Key findings
People from ethnic minority groups in the UK face poorer physical health outcomes, including greater risk of COVID-19 infection and COVID-related bereavement; however, people from ethnic minority groups fared better than the White majority in relation to mental health.
We found a higher risk of COVID-19 infection among people from many ethnic minority groups compared with the White British group; COVID-19 related bereavement was also more likely among most ethnic minority groups.
There was a higher risk of physical multimorbidity among Bangladeshi and Black Caribbean women, and Gypsy/Traveller and Roma men, compared with their White British counterparts.
A higher risk of depression and anxiety was found for the Arab group. A higher risk of anxiety was also seen for people in the Any other Black background and White Irish groups. The White Irish group had a higher risk of experiencing an increase in loneliness during the pandemic. The risk of loneliness was also higher for people from the Mixed White and Black Caribbean group, and those from any other ethnic group.
People from the Roma and Chinese groups reported more difficulty in accessing health services, compared with the White British group.
However, there were some outcomes for which ethnic minority groups fared better than the White British group:
Levels of anxiety and depression were lower among people in the Black African, Chinese, White Eastern European and Any other Asian groups compared with the White British group.
People from Gypsy/Traveller, Roma, Chinese and Black African ethnic groups were less likely to experience loneliness during the pandemic than the White British group, and those from the Roma, Bangladeshi, Black African, Pakistani and Indian groups
had a lower risk of their loneliness increasing compared to before the pandemic than the White British group. People from the White Irish and Black African groups were able to access health services during the pandemic more readily than the White British group.
Introduction
There is a considerable body of evidence demonstrating ethnic inequalities in health in the UK (Nazroo, 1997; Erens et al, 2000; Sproston and Nazroo 2002; Sproston and Mindell, 2006; Bécares, 2015; Darlington et al, 2015; Stopforth et al, 2021a). When the first measures to tackle COVID-19 appeared in the UK in March 2020, the initial messaging from the government and beyond was that the virus does not discriminate. However, people from ethnic minority groups suffered greater levels of infection, hospitalisation and death during the pandemic compared with the White British majority (Pan et al, 2020; Public Health England, 2020; Mathur et al, 2021). This chapter explores ethnic inequalities in health and health-related outcomes in the UK during the COVID-19 pandemic, as well as inequalities in experiences of loneliness and bereavement. It also examines whether people from ethnic minority groups were able to access health services as readily as the White British majority during the pandemic.
Past research has shown the persistence of ethnic inequalities in health in the UK over a number of decades. Additionally, there is considerable evidence to show that racism is a fundamental cause of poor physical and mental health in ethnic minority groups (Karlsen and Nazroo, 2002; Williams et al, 2003; Williams, Neighbours and Jackson, 2003; Wallace et al, 2016; Nazroo et al, 2020 ‒ see also Chapter 4 for a further discussion of the Evidence for Equality National Survey [EVENS] findings regarding racism). Most recently, data from the 2015/17 wave of the UK Household Longitudinal Study (also known as ‘Understanding Society’) show that the chances of having a limiting long-term illness (LLTI) are increased among Black Caribbean, Black African, Indian, Pakistani and Bangladeshi groups compared with the White British group (Stopforth et al, 2021a). Similarly, data from three pooled years (2009, 2010 and 2011) of the Health Survey for England showed that Pakistani or Bangladeshi people had higher age-adjusted rates of limiting long-term illness compared to the White British majority, whereas those in Black ethnic groups showed lower LLTI rates (Darlington et al, 2015). There is also evidence that ethnic inequalities in health are worse in later life due to the disadvantage that has accumulated for ethnic minority people across the life course (Dannefer, 2003; Stopforth et al, 2021b). For example, in the UK, data from the 2011 Census show that
The effect of COVID-19 upon the health of the UK’s ethnic minority groups has been well documented. Evidence showing increased rates of COVID-19 infection among ethnic minority groups was published only a few months into the pandemic (Pan et al, 2020). Repeated studies have found higher levels of infection among people from ethnic minority groups (Public Health England, 2020). These higher levels of infection translated into higher rates of mortality among ethnic minority groups; for example, people from the Bangladeshi ethnic group had a mortality rate around five times higher than the White British group in the period from December 2020 to December 2021 (Mathur et al, 2021; ONS, 2022). The impact of COVID-19, and the resulting restrictions, also impacted on the mental health of people from ethnic minority groups in the UK. Levels of psychological distress were higher among non-White respondents to the Understanding Society COVID-19 survey (Understanding Society, 2022), and remained steady between Wave 8 (31.1%) and Wave 9 (30.7%) for the non-White group, but psychological distress levels reduced for the White group (24.2% in Wave 8 reducing to 20.3% in Wave 9). The UCL COVID-19 Study also reported higher rates of depression, anxiety, unemployment stress and financial stress among people from ethnic minority groups (Fancourt et al, 2020). In addition to the direct effect of COVID-19 infection on health, the effect of lockdowns and the government’s wider response to the pandemic greatly reduced people’s access to healthcare services (Mansfield et al, 2021). Furthermore, disruption to hospital admissions was greatest in areas with the largest proportions of ethnic minority people (Warner et al, 2021).
