Appendix D: Further data and evidence

Question 1

Differences between the first, second and third waves

Accepted Answer

The third quarterly report set out the differences between the first wave of the pandemic (24 January 2020 to 31 August 2020) and the second wave (1 September 2020 to 31 January 2021) using ONS analysis of COVID-19 mortality. Their analysis modelled hazard ratios, which measure the relative risk of death for ethnic minority groups compared with the majority White British ethnic group. The model was adjusted iteratively to explore the change to the hazard ratios once different factors were accounted for.

This section updates the analysis of relative risk set out in the third quarterly report using hazard ratios. It also describes the absolute risk of death using annualised age-adjusted mortality rates (AAAMRs).

Hazard ratios

ONS’ latest analysis, ^{[footnote 1]} ^{[footnote 2]} incorporates COVID-19 deaths during February and March 2021 and updates the time periods for the 2 waves. The first wave is considered to be the period from 24 January 2020 to 11 September 2020 ^{[footnote 3]} while the second wave is from 12 September 2020 to 31 March 2021. ONS also expanded the population of the study to include people who live in different residence types (care homes and other communal establishments), where previously the analysis had only looked at individuals who lived in private households ^{[footnote 4]}. The model adjusted only for age quantifies an ethnic group’s risk of mortality compared with the White British population. The model adjusted for residence type (private residence, care home or other communal residence), geography, socioeconomic factors and pre-existing health conditions quantifies how much risk is due to these factors. Any residual risk is currently unexplained by measurable risk factors.

Figure 2a: Risk of death involving COVID-19 compared with White British people, expressed as hazard ratios, by ethnicity and sex during the first wave of the pandemic (24 January to 11 September 2020)

Bar chart showing that, adjusting only for age, men and women from the Bangladeshi group were 3.0 and 1.9 times as likely to die with COVID-19 as White British people in the first wave.

Figure 2a: see this data in a table.

Source: Office for National Statistics

Figure 2b: Risk of death involving COVID-19 compared with White British people, expressed as hazard ratios, by ethnicity and sex during the second wave of the pandemic (12 September 2020 to 31 March 2021)

Bar chart showing that, adjusting only for age, men and women from the Bangladeshi group were 5.0 and 4.1 times as likely to die with COVID-19 as White British people in the second wave.

Figure 2b: see this data in a table.

Source: Office for National Statistics

The study shows that, adjusting only for age, people in the Bangladeshi and Pakistani ethnic groups had a larger excess risk of death compared with White British people in the second wave than in the first wave:

in the first wave, men and women from the Bangladeshi group were 3.0 and 1.9 times as likely to die as their White British counterparts, compared with 5.0 and 4.1 times as likely in the second wave
for men and women in the Pakistani ethnic group, first wave hazard ratios were 2.2 and 2.0, compared with 3.4 and 2.8 in the second wave

Adjusting for residence type, geography, socio-economic factors and pre-existing health conditions ^{[footnote 5]} attenuated the risk but people in Pakistani and Bangladeshi ethnic groups still experienced excess risk compared with White British people:

second wave fully adjusted hazard ratios for men and women in the Bangladeshi ethnic group were 2.5 and 1.9, compared with 5.0 and 4.1 adjusted only for age
second wave fully adjusted hazard ratios for men and women in the Pakistani ethnic group were 2.0 and 1.5, compared with 3.4 and 2.8 adjusted only for age

In addition, after adjusting for these additional factors, men and women in the Indian ethnic group also had a larger excess risk of death in the second wave than in the first wave compared with White British people.

Adjusting only for age, compared with White British counterparts, Black African and Black Caribbean people, White Other men and men from other ethnic groups had a smaller excess risk in the second wave than in the first wave:

in the first wave, men and women from the Black African group were 3.7 and 2.6 times as likely to die as their White British counterparts, compared with 2.2 and 1.6 times as likely in the second wave
for Black Caribbean men and women, first wave hazard ratios were 2.7 and 1.8, compared with 1.7 and 1.4 in the second wave
for White Other men, the first wave hazard ratio was 1.3, compared with 1.1 in the second wave
for men from the other ethnic group, the first wave hazard ratio was 2.1, compared with 1.7 in the second wave

After adjusting for residence type, geography, socio-economic factors and pre-existing health conditions:

second wave fully adjusted hazard ratios for Black African men and women were 1.7 and 1.2, compared with 2.2 and 1.6 adjusted only for age
second wave fully adjusted hazard ratios for Black Caribbean men and women were 1.2 and 1.0, compared with 1.7 and 1.4 adjusted only for age
the second wave fully adjusted hazard ratio for White Other men was 0.9, compared with 1.1 adjusted only for age
the second wave fully adjusted hazard ratio for men from the other ethnic group was 1.4, compared with 1.7 adjusted only for age

After adjusting for these additional factors, the data suggests that compared with White British counterparts, only Black African men had a significantly smaller excess risk in the second wave than in the first wave. Black Caribbean people and Black African women had similar excess risk compared with White British counterparts in the second and first wave.

ONS analysis of hazard ratios shows that adjusting for geographic variables (local authority district and population density) brings the largest reduction in excess risk for ethnic minority groups compared with adjusting for other groups of variables.

Table 1a: Hazard ratios of death involving COVID-19 for men, by ethnic group and pandemic wave

Ethnic group	Age: Wave 1	Age: Wave 2	+Residence type: Wave 1	+Residence type: Wave 2	+Geography: Wave 1	+Geography: Wave 2	+Socio-economic, household, occupation: Wave 1	+Socio-economic, household, occupation: Wave 2	+Health: Wave 1	+Health: Wave 2
Bangladeshi	3.00	4.96	3.48	5.51	2.14	4.04	1.66	3.19	1.29	2.45
Black African	3.70	2.16	3.91	2.21	2.35	1.79	2.05	1.58	2.24	1.72
Black Caribbean	2.66	1.69	2.62	1.66	1.59	1.33	1.46	1.21	1.40	1.19
Chinese	1.39	1.11	1.49	1.15	1.08	1.01	1.12	1.06	1.20	1.13
Indian	1.85	1.95	2.05	2.10	1.38	1.76	1.42	1.81	1.34	1.68
Mixed	1.58	1.39	1.64	1.42	1.23	1.25	1.10	1.11	1.12	1.11
Other	2.11	1.73	2.26	1.79	1.54	1.54	1.38	1.39	1.39	1.39
Pakistani	2.21	3.40	2.57	3.79	1.70	2.85	1.46	2.45	1.22	2.04
White other	1.31	1.06	1.30	1.05	1.01	0.96	0.99	0.94	0.95	0.91

Table 1b: Hazard ratios of death involving COVID-19 for women, by ethnic group and pandemic wave

Ethnic group	Age: Wave 1	Age: Wave 2	+Residence type: Wave 1	+Residence type: Wave 2	+Geography: Wave 1	+Geography: Wave 2	+Socio-economic, household, occupation: Wave 1	+Socio-economic, household, occupation: Wave 2	+Health: Wave 1	+Health: Wave 2
Bangladeshi	1.93	4.11	2.27	4.67	1.35	3.24	0.86	2.09	0.82	1.93
Black African	2.61	1.62	2.93	1.77	1.8	1.45	1.36	1.12	1.52	1.22
Black Caribbean	1.81	1.35	2.03	1.49	1.21	1.18	1.11	1.09	1.04	1.01
Chinese	1.2	0.92	1.34	1	1.03	0.91	1.05	0.94	1.38	1.25
Indian	1.62	1.57	1.85	1.75	1.24	1.51	1.15	1.42	1.16	1.42
Mixed	1.52	1.36	1.6	1.41	1.27	1.31	1.14	1.18	1.08	1.1
Other	1.71	1.5	1.89	1.61	1.32	1.42	1.12	1.23	1.17	1.29
Pakistani	2.01	2.84	2.43	3.28	1.64	2.47	1.22	1.83	1.02	1.5
White other	1.03	0.9	1.07	0.93	0.86	0.87	0.85	0.86	0.82	0.83

Annualised age-adjusted mortality rates

PHE have analysed the deaths of individuals who had a laboratory-confirmed positive COVID-19 test between 31 July 2020 and 31 July 2021, and have calculated mortality rates per 100,000 of the population for each broad ethnic group ^{[footnote 6]}. These mortality rates have subsequently been age-standardised, to account for differences in the age structures of ethnic groups, and annualised, to account for differences in the size of time periods studied and to therefore allow comparisons between time periods ^{[footnote 7]}.

