House TA, et al. Social encounter networks: characterizing Great Britain. Proc Biol Sci 2013;280(1765):20131037. doi: 10.1098/rspb.2013.1037 [published Online First: 2013/06/28]
  16. Gill M, Sridhar D, Godlee F. Lessons from Leicester: a covid-19 testing system that’s not fit for purpose. BMJ 2020;370:m2690. doi: 10.1136/bmj.m2690 [published Online First: 2020/07/09]
  17. Syal R. Leicester mayor accuses Matt Hancock of not sharing COVID-19 data. Published 8th July 2020. Available at: Accessed 14th July 2020. The Guardian 2020.
  18. Adam D. Special report: The simulations driving the world’s response to COVID-19. Nature 2020;580(7803):316-18. doi: 10.1038/d41586-020-01003-6 [published Online First: 2020/04/04]
  19. Doctors of the World. Coronavirus: Latest NHS guidelines translated into 60 languages. Available at: Accessed 31st May 2020. 2020
  20. Netto G, Bhopal R, Lederle N, et al. How can health promotion interventions be adapted for minority ethnic communities? Five principles for guiding the development of behavioural interventions. Health Promot Int 2010;25(2):248-57. doi: 10.1093/heapro/daq012 [published Online First: 2010/03/20]
  21. 21. PHE. PHE report. Local action on health inequalities: Understanding and reducing ethnic inequalities in health. Available at: Accessed 4th June 2020. 2018