This chapter adds new evidence to the literature on ethnic inequalities of health in the UK, beginning with an investigation of ethnic inequalities in COVID-19 infection, before moving on to limiting long-term illness, mental health, loneliness and access to health services. This chapter will address to what extent the well-documented inequalities in COVID-19 infection are mirrored in other health outcomes, in terms of both physical and mental health.
In this chapter, we present the findings from a selection of the health measures collected in the EVENS data. All results presented here are the outcomes of logistic regression modelling, which was used to adjust for differences in the underlying age and sex structure of the different ethnic groups in the UK. More details can be found in Box 5.1 at the end of this chapter. The results of the logistic regression modelling are presented in charts, each of which compares outcomes for ethnic minority groups with the White British group. The red dotted line in each chart represents the White British group. Each ethnic minority group has a point estimate (represented by a dot) reported in an ‘Odds Ratio’ (OR) scale. Taking COVID-19 infection as an example, an OR of 2 means that the ethnic group in question experienced twice the levels of infection of the White British group, while an OR of 0.5 means that the ethnic group experienced half the levels of infection of the White British group. The horizontal lines either side of the dots on the chart represent the 95% confidence interval (CI), or the certainty of the estimate. Where these horizontal lines cross the red dotted line, it is unclear whether there is any difference between the ethnic minority group and the White British group.
COVID-19 infection
EVENS participants were asked if they had ever received a positive COVID-19 test. Given the increased likelihood of COVID-19 infection among older people and among men, the results presented here control for age and sex, as well as a squared age term to represent the non-linear effects of age, thereby accounting for the possibility that infection risk grew at an increasing rate with higher ages. Incorporating this adjustment means that we can be confident that any differences observed between ethnic groups are not simply due to differences in the age and sex structure of the population of each ethnic group. Higher levels of COVID-19 infection were seen among people from the Gypsy/Traveller, Bangladeshi, Mixed White and Black African, Pakistani, Black African, White Eastern European, White Irish and Indian groups (see Figure 5.1). The largest inequalities were seen for the Gypsy/Traveller group (OR 2.82, 95% CI 1.31–6.07) and the Bangladeshi group (OR 2.80, 95% CI 1.67‒4.70).
Bereavement
Physical multimorbidity
EVENS participants were asked if they had any physical health conditions, drawing from a list of five conditions (high blood pressure, diabetes, heart disease, lung disease and cancer) and were given the opportunity to specify if they had a health condition not included in the list. Here, we define ‘multimorbidity’ as having two or more physical conditions. Given the reported sex differences in LLTI (Bécares, 2015), separate analytical models were run for men and women. This analysis controlled for age, age squared and sex in order to account for the way in which physical multimorbidity becomes increasingly more prevalent in the most elderly (Barnett et al, 2012). Several ethnic minority groups had higher odds of having physical multimorbidity than the White British group (see Figure 5.3), namely Bangladeshi women (OR 4.91, 95% CI 2.40‒10.05), Black Caribbean women (OR 2.54, 95% CI 1.47‒4.39), Gypsy/Traveller men (OR 12.42, 95% CI 4.98‒30.94) and Roma men (OR 5.08, 95% CI 1.75‒14.77). White Eastern European men were less likely to have physical multimorbidity (OR 0.09, 95% CI 0.01‒0.84).