The annualised age-adjusted mortality rate (AAAMR) was highest for the Asian ethnic group, at 354.6 per 100,000 of the population per year (based on 6,644 deaths), followed by the black group (284.6 per 100,000 of the population per year, based on 2,478 deaths) and the other ^{[footnote 8]} group (220.8 per 100,000 of the population per year, based on 468 deaths). The

AAAMRs for these groups were significantly different to one another and significantly higher than the AAAMRs for the white (159.5, based on 80,461 deaths) and mixed (158.6, based on 432 deaths) groups, who had similar rates.

Table 2: Number of deaths and age-adjusted annualised mortality rate (AAAMR) per 100,000 population in laboratory-confirmed cases of COVID-19, by ethnicity (PHE table)

Ethnic category ***	Deaths (31 July 2020 to 31 July 2021 )	Age-adjusted mortality rate** (95% CI) (31 July 2020 to 31 July 2021)
White/White British	80,461	159.5 (158.4-160.7)
Black/Black British	2,478	284.6 (272.8-296.7)
Asian/Asian British	6,644	354.6 (345.7-363.7)
Mixed	432	158.6 (142.8-175.4)
Other±	468	220.8 (199.5-243.5)
Unknown	1,782	-

** Rates are annualised and expressed as the number of deaths per 100,000 population per year *** Ethnic categories are based on ONS classifications

Source: Public Health England

Cumulative age-adjusted mortality rates

PHE have published cumulative mortality rates for deaths involving COVID-19 by ethnic group on the CHIME platform. This data shows that cumulative ASMRs are higher in every ethnic minority group than they are in the white group (231.3) – ASMRs are highest in Black Other (926.9), Pakistani (736) and Bangladeshi (716.3) populations.

Table 3: Cumulative age-standardised mortality rate for deaths involving COVID-19 in England, by ethnicity (March 2020 to August 2021)

Ethnic group	Deaths (March 2020 to August 2021)	Age-standardised mortality rate (March 2020 to August 2021)	Lower CI	Upper CI
White	117920	231.3	229.9	232.6
Pakistani	2831	736	705.7	767.1
Bangladeshi	1039	716.3	668	766.9
Chinese	379	267.3	239.1	297.8
Indian	3524	399.4	385.7	413.4
Asian Other	1895	544.6	516.4	573.7
Black African	1427	493.7	461.3	527.5
Black Caribbean	2278	408.2	390.7	426.1
Black Other	753	926.9	849.1	1009.1
Mixed or multiple ethnic groups	721	267.4	246.9	289.1
Any other ethnic group	694	286.3	263.3	310.7

Source: Public Health England

Age-adjusted case rates

ONS analysis shows that case rates in the second wave (1 September 2020 to 22 May 2021) were highest in the Bangladeshi and Pakistani ethnic groups, at 390.6 and 378.1 cases per 100,000 person-weeks respectively, while case rates from the third wave to date ^{[footnote 9]} (23 May to 25 July 2021) were highest in White British people at 234.7 cases per 100,000 person-weeks. Most ethnic groups have a lower case rate in the third wave to date, compared with the second wave – this is very pronounced in the Bangladeshi, Pakistani and Indian groups. However, the case rate is notably higher in the third wave than in the second in the mixed and White British ethnic groups.^{[footnote 10]}

Table 4: Age-adjusted COVID-19 case rate per 100,000 person-weeks in England, by ethnicity and time period (wave 2 and early wave 3)

	Wave 2 (1 September 2020 to 22 May 2021)				Wave 3 (23 May 2021 to 25 July 2021)
Ethnic group	Number of cases	Rate	Lower CI	Upper CI	Number of cases	Rate	Lower CI	Upper CI
Bangladeshi	44,428	390.6	386	395.2	4,963	138.7	134.1	143.4
Black African	48,417	202.5	200.3	204.7	9,650	137.2	134.2	140.2
Black Caribbean	27,950	186	183.7	188.2	5,986	170.3	165.9	174.7
Chinese	6,985	93	90.7	95.3	1,694	97.5	92.7	102.3
Indian	102,805	269.5	267.8	271.2	14,845	152.7	150.2	155.2
Mixed	56,116	184.7	182.7	186.8	22,299	208.6	205.3	211.8
Other	88,751	240.7	239	242.4	14,552	144.3	141.9	146.7
Pakistani	111,920	378.1	375.5	380.7	15,335	169	166	172
White British	1,859,074	166	165.8	166.3	613,531	234.7	234.1	235.3
White Other	116,091	168.1	167	169.1	24,834	158.1	156	160.2

Source: Office for National Statistics

This change in trend in the third wave is corroborated by more recent age-standardised case rates from CHIME. In October 2021, the white population had the highest case rate at 22,784.8 cases per 100,000 person-years, followed by the Indian ethnic group at 21,668.8 per 100,000 person-years. The Pakistani and Bangladeshi ethnic groups had lower case rates at 15,804.6 and 14,179.8 per 100,000 person-years respectively. The Black Other ethnic group had the lowest case rate, at 9,200.1 per 100,000 person-years.

A similar change can be seen in case rates by deprivation decile. The most deprived deciles generally had the highest case rates during most of the pandemic to date but October 2021 data shows a clear reversal. The least deprived decile now has the highest case rate at 27,183.9 per 100,000 person-years, while the most deprived decile has the lowest case rate at 17,218.4 per 100,000 person-years.

The cause for this change in trend is as yet unclear but it comes after prolonged periods of high infection and antibody development in some communities, and at a time with greater opportunity for international travel among the affluent.

Figure 3: Monthly age-standardised confirmed COVID-19 case rate per 100,000 person-years in England, by ethnic group (March 2020 to October 2021)

2 line graphs showing the monthly age-standardised confirmed COVID-19 case rate per 100,000 person-years in England, by ethnic group from March 2020 to October 2021.

Figure 3: see this data in a table.

Source: Public Health England

Question 2

Excess deaths

Accepted Answer

The number of excess deaths over a given period of time refers to the number of deaths which have occurred in addition to the deaths expected for that time of year, as determined by mortality rates from earlier years. The metric gives a broader sense of the impact of the pandemic, because it considers all deaths, not just those attributed directly to COVID-19. For example, it captures COVID-19 deaths which were not correctly identified and reported, as well as deaths which may have occurred indirectly as a result of strain placed on healthcare services by the pandemic. It also captures deaths of people not presenting to hospital.

PHE’s analysis of excess deaths, ^{[footnote 11]} ^{[footnote 12]} by ethnicity shows that, for the period between the week ending 27 March 2020 and the week ending 11 September 2020, the ratio of the number of all registered deaths to the number of expected deaths was highest for the black ethnic group, at 1.75. The ratio was similar for the other, Asian and mixed ethnic groups (1.52, 1.50, and 1.49 respectively). It was lowest for the white ethnic group, at 1.22. For the period between the week ending 18 September 2020 and the week ending 02 April 2021, the ratio was highest for the Asian ethnic group (1.60) followed by the mixed ethnic group (1.46). The ratio for the black ethnic group was 1.43, while for the white ethnic group it was 1.13.

Research has looked at excess deaths by area-based deprivation and ethnicity among those aged under 75 years.

Between 21 March 2020 and 26 February 2021, there were 1.17 times as many deaths as expected for white people and 1.63 and 1.58 times as many deaths as expected for Asian and black people – the lowest and highest ratios of excess death respectively.

Among white people, there was a clear positive gradient between excess deaths and increased deprivation – that is, white ethnic groups in more deprived areas experienced higher excess death from all causes.

For people from the Asian and black ethnic groups, there were no clear gradients of excess deaths between the most and least deprived areas – that is, there is no clear link between area-based deprivation and excess deaths.

The analysis showed that there were separate associations between excess mortality and deprivation and excess mortality and ethnicity in people aged under 75 years over the period studied. This suggests that some aspect of ethnicity was a determinant of excess mortality, regardless of area-based deprivation.

Analysis from ONS found that at a neighbourhood^{[footnote 13]} level, there were around 200 neighbourhoods where the number of deaths were at least double what would have been expected in the 5 months between March and July 2020. Between September 2020 and March 2021, West St Leonards (Hastings), Hadleigh (Babergh) and Old Oak and Wormwood (Hammersmith and Fulham) had the highest numbers of excess deaths compared with the average for the same months between 2015 and 2019 (65, 57 and 56 excess deaths above the respective averages of 119, 79 and 28).

Risk factors

The increased risk of COVID-19 mortality in men and women from black and South Asian groups compared with people from the White British group is mainly driven by an increased risk of infection. According to REACT-2 COVID-19, the odds of testing positive for antibodies were higher in black (1.6), Asian (1.7), mixed (1.2) and other (1.4) ethnic groups compared with the white population after adjusting for age and sex between 12 May and 25 May 2021. This was due to higher prior infection rates in these groups.^{[footnote 14]}

Post infection, or vaccination, developing antibodies acts as a protective factor, reducing risk of severe COVID-19 and slightly minimising transmission. It is for this reason that vaccinations and boosters are so important, particularly among populations at increased risk of severe COVID-19.

In the first and second waves, the reasons for higher infection rates included living in densely populated urban areas, larger and multi-generational households or working in occupations with higher exposure risk.

As well as differences in infection rates, there were some differences in survival, once infected, between ethnic groups. Analysis of deaths earlier on in the pandemic, between 20 March and 13 July 2020, suggested that once testing positive for COVID-19, Asian and black people had odds of death 1.2 and 1.1 times higher than white people after adjusting for sex, age, deprivation, pre-existing health conditions and region. More analysis is required to understand survival across the pandemic, due to higher infection rates and the emergence of the delta variant in the third wave.

Younger age is a risk factor for infection and older age is the most significant risk factor for severe illness and mortality from COVID-19.

People aged 25 to 49 years had the highest rates of infection per 100,000, between March 2020 and September 2021 (14,837 infections per 100,000 people).

Between 26 September and 6 November, estimated daily infections in England were highest among the youngest age groups – age 2 to school year 6 and school years 7 to 11. ^{[footnote 15]} ^{[footnote 16]}

The average age of death for people whose death involved COVID-19 in the 19 months from March 2020 to September 2021 was 80 years old. ^{[footnote 17]} ^{[footnote 18]}

72% of deaths involving COVID-19 in the same time period were of people aged 75 years and over. It is likely that this is due to an ageing immune system, or an increased likelihood of comorbidities at older ages.

In the period between 24 January 2020 and 31 March 2021, the age-standardised mortality rates for deaths involving COVID-19 were higher among those aged between 65 and 100 years than those aged between 30 and 64 years in every ethnic group.

Figure 4 also shows that nearly all ethnic minority men (excluding the Chinese ethnic group but including the White Other ethnic group) had significantly higher mortality rates per 100,000 person-years ^{[footnote 19]} than White British men across both age groups. Among women, mortality rates were significantly higher for most ethnic minority groups than White British women in both age groups, with the exception of women in the White Other and Chinese ethnic groups, where rates were either not significantly different (women in the Chinese ethnic group, and women in the White Other ethnic group aged over 65 years) or lower than White British women (women in the White Other ethnic group aged 30 to 64 years).

For Chinese men and women, although mortality rates were among the lowest out of all ethnic groups, relative rates per 100,000 person-years were 26 and 36 times higher (respectively) for those aged 65 and over than those aged between 30 and 64 years. These were the highest relative mortality rates between these 2 age groups for any ethnic group.

For men in the Bangladeshi ethnic group, who had the highest overall mortality rates out of all ethnic groups for both age groups, relative rates per 100,000 person-years were 13 times higher for those aged over 65 than for those aged between 30 and 64 years.

Figure 4: Age-standardised mortality rates for deaths involving COVID-19 for men and women, by ethnicity and age group (24 January 2020 to 31 March 2021)

Bar chart showing that the Bangladeshi and Pakistani ethnic groups had the highest age-standardised mortality rates for deaths involving COVID-19 for men and women from 24 January 2020 to 31 March 2021

Figure 4: see this data in a table.

Source: Office for National Statistics

Sex has also been found to be a risk factor for mortality:

Between March 2020 and August 2021, COVID-19 death rates for men were 1.6 times higher than for women.

Global data also indicate higher COVID-19 mortality in men than women, and it is thought that a difference in immune system response could be an important factor in explaining this., ^{[footnote 20]} ^{[footnote 21]}

South Asian ethnic groups are more likely to live in large and multigenerational households:

ONS data shows that in 2018, over-70s in the Bangladeshi and Pakistani ethnic groups were much more likely to have contact with adults and school age children within the same household (56.4% and 34.7% respectively, compared with 1.5% of white adults).

The average household size for people aged 70 and over was 1.7 people in 2018. 27.1% of people aged 70 in the Pakistani ethnic group and 26.8% of those in the Bangladeshi ethnic group lived in households that contained other people aged 70 and over only, the least likely compared with all other ethnic groups. ^{[footnote 22]}

CoMix reported that, in England, there were larger increases in the R rate (reproduction rate) when schools were opened (4 September to 24 October 2020). Over-70s in the Pakistani and Bangladeshi ethnic groups will be disproportionately impacted by this increased source of transmission.

In the first wave (1 February to 31 August 2020), there was no association between living with children and COVID-19 outcomes in all people aged over 65, but in the second wave (defined here as 1 September to 18 December 2020) there was an associated increased risk of infection (adjusted hazard ratio of 1.3), ICU admission (1.9) and COVID-19 mortality (1.4) for adults aged over 65 living with children, ^{[footnote 23]} ^{[footnote 24]} likely related to schools being opened in the second wave (until end December 2020).