Mental health
Figure 5.4 shows that the Arab group had higher odds of both depression (OR 2.18, 95% CI 1.22‒3.90) and anxiety (OR 3.19, 95% CI 1.80‒5.66) compared to the White British group. The odds of having anxiety were also higher in the Any Other Black background and White Irish groups. To further explore these observations, we created additional separate models for men and women, and found that Arab women, but not Arab men were at higher risk of depression. Furthermore, only women from the Any Other Black background group had higher odds of anxiety compared with the White British group.
People from the Chinese, Any Other Asian, Black African and White Eastern European groups had lower odds of both anxiety and depression when compared with the White British group. Some ethnic minority groups, namely the Mixed White and Asian, and Roma groups, as well as Indian women, had lower odds of depression than the White British group, but had odds of anxiety that were not significantly different from the White British group.
As previously detailed, the UCL COVID-19 Study found higher rates of depression, anxiety, unemployment stress and financial stress among people from ethnic minority groups during the pandemic. This was generally not reflected in the age and sex-adjusted EVENS analysis. However, it should be noted that the UCL Social Survey results do not adjust for age and that when controlling for age, these ethnic differences are reduced. Indeed, when observing EVENS data that do not adjust for age and sex, there are higher levels of anxiety among people from the Arab, Any Other Black background, White Irish, Mixed White and Black Caribbean, Pakistani and Any Other White background groups compared with the White British group. The reason for this is that anxiety and depression appear to be more common in younger people, and the age structure of the UK’s ethnic minority groups is generally younger than the White British group
Social isolation and loneliness
Access to services
EVENS participants were asked about how readily they were able to access health and social care services during the pandemic. Here, we report the OR of being able to access required services ‘never’ or ‘hardly ever’, or not trying to access services despite having a need to do so. The results given in Figure 5.6 show that access was poorer for people from Roma (OR 2.45, 95% CI 1.31‒4.58) and Chinese (OR 1.71, 95% CI 1.18‒2.46) ethnic groups compared with the White British group. Conversely, the White Irish (OR 0.57, 95% CI 0.37‒0.87) and Black African (OR 0.72, 95% CI 0.53‒0.97) groups appeared to be able to more readily access services than the White British group.
Discussion
The results from the EVENS data give a comprehensive picture of the health of people from ethnic minority groups in Britain during the COVID-19 pandemic, with ethnic inequalities being present for physical health outcomes, coupled with mixed findings around inequalities in mental health outcomes. ONS data showed higher levels of COVID-19 infection and mortality for people from many ethnic minority groups, an observation mirrored in the EVENS data on coronavirus infection, and this is also suggested by the EVENS data on experiences of bereavement. Additionally, people from some ethnic minority groups were more likely to have physical multimorbidity compared with the White British group. However, while certain ethnic minority groups, including the Arab, Any other Black background, and White Irish groups, had increased odds of poorer mental health outcomes, the EVENS data shows lower odds of depression, anxiety and loneliness among people from several ethnic minority groups. Finally, there was evidence of inequitable access to services for people from Roma and Chinese groups during the COVID-19 pandemic.
The odds of COVID-19 infection were higher among EVENS participants from the Gypsy/Traveller, Bangladeshi, Mixed White and Black African, Pakistani, Black African, White Eastern European, White Irish and Indian groups, mirroring official statistics during the second wave of the pandemic (September 2020 to May 2021) (ONS, 2021b). Multiple reasons have been proposed for these ethnic inequalities including differential exposure to COVID-19 (for example, through occupation or working conditions), increased vulnerability to infection (for example, due to pre-existing health problems) and differential consequences of control measures (for example, employment insecurity and lack of sick pay) (Katikireddi et al, 2021).
The higher levels of COVID-related bereavement found among many ethnic minority groups, when compared with the White British group, mirror the higher rates of mortality seen in many of these groups according to the official statistics (ONS, 2021d). It should be noted, of course, that a person’s networks are likely to stretch beyond their own ethnic group. The detailed ethnic group categorisation used in EVENS facilitated the observation of COVID-related outcomes for ethnic groups that are not usually covered in national surveys. For example, the EVENS data showed that people from the White Eastern European group had higher rates of infection than the White British group, and also showed a higher level of COVID-related bereavement in the Jewish group compared with the White British group, in line with ONS analysis from the pandemic period (ONS, 2021c).