Analysis of household composition found that during the second wave of the pandemic (between 1 September 2020 and 31 January 2021) living with younger generations was associated with an increased risk of severe COVID-19 outcomes for people aged 67 years and older in South Asian ethnic groups ^{[footnote 25]}. In contrast, in both the first and second wave, living with younger generations was associated with a lower risk of severe COVID-19 outcomes for people aged over 67 years in the white ethnic group.

OpenSAFELY data indicates that a larger proportion of South Asian people aged 67 years and older lived in households with 1 or more other generations than white people aged 67 years and older (69% compared with 31%). Within the South Asian groups, higher percentages of people aged 67 years and older in Pakistani and Bangladeshi ethnic groups (82% and 84%) lived with one or more generations than people in the Indian ethnic group (66%).

Data from the ONS COVID-19 Infection Survey from 11 October to 7 November 2020 shows a positive association between household size and COVID-19 infection – as household size increased, COVID-19 positivity increased. This association was stronger for ethnic minority people (excluding white minorities) than for white people. At this time schools were open.

The data also shows that between 31 January and 27 February 2021, for white people, living in a multigenerational household was associated with a lower likelihood of testing positive for COVID-19 compared with not living in a multigenerational household. However, among ethnic minority people (excluding white minorities), living in a multigenerational household was associated with a higher likelihood of testing positive compared with not living in a multigenerational household, although this difference was not statistically significant. At this time schools were closed.

PHE analysis of cumulative COVID-19 case rates from March 2020 to October 2021 showed that as age increases from 0 to 24 years, to 65 years and over, the risk of COVID-19 infection for most ethnic minority groups relative to the risk experienced by the white ethnic group increases:

among those aged 0 to 24, people in the Pakistani and Bangladeshi ethnic groups were 0.9 and 0.8 times as likely to become infected as white people
among those aged 25 to 49, people in the Pakistani and Bangladeshi ethnic groups were 1.5 times as likely to become infected as white people
among those aged 50 to 64, people in the Pakistani and Bangladeshi ethnic groups were 2.1 and 2.0 times as likely to become infected as white people
among those aged 65 and over, people in the Pakistani and Bangladeshi ethnic groups were 3.1 and 2.5 times as likely to become infected as white people

As people aged 65 years and over are generally above the working age, the higher relative rate of infection among this age group is more likely to be explained by household transmission. Living in larger, multigenerational households may explain why this relative risk was highest among over-65s in the Pakistani and Bangladeshi ethnic groups.

Figure 5. Cumulative age-standardised COVID-19 case rates per 100,000 people in England, expressed as a relative rate compared with the white ethnic group, by ethnicity and age group (March 2020 to October 2021)

A heatmap showing that people aged 50 and over in the Pakistani and Bangladeshi ethnic groups had the highest cumulative age-standardised COVID-19 case rates per 100,000 people relative to the white ethnic group.

Figure 5: see this data in a table.

Source: Public Health England

However, there are also significant positive impacts of living in a multi-generational household as well as disadvantages of living on your own. According to academic research, multigenerational households among ethnic minorities in the UK were associated with lowest levels of loneliness and greatest quality of life. Elderly relatives often assume child care arrangements, which can be vital for working parents on lower incomes. ^{[footnote 26]} ^{[footnote 27]}

In contrast, research shows that living alone can be associated with common mental health disorders, such as depression.

Ethnic minority groups have higher life expectancies. In the 3 years to March 2014, women from Black African, Bangladeshi and Asian Other ethnic groups had the highest life expectancies (88.9, 87.3 and 86.9 years). The highest life expectancies among men were in the Asian Other, other and Black African groups (84.5, 84.0 and 83.8 years). In the Bangladeshi ethnic group, men had life expectancies 6.2 years lower than women – the largest gap between men and women in any ethnic group.

Geographic factors have been identified as risk factors.

As mentioned earlier, analysis from ONS suggested that living in areas of high population density and local authority district (that is, the area where someone lives) explain the largest part of the disparities in COVID-19 mortality experienced by ethnic minorities.

Areas with high population density, such as major urban conurbations (the most built-up areas), had the highest COVID-19 death rates. From March 2020 to May 2021, the cumulative age standardised death rate was 2.4 times as high in the most densely populated areas in England than the least populated areas (359.2 deaths per 100,000 people in the most densely populated compared with 152.8 deaths per 100,000 in the least populated). ^{[footnote 28]}

These are the type of areas that ethnic minority people were most likely to live in. In 2011 in England and Wales, 76.0% of black people, 69.6% of people in the Other ethnic group, 63.9% of Asian people and 53.9% of people in mixed ethnic groups lived in major urban conurbations. This compares with 28.3% of white people. People from Asian ethnic groups made up the second largest percentage of the population of England and Wales (at 7.5%), followed by black ethnic groups (at 3.3%) and mixed ethnic groups (at 2.2%).

People from ethnic minorities were also more likely to be living in places of enduring COVID-19 prevalence. Enduring prevalence is a term used to describe a repeating pattern of early increasing prevalence of COVID-19 at a local authority level and a slower decline in prevalence than the surrounding local authorities. The particular factors that contribute to enduring prevalence are complicated and vary by area, although there seems to be a strong link with deprivation. Living in an area where there are more infected people, for longer, may create greater capacity to become infected.

For example, Bradford, in which 20% of the population is from the Pakistani ethnic group, had one of the highest number of days spent in the epidemic phase ^{[footnote 29]}. Other local authorities with enduring prevalence also have a notable percentage of Asian people living in them, such as Peterborough, Kirklees, Rochdale, Leicester, Blackburn with Darwen, Luton and Oldham.

Between March and July 2020, COVID-19 mortality rates were 1.7 times higher in neighbourhoods with worse overall air quality than areas with better air quality, after accounting for socio-demographic factors, although some pollutants might be acting as proxies for increasingly urban areas. Poor air quality was also strongly associated with the ethnic diversity of an area – this analysis shows that on average, areas with less diverse ethnic minority populations ^{[footnote 30]} had worse air quality.

Additionally, deprivation was associated with both COVID-19 infection and risk of mortality.

Analysis from ONS shows that between 1 January and 31 December 2020, among those living in the most deprived areas, the age-adjusted mortality rates (AAMRs) for the Pakistani (456.3), Bangladeshi (420.4), Indian (324.0), Black African (307.3), Black Caribbean (304.1) and other (292.9) ethnic groups were significantly higher than the AAMR for the White British (238.5) ethnic group. ^{[footnote 31]} ^{[footnote 32]} ^{[footnote 33]}

Among those living in the least deprived areas, the disparities were similar – the AAMRs for the Bangladeshi (333.1) ^{[footnote 34]}, Black African (237.7), Pakistani (207.9), Black Caribbean (207.6) and Indian (145.2) ethnic groups were significantly higher than the AAMR for the White British (104.6) ethnic group.

For the White British ethnic group ^{[footnote 35]}, AAMRs were significantly different between deprivation quintiles – the more deprived the area, the higher the AAMR (from least to most deprived: 104.6, 117.6, 130.4, 171.7 and 238.5).

Asian populations are over-represented in deprived areas in England, with 15.7% of Asian people living in the most deprived 10% of neighbourhoods. This is more evident in certain Asian populations – 31.1% of people in the Pakistani ethnic group and 19.3% of people in the Bangladeshi ethnic group live in the most deprived 10% of neighbourhoods. According to analysis presented in the third quarterly report, the variance in mortality rate by deprivation appeared greater during ‘peaks’ of the first and second waves.

Looking at different aspects of deprivation and ethnicity, an area’s ethnic minority population and health deprivation ^{[footnote 36]} had significant impacts on COVID-19 mortality between March and July 2020. Lower income appeared to have little effect on COVID-19 mortality when separated from health deprivation.

Occupation is known to be a risk factor for infection.

VirusWatch research shows that certain occupations, such as healthcare workers, indoor trade or transport and mobile machine workers, had at least twice the total odds of seropositivity (presence of antibodies) compared with people employed in other occupations.

In December 2020, research using UK BioBank data and PHE infections data from March to June 2020 found that, after adjusting for age, sex, ethnicity, country of birth, deprivation, education level, shift work, manual work, job tenure, working hours, chronic conditions, long-standing illness or disability, BMI, smoking status and alcohol consumption, healthcare

workers and social and education workers ^{[footnote 37]} had increased odds of getting severe COVID-19 ^{[footnote 38]} ^{[footnote 39]}: compared with people who were not essential workers (7.7 and 1.9 times as likely, respectively). Ethnic minority essential workers, ethnic minority non-essential workers and white essential workers were more at risk of severe COVID-19 than white non-essential workers (6.2, 3.0 and 3.0 times respectively).

In April 2021, analysis of UK BioBank data and PHE infections data from March to August 2020 found again that there were significant differences in the odds of getting severe COVID-19 for healthcare workers compared with people who weren’t. White and South Asian health workers were more likely (2.0 and 6.0 times respectively) than white and South Asian people who are not health workers to get severe COVID-19. ^{[footnote 40]}

Shift working was also found to be a risk factor for infection, ^{[footnote 41]} ^{[footnote 42]}. People who worked shifts were found to be over 4 times as likely to contract COVID-19 as those who didn’t work shifts. ^{[footnote 43]}

Disabled people in England have had a markedly increased risk of mortality involving COVID-19 compared with non-disabled people. ^{[footnote 44]} ^{[footnote 45]}

Disability was identified as a risk factor in the second quarterly report using data referring to a period of 2020 – this analysis has been updated to include the first and second wave, up to February 2021.

Between 24 January 2020 and 28 February 2021, ‘more-disabled’ and ‘less disabled’ ^{[footnote 46]} men had mortality rates 3.1 and 1.9 times higher than non-disabled men. ^{[footnote 47]}

‘More-disabled’ and ‘less disabled’ women had rates 3.5 and 2.0 higher than non-disabled women.

There are some explanations available for the association between disability and COVID-19 mortality. Disabled people are on average older, more likely to become infected as a result of contact in care homes or with carers , ^{[footnote 48]} ^{[footnote 49]} more likely to experience other known risk factors, such as diabetes, more likely to live in socioeconomically disadvantaged conditions or areas and more likely to experience barriers in accessing care. ^{[footnote 50]} ^{[footnote 51]}

While including other risks such as pre-existing health conditions in the analysis did reduce the excess mortality rates for disabled people, some excess risk remains unexplained.

As at the Census 2011, 9.0% of white people, 6.0% of people in other ethnic groups, 5.5% of black people, 5.4% of Asian people, 4.4% of people in mixed ethnic groups in England and Wales were living with a long term health condition or disability that ‘limited day-to-day activities a lot’. Gypsy or Irish Traveller (14.4%), White Irish (13.7%), White British (9.3%) and Black Caribbean (9.0%) people had particularly high proportions of people whose day-to-day activities were ‘limited a lot’.

Between 21 March and 5 June 2021, data from the Learning Disabilities Mortality Review (LeDeR) suggests that the proportions of COVID-19 deaths among people with learning disabilities for Asian (6.5%) and black (3.3%) people with learning disabilities were around 3 times higher than the proportions of average deaths in 2018 and 2019 (2.1% and 1.3% respectively). ^{[footnote 52]} ^{[footnote 53]}

Proportions of deaths from COVID-19 among people with learning disabilities were also higher for Asian and black people than for deaths from all other causes in the same time period (4.4% and 2.7% respectively).

Some pre-existing health conditions are known to impact COVID-19 mortality.

PHE analysed infections and deaths in England between 21 March and 17 July 2020 among people who also had conditions such as diabetes, hypertension, chronic kidney conditions, cardiovascular conditions, respiratory conditions and dementia. Among people with any pre-existing health conditions, the age-standardised mortality rates were higher for all ethnic minority groups compared with the overall death rate for all people with the same conditions ^{[footnote 54]} ^{[footnote 55]}. This analysis only controls for age so it is likely the mortality rates would be lower if also adjusted by other factors.

In the second quarter of 2020 ^{[footnote 56]} (during the first wave of infection and deaths in the UK), COVID-19 age and sex standardised mortality ratios increased among people living with mental health disorders in London, when compared with London’s population (3.8 for people with dementia, 3.3 for people with schizophrenia-spectrum disorders, 4.8 for eating disorders, 5.0 for pervasive developmental disorders, 9.2 for people with learning disabilities and 4.6 for personality disorders). However by the last quarter of 2020 ^{[footnote 57]} mortality ratios were no longer elevated across most psychiatric diagnoses except for dementia, where an increased risk of COVID-19 mortality remained (1.5).

Sickle cell disease and trait were observed to be associated with increased risks of severe COVID-19. Sickle cell disease was associated with a 4.1-fold increased risk of COVID-19 hospitalisation, and a 2.6-fold increased risk of dying due to COVID-19, adjusting for age, ethnicity and sex. In the UK, most people who carry the sickle cell trait have an African or Caribbean family background. ^{[footnote 58]} ^{[footnote 59]}

In the third quarterly report, walking pace (a proxy measure of physical activity) was identified as a risk factor associated with severe COVID-19 infection and COVID-19 mortality.

Between November 2019 and November 2020, the percentage of physically active people in the Asian (excluding Chinese) and black ethnic groups (49.5% and 53.3% respectively) was lower than the percentage of physically active people in the White British group (63.1%).

For White British people, activity levels were 1.5% lower compared with the previous 12 months. For black and Asian people, activity levels were reduced by 4.5% and 4.4% respectively, suggesting that the pandemic has had an impact on physical activity levels and has widened the pre-existing inequality. ^{[footnote 60]} ^{[footnote 61]}

There is some evidence to suggest that lifestyle factors such as smoking are associated with higher risk.

Former and current smokers had higher rates of hospitalisation and death than people who had never smoked ^{[footnote 62]} ^{[footnote 63]} with current smokers experiencing risk of mortality almost 5 times higher than people who had never smoked. There was no evidence to suggest that current smoking increased risk of infection compared with those who had never smoked. ^{[footnote 64]}

In the third quarterly report, religious identity was identified as a potential factor associated with increased risk from COVID-19.

Between 24 January 2020 and 28 February 2021, people identifying as Muslim (men HR: 1.7, women HR: 1.3) or Hindu (men HR: 1.3, women HR: 1.2) and Jewish men (HR: 1.2) and Sikh men (HR: 1.1) had higher rates of death than Christian men and women, ^{[footnote 65]} ^{[footnote 66]} after adjusting for location variables, measures of disadvantage, occupation, living arrangements and pre-pandemic health status. Men and women who identified as ‘no religion’ had lower rates of death than Christian men and women (HR for both men and women: 0.9).