Physical health
Pre-pandemic literature points to poorer health among certain ethnic minority groups in the UK (Nazroo, 1997; Erens et al, 2000; Sproston and Nazroo, 2002; Sproston and Mindell, 2006; Bécares, 2015; Darlington et al, 2015; Stopforth et al, 2021b), and our sample shows this trend continuing, with evidence of ethnic inequalities in COVID-19 infection, COVID-related bereavement and physical multimorbidity. It should be noted that EVENS is the first to have sufficient data to identify poor health among Roma men.
There are some considerations around the use of the EVENS data when looking at physical health. First, existing evidence shows that ethnic inequalities in rates of LLTI are highest among older people (Bécares, 2015), whereas the EVENS data have relatively few participants aged 65 or older. Additionally, it is necessary to consider the timeline of the recruitment of the EVENS sample. The White British sample was recruited mainly through survey panels, three waves of which were conducted at the start of the EVENS data collection (during the second lockdown, in early 2021), with an additional panel being conducted at the end of the data collection, in late 2021. This is in contrast
Mental health and loneliness
Although on the whole, people from ethnic minority groups in the EVENS data had relatively good mental health outcomes compared with the White British group, some ethnic minority groups had poor mental health outcomes. The Arab group had higher odds of anxiety and depression than the White British group. There are very little data on the mental health of the UK Arab population, so this represents a novel finding. In addition, the White Irish group had higher levels of anxiety and higher odds of having experienced an increase in loneliness during the pandemic period than the White British group. This observation is consistent with other literature showing poorer mental health outcomes for Irish people living in England (Delaney et al, 2013).
There were not large differences in loneliness across the ethnic groups included in our sample; however, certain groups (Gypsy/Traveller, Roma, Chinese and Black African people) appeared to be less lonely than the White British group. These results were in contrast to those seen in the July 2020 findings from the UCL COVID-19 Study, where people from ethnic minority groups were more likely to have experienced loneliness since the beginning of the pandemic (Fancourt et al, 2020). In the EVENS data, the Roma, Bangladeshi, Black African, Pakistani and Indian groups were all less likely to have reported an increase in feelings of loneliness compared with the White British group. One potential explanation is that people living in multigenerational housing may have been less susceptible to loneliness when compared with those living alone or with one other person. This may be particularly relevant to the Gypsy/Traveller and Roma groups; in the EVENS data, people in both of these groups were less likely to be lonely than people in the White British group. As the Roma group are often excluded from social research, this represents a novel finding from EVENS.
Access to services
Among EVENS participants, there was some evidence of ethnic inequalities in access to health and social care services during the pandemic, with
It should also be noted that in the EVENS data, those from White Irish and Black African groups appeared to be able to access services more readily during the pandemic. The reasons for this are unclear and would benefit from additional research to understand what factors may be influencing these positive outcomes.
Conclusion
The long-term effects of COVID-19 on the health of the British population, and on ethnic health inequalities, are as yet unknown. In addition to the direct effects of COVID-19 on health, it is also important to consider the consequences of the measures taken to manage the pandemic and the emerging economic downturn. What is generally known is that periods of financial insecurity often affect the most socioeconomically deprived in society most acutely, and that socioeconomic deprivation, racial minority status and poor health are tightly interwoven. Additionally, the widespread levels of bereavement experienced by people from ethnic minority groups reflected in the EVENS data, termed the silent ‘pandemic of grief’, may have long-term mental health consequences which may not yet be fully apparent.
Health and wellbeing: measures and methods
All figures reported in this chapter were created using logistic regression models. Data were analysed using R version 4.2.0 (R Core Team, 2022). Analyses adjusted for benchmarking and propensity weights, which were implemented using a weights argument specified in the R glm library. Each model corrected for age (expressed as an integer) and sex. Some models also included an age squared term to account for the non-linearity of the effects of age, whereby its effects on health are often amplified at the oldest ages. The main variables used for each question were taken from the EVENS Health module and the Social Isolation module:
COVID-19 infection: The COVID-19 infection results draw upon the EVENS question HLTH13. ‘Have you ever received a positive result for a coronavirus (COVID-19) test?’; we considered the participants who responded ‘Yes’.