For some religious groups, there is considerable overlap with ethnic background (for example, in the 2011 Census, about 9 in 10 residents identifying with the Pakistani and Bangladeshi ethnic groups also identified as Muslim ^{[footnote 67]}), which makes it difficult to separate the observed association between COVID-19 mortality risk and religion from the risk associated with ethnic background.

Some research suggests that sociability should be considered as a separate factor in the context of increased COVID-19 mortality among religious groups. Sociability in this context is defined as attending communal events such as regular group prayer or on religious celebrations or holidays. Using the number of deaths overseen by British Jewish burial societies ^{[footnote 68]}, analysis showed that in 2020 the number of deaths among Jewish people in England was 28% higher than expected.

Researchers stated that in the past, societies with higher poverty and/or overcrowding have been associated with the spread of communicable diseases. However in the context of COVID-19 and excess mortality among Jewish people, they considered the impact of social and religious involvement on greater exposure to COVID-19 and concluded that sociability should be considered independently from the poverty or socioeconomic disadvantage often associated with ethnic minority, and migrant, communities.

It is also worth noting that places of worship were closed in the first lockdown and open in the second therefore the extent to which religious sociability was a risk factor may have changed between the time periods.

Meeting or socialising with people outside of the household could also be a risk factor. ONS analysis of the Opinions and Lifestyle survey (OPN) shows that between 21 July and 15 August 2021. ^{[footnote 69]} ^{[footnote 70]}

Higher proportions of white people socialised indoors with people who were not in their household (64%) than people from other (39%) and mixed ethnic groups (49%), black people (45%) and Asian people (53%).

Higher proportions of white people socialised outdoors with people who were not in their household (66%) than black (45%) and Asian people (54%). ^{[footnote 71]}

Among those who met up or socialised with someone not in their household, the majority of people in each ethnic group met up or socialised with between approximately 1 and 5 other people.

Among those who met up or socialised with someone not in their household, people in other, Asian and black ethnic groups were more likely to meet in a public outdoor space such as a park (82%, 72% and 66%) than a private outdoor space such as a private garden (35%, 41% and 34%) ^{[footnote 72]}. The inverse is true for white people, who were more likely to meet in private outdoor spaces than public ones (56% compared with 50%).

Between the period 31 March to 25 April 2021 and 21 July to 15 August 2021, socialising indoors and outdoors with someone from outside of the household increased for all ethnic groups. ^{[footnote 73]}

The first and second waves of the COVID-19 pandemic has shown some of the same pattern as the flu pandemic in 2009.

In the flu pandemic in 2009, people in the Pakistani ethnic group in England experienced death rates 3.4 times higher than white people., ^{[footnote 74]} ^{[footnote 75]}

Between 2008 and 2018, rates of flu or flu-like illness were higher among Bangladeshi (2.1 times higher), Pakistani (1.9) and Black African (1.5) ethnic groups, compared with the White British group. ^{[footnote 76]} ^{[footnote 77]}

Research into why also found deprivation to be strongly associated with death from flu and larger or multigenerational households to be associated with flu infections, which potentially went some way to explain the increased risk for ethnic minorities, in particular Asian groups.

Most of the risk factors already identified explain a lot of the increased mortality risk observed for some ethnic groups, but not all of it. The remaining unexplained risk will likely be due to factors not known or not able to be included in analyses – for example, factors such as transport use by ethnic groups. There is also some evidence to suggest that genetic differences may play a role for some ethnic groups.

Previous quarterly reports stated that ethnicity is not considered a risk factor in and of itself. We have explored this further and considered whether ethnicity should be viewed as a risk factor in Annex E.

Question 3

Long COVID

Accepted Answer

NICE has identified 3 phases post COVID-19 infection, the latter 2 of which are commonly described as long COVID:

Acute COVID-19 – signs and symptoms of COVID-19 for up to 4 weeks
Ongoing symptomatic COVID-19 – signs and symptoms of COVID-19 for between 4 and 12 weeks
Post COVID-19 syndrome – signs and symptoms of COVID-19 that continue for more than 12 weeks and are not explained by an alternative diagnosis ^{[footnote 78]}

The ONS research summarised here is based on people’s self-reported experiences of long COVID, as opposed to clinically-diagnosed long COVID ^{[footnote 79]}. People can experience long COVID for different periods of time. To measure this, ONS typically uses 3 time periods in their analysis:

the number of people with long COVID who first experienced COVID-19 symptoms 12 weeks previously
the number of people with long COVID who first experienced COVID-19 symptoms 12 months previously
the number of people with long COVID of any duration (having first experienced COVID-19 symptoms at least 4 weeks previously)

The self-reported prevalence rate of long COVID reported by the ONS remains slightly higher for the white ethnic group than the Asian group, as was reported in the third quarterly report ^{[footnote 80]}:

for the period between 4 October and 31 October 2021, 1.36% of the white population of the UK reported that they had long COVID after first experiencing COVID-19 symptoms at least 12 weeks previously, significantly higher than 1.04% of the Asian population but not black (1.02), other (1.02%) or mixed (1.25%) ethnic groups
for the period between 5 September and 2 October 2021, the prevalence rates for the white and black ethnic groups were 1.33% and 0.83% respectively
for the period 9 August to 5 September, there were no significant differences in prevalence rates between any of the ethnic groups
for the period 5 July to 1 August 2021, the prevalence rates for the white and Asian ethnic groups were 1.28% and 0.98% respectively
for the period 7 June to 4 July 2021, the prevalence rates for the white and mixed ethnic groups were 1.31% and 0.90% respectively
for the period 10 May to 6 June 2021, the prevalence rates for the white and Asian ethnic groups were 1.34% and 1.06% respectively
for the period 5 April to 2 May 2021, the prevalence rates for the white and Asian ethnic groups were 1.37% and 0.95% respectively

For the period between 4 October and 31 October 2021, 0.69% of the white population self reported long COVID after first experiencing COVID-19 symptoms at least 12 months previously. 0.47% of the other ethnic group, 0.54% of the Asian ethnic group, 0.55% of the black ethnic group and 0.57% of the mixed ethnic group self reported long COVID after first experiencing COVID-19 symptoms at least 12 months previously but there were no significant differences between any ethnic groups.

0.37% of white, 0.30% of Asian, 0.29% of other, 0.23% of mixed and 0.22% of black people said that they had long COVID of any duration and reported that their day-to-day activity was limited a lot, though there were no significant differences between any ethnic groups

Between 4 October and 31 October 2021, women were significantly more likely than men to report that they had experienced long COVID of any duration (2.17% of women compared with 1.56% of men).

Socio-demographic characteristics such as ethnicity, age and occupation can sometimes be highly correlated with one another (for example, health status is related to age). Furthermore, patterns in the prevalence rates of long COVID could reflect differences between ethnic groups in the risk of COVID-19 infection and in the risk of developing prolonged symptoms following infection. For these reasons, ONS previously calculated the “odds” of ethnic minority groups reporting long COVID after first experiencing COVID-19 symptoms at least 12 weeks previously compared with the white group in 3 different models, for data collected between the 5 April and 2 May 2021 ^{[footnote 81]}:

Unadjusted Adjusted (for other socio-demographic characteristics) Adjusted and confirmed COVID-19 cases only ^{[footnote 82]}

In the unadjusted model, people from the Asian, black and mixed groups are less likely than white people to report long COVID. In both the second and third models, Asian people were less likely than white people to report long COVID (0.80 and 0.59 times as likely, respectively). All other results for each ethnic group and model were not significantly different from the result for the white group. The results from separate analysis conducted by University College London also show that people of “non-white” ethnicity are less likely than white people to report long COVID after first experiencing symptoms at least 12 weeks previously, with an odds ratio of 0.32. ^{[footnote 83]}

Figure 6: Odds ratios ^{[footnote 84]}, for self-reported long COVID of at least 12 weeks compared with the white ethnic group, by ethnicity: detailed employment sector data ^{[footnote 85]} (5 April to 2 May 2021)

Graph showing that people of “non-white” ethnicity are less likely than white people to report long COVID after first experiencing symptoms at least 12 weeks previously.

Figure 6: see this data in a table.

Source: Office for National Statistics

From Autumn 2021, UCL’s ‘Convalescence Long-COVID Study’ will conduct in-depth interviews with 80 participants from across the UK who have experienced long COVID to understand their experiences of and perspectives on the condition (including 40 participants from the Born in Bradford study). UK Research and Innovation (UKRI) and the National Institute for Health Research (NIHR) are funding 4 research studies into long COVID in the community, including projects testing therapies for long COVID, and assessing the characteristics, determinants, mechanisms and consequences of long COVID. This will provide an evidence base for health care services and response. In total, the government is providing £50 million in research funding on long COVID.

Question 4

Vaccinations

Accepted Answer

The UK’s COVID-19 vaccination rollout programme began in December 2020, with older adults, care home residents and frontline health and social care workers prioritised for vaccination. As of 18 June 2021, all adults aged 18 and over became eligible to receive a COVID-19 vaccination in England. This was extended to those aged 16 and over in August.

Vaccine sentiment

Vaccine confidence has increased in every ethnic group from the period between December to January and June to July 2021 ^{[footnote 86]} ^{[footnote 87]}, although black ethnic groups have been consistently shown to have greater vaccine hesitancy than other broad ethnic groups. The vaccine confidence gap between black people and people from other broad ethnic groups has narrowed since December.

96% of adults aged 16 and over in Great Britain reported a positive vaccine sentiment between 23 June and 18 July 2021, according to the latest survey data from ONS. Vaccine confidence ranged from 93% to 96% in the mixed, Asian, other and white ethnic groups, and was lowest among the black ethnic group (79%).

Figure 7: Percentage of people who said they were likely to accept or had already accepted the COVID-19 vaccine, by ethnicity and research period

Bar chart showing that vaccine confidence went up in every ethnic group in the 6 months to July 2021, and the increase was biggest in the black and other ethnic groups.

Figure 7: see this data in a table.

Source: Office for National Statistics

Research from the University of Birmingham found that between 4 September and 9 October 2020, people from an ethnic minority background had 5.5 times the odds of white people to be interested in taking an approved COVID-19 vaccine.

Online polling from 25 January to 1 February 2021 found that 9% of people from an ethnic minority background and 8% of those from a white background said they would not get a COVID-19 vaccine. A higher proportion of ethnic minority people than white people said they would “wait before getting a COVID-19 vaccine to see what happens” (26% and 17% respectively).

Virus Watch analysis found that, of those who said that they would not, or were unsure about, accepting a vaccine in December 2020, 86% reported that they would accept or had already accepted a vaccine in February 2021. The magnitude of this shift was consistent across all ethnic groups measured, ranging from 72% of people from mixed ethnic backgrounds to 90% of people from South Asian backgrounds.

Data from REACT-2 suggest that from 12 to 25 May 2021, compared with the white population, the odds of vaccine hesitancy were higher in people from black, other and mixed ethnic backgrounds (2.1, 1.9 and 1.6 respectively) and lower in Asian people (0.8) after adjusting for age and sex. This suggests there is no uniform ethnic minority trend in vaccine hesitancy.

Recent ONS analysis shows that among adults who were previously vaccine hesitant, a similar percentage of black (47%) and white (42%) adults had received at least one dose of the COVID-19 vaccine by September 2021. 54% of Asian, 53% of other and 33% of mixed ethnicity adults who were vaccine hesitant went on to receive at least one dose.

Vaccine uptake

The COVID-19 vaccines currently available are administered in 2 doses. The second dose is usually given 8 to 12 weeks after the first dose. As at 10 November 2021, 79.9% of the over-12 population in the UK had received 2 doses of the COVID-19 vaccination, and a further 7.8% of the population had received one dose. 19.9% of the population had received a booster or third dose.

Vaccine uptake estimates from NHS England suggest that 89.6% of White British over-18 year olds had received at least one COVID-19 vaccination by 31 October 2021 – the highest rate of uptake out of all ethnic groups. The lowest rates of vaccine uptake in over-18 year olds were among the Chinese (50.8%), Black Caribbean (54.6%) and Black Other (56.8%) ethnic groups. Vaccination estimates for Gypsy, Roma and Irish Traveller populations are not available as NHS systems do not have these categories available to choose.

The pattern of vaccine uptake is similar among over-18 year olds who had had both doses of a COVID-19 vaccination by 31 October 2021. The highest rate of uptake was among the White British ethnic group (86.1%) and the lowest was among the Chinese (47.9%), Black Other (49.4%) and Black Caribbean (49.5%) ethnic groups.

It is important to note that older age cohorts have been prioritised for vaccine administration, and that certain ethnic groups (such as the Chinese group) have a younger age profile than other ethnic groups. Therefore, when considering how vaccine uptake varies by ethnicity, looking at rates among over-50s potentially provides greater insight, at least until younger people have been vaccinated to the same rate as older people.

Among over-50 year olds, the White British ethnic group had the highest rate of uptake out of all ethnic groups for both those who have had at least one COVID-19 vaccination (95.6%) and those who have had both doses (94.3%) by 31 October 2021. The lowest rates of vaccine uptake were among the black ethnic groups when looking at both over-50 year olds who have had at least one COVID-19 vaccination (Black Caribbean (68.3%), Black Other (71.9%), Mixed and White Black Caribbean (74.8%), Black African (75.1%%)), and over-50 years old who have had both doses (Black Caribbean (64.4%), Black Other (67.4%), Black African (70.6%)).

Between 7 April and 31 October 2021, the percentage of over-50s who had received at least one COVID-19 vaccine increased in all ethnic groups. The largest percentage point increases were in the Pakistani ethnic group (from 73.1% to 83.7%, up by 10.6 percentage points) and Black African ethnic group (from 64.9% to 75.1%, up by 10.2 percentage points).

Figure 8a: Percentage of people aged over 50 years who had received at least one COVID-19 vaccine by 7 April and by 31 October 2021, by ethnicity

Bar chart showing that, among over-50s, Black Caribbean people were least likely to have had at least one COVID-19 vaccine, and White British people were most likely.

Figure 8a: see this data in a table.

Source: NHS England

Between 31 May and 31 October 2021, the percentage of over-50s who had received both doses of the COVID-19 vaccine increased in all ethnic groups. The largest percentage point increases were in the Pakistani ethnic group (from 54.2% to 78.8%, up by 24.6 percentage points) and Bangladeshi ethnic group (from 63.7% to 87.0%, up by 23.3 percentage points).

Figure 8b: Percentage of people aged over 50 years who had received both COVID-19 vaccines by 31 May and by 31 October 2021, by ethnicity

Bar chart showing that, among over-50s, black people were least likely to have had both COVID-19 vaccines.