Bereavement: To explore bereavement, we utilised two questions from the EVENS survey: (i) HLTH16. ‘Have you experienced any bereavement of someone close to you (for example, a partner, family member or close friend) since February 2020? (Yes/No/Prefer not to say)’; and (ii) ‘HLTH17. Did the person, or any of the people, you lost die with coronavirus? (Yes/No/Don’t know/Prefer not to say)’
Physical multimorbidity: Physical multimorbidity was defined according to respondents’ answers to question HLTH06: ‘Do you currently have or have you ever had any of the following medical conditions? (Please select all that apply): 1. High blood pressure, 2. Diabetes, 3. Heart disease, 4. Lung disease (e.g., asthma or COPD), 5. Cancer, 6. Another clinically-diagnosed chronic physical health condition (please specify).’ Physical multimorbidity was defined as those respondents who responded ‘Yes’ to two or more of these conditions.
Mental health: The measure of depression was calculated using EVENS question HLTH04: ‘Now think about the past week and the feelings you have experienced. Please tell me if each of the following was true for you much of the time during the past week.’ Participants were then invited to respond (Yes/No/Prefer not to say) according to eight measures. A score was calculated by giving 1 point for ‘yes’ and 0 points for ‘no’. Two scale items (4 and 6) asked about positive symptoms (being happy or enjoying life) so were reverse-coded, whereby a ‘no’ response received 1 point and a ‘yes’ response received 0 points. In the analyses presented here, participants were said to have symptoms of depression if they scored 3 or more points. The measure of anxiety was calculated using EVENS question HLTH05: ‘Over the last two weeks, how often have you been bothered by any of the following problems?’ Participants were asked to respond (Not at all/Several days/More than half the day/Nearly every day/Prefer not to say) to seven separate measures aimed at evaluating symptoms of anxiety, such as trouble relaxing, or not being able to stop or control worrying. For each measure, participants were given a score, with ‘Not at all’ receiving a score of 0 and ‘Nearly every day’ receiving a score of 3. The score was summed; participants with a total score of 10 or more indicated symptoms of anxiety.
Social isolation and loneliness: The results on loneliness presented here refer to EVENS question ISOL01L: ‘The next questions are about how you feel about different aspects of your life. For each one, please say how often you feel that way at the moment.’ Participants were then invited to respond (Hardly ever or never/Some of the time/Often/Prefer not to say) to three questions regarding loneliness and isolation. Each question was scored, with ‘Hardly ever or never’ receiving a score of 1 and ‘Often’ receiving a score of 3. Participants with a total score of 6 or more were deemed to be exhibiting symptoms of loneliness. The results on change in levels of loneliness refer to question ISOL03: ‘Have your feelings of loneliness and isolation changed since the coronavirus outbreak began in February 2020?’ (they have: increased/decreased/stopped/stayed the same).
Access to services: The results pertaining to access to services consider responses to EVENS question HLTH07: ‘Since the coronavirus outbreak began in February 2020, have you always been able to access the community health and social care services and support you need, for instance your GP, a dentist, podiatrist, nurse, counselling for depression or anxiety or personal care?’ Participants responded on a scale ranging from ‘Yes, always’ to ‘No, never’. The following results consider those who responded ‘No, hardly ever’, ‘No, never’ or ‘I did not attempt to contact them’ to the question. The interpretation of this measure aims to evaluate whether respondents were able to get help if needed for any health problems they may have, and so excludes those who said they did not need to access services.
Key to interpreting the results in this chapter is an understanding of the context in which the EVENS data were collected. At the beginning of the data collection in February 2021, the UK was almost one year into the pandemic. England was one month into its third national lockdown, with the stay-at-home order remaining in place until 31 March 2021. It was not until 19 July 2021 that the majority of limitations on social contact were lifted. The initial effects of the pandemic on the mental health of UK residents were sudden and profound (Pierce et al, 2020), and although levels of anxiety and depression have stabilised, at the time of writing they have not yet returned to pre-pandemic levels (OHID, 2022). The effects on physical health are less clear; however, the potential effect of the COVID-19 pandemic should be considered when interpreting these results, especially where comparisons are made with pre-pandemic findings.