Figure 8b: see this data in a table.

NHS vaccinations data also shows regional variations in vaccine uptake up to 31 October 2021. Among over-50 year olds, 96.4% of White British people had received at least one dose of the vaccination in the East of England and 65.6% of Black Caribbean people had received at least one dose in London – the highest and lowest rates of uptake respectively (out of all combinations of ethnic groups and regions). For people in the white, mixed, black and Chinese ethnic groups, the lowest rate of uptake was in London. The Bangladeshi group was the only ethnic group for which London was the region with the highest rate of vaccine uptake, at 91.7%. For most other ethnic groups, the highest rates of vaccine uptake were in the East of England, the South West or the South East.

According to Our World in Data, the UK has one of the highest rates of COVID-19 vaccination in the world, in terms of the share of the population that has received at least one dose of the vaccine. As at 9 November 2021, 73.8% of the UK population ^{[footnote 88]} had received at least one dose of the COVID-19 vaccination compared with vaccination rates in Canada (78.7%), Japan (78.5%), Italy (77.5%), France (76.1%), Germany (69.2%) and the United States (66.8%).

Analysis by the American non-profit organisation Kaiser Family Foundation (KFF) estimates that at 1 November 2021, 48.0% of the total black population in 43 US states had received at least one dose of a vaccine ^{[footnote 89]} ^{[footnote 90]}. In comparison, 63.7% of the Black African population, 54.6% of the Black Caribbean population and 56.8% of the Black Other population aged 18 and over had received at least one dose in England, as estimated in NHS vaccine statistics up to 31 October 2021.

KFF’s analysis also found that 55.0% of the total white population of these states had received at least one dose of the vaccine. This compares to 89.6% of the White British population, 80.0% of the White Irish population and 62.5% of the White Other population, as estimated by the NHS.

Outcomes following vaccination

Research suggests that the risk of COVID-19 infection after a first dose of the vaccine is associated with deprivation (and deprivation is associated with ethnic diversity). The risk of post-vaccination infection was highest for people living in areas with higher levels of deprivation and lowest for those living in areas with lower levels of deprivation. Almost all COVID-19 symptoms were reported less in people infected and vaccinated, compared with those infected and unvaccinated.

Recent analysis from ONS suggests that between 3 February and 5 September 2021, receiving a first dose of the COVID-19 vaccination was associated with a 12.8% decrease in the odds of self-reported long COVID (symptoms persisting for at least 12 weeks after first having COVID-19 that were not explained by something else) among study participants aged 18 to 69 years in the UK. Receiving a second dose was associated with an 8.8% decrease in the likelihood of self-reported long COVID, relative to having received the first vaccination, and there was statistical evidence of a sustained improvement after this. There was no statistical evidence of differences in post-vaccination long COVID trends according to socio-demographic characteristics (including ethnicity) or health-related factors.

Analysis found that a first dose of a COVID-19 vaccination reduced the risk of COVID-19 death by 52.6% (95% CI 26.6% to 84.2%) in those aged 80 to 84, compared with those aged 75 to 79 who had lower vaccination rates at the time, suggesting the vaccination is effective in protecting against COVID-19. Recent ONS analysis found that, of all deaths involving COVID-19 between 2 January and 2 July 2021, 1.2% occurred in people who were fully vaccinated.

Compared with the white ethnic group, Pakistani and Indian groups were 2.5 and 1.3 times as likely to be admitted to hospital with COVID-19 and 2.3 and 1.6 times as likely to die from COVID-19 after vaccination ^{[footnote 91]} ^{[footnote 92]}. The research concludes that “These ethnic disparities in COVID-19 outcomes could represent residual differential exposure (for example, linked to behaviour, lifestyle, household size, and occupation) more than differential susceptibility mechanisms, although we also acknowledge that being vaccinated could change behaviour (and exposure) in some groups more than in others.”

Question 5

Progress with the first batch of NIHR-funded research projects

Accepted Answer

A team led by Professor Julia Hippisley-Cox (University of Oxford) and Dr Hajira Dambha-Miller (University of Southampton) have undertaken several analyses to quantify the association between ethnicity and COVID-19.

In a cohort study of over 2.5 million children, testing varied across ethnic groups (17.1% white children were tested compared with 13.6% of Asian, 8.3% of black and 12.9% of mixed or other children). Asian children were more likely to be admitted to hospital for confirmed COVID-19 (adjusted OR 1.62 (1.12-2.36) Asian) and children from ethnic minority groups were significantly more likely to be admitted to intensive care due to COVID-19 (adjusted OR 2.11 (1.07 to 4.14) Asian, 2.31 (1.08 to 4.94) black, 2.14 (1.25 to 3.65) mixed or other).

Sickle cell disease and trait were observed to be associated with increased risks of severe COVID-19. Sickle cell disease was associated with a 4.1-fold increased risk of COVID-19 hospitalisation, and a 2.6-fold increased risk of dying due to COVID-19, adjusting for age, ethnicity and sex. In the UK, most people who carry the sickle cell trait have an African or Caribbean family background. ^{[footnote 93]} ^{[footnote 94]}

The main analysis of COVID-19 risks across ethnic groups in both waves of the pandemic has been concluded, with comparisons made between England (9.8 million adults) and data from Ontario, Canada (10.3 million). Meta-analysis was undertaken. Those from South Asian ethnic groups had higher risks of COVID-19 death (HR 1.63, 95% CI: 1.09-2.44), hospitalisation (1.53, 1.32 to 1.76) and ICU admission (1.67, 1.23 to 2.28). In the English data, it was estimated that sociodemographic, lifestyle and clinical factors accounted for 60.3% (mixed), 43.8% (South Asian) and 39.6% (black) of the excess risks of COVID-19 death.

Other work is currently under peer review, including ethnic patterns in uptake of non-COVID vaccines in older adults (aged over 65) across ethnic groups, and effects of prior receipt of influenza, pneumococcal and shingles vaccines on risks of COVID-19 hospitalisation and mortality. The group is continuing to evaluate the effects of specific drugs – for example, medications for diabetes – on COVID-19 outcomes among ethnic minorities.

The Virus Watch project, led by Dr Robert Alrdridge (University College London), began recruiting in August 2020, with a focus on the recruitment of people from ethnic minority backgrounds and migrants. As of 14 May 2021, 24,322 households and 50,774 people across England and Wales had joined the study, with 7,839 (15%) of the cohort from ethnic minority backgrounds and 5,318 (10%) people born outside of the UK. 126,414 weekly surveys have been completed by Virus Watch participants from ethnic minority backgrounds since June 2020, and as a result there are just under 1-million-person days of follow-up of people from ethnic minority backgrounds.

Between October 2020 and January 2021, full venous samples were collected from 6,243 participants (of which 18% were from ethnic minority backgrounds) and tested for COVID-19 spike (S) antibodies. 30,437 finger prick samples were collected from 11,538 participants (of which 9% were from ethnic minority backgrounds) and tested for COVID-19 spike (S) and nucleocapsid (N) antibodies. Virus Watch has experienced delays from NHS Digital in processing the linked data from participants in the community to hospital and death data. As a result, they are seeking an extension from UKRI.

The team have also completed the following analyses using unlinked data:

Appendix D: Further data and evidence

Differences between the first, second and third waves

Hazard ratios

Annualised age-adjusted mortality rates

Cumulative age-adjusted mortality rates

Age-adjusted case rates

Excess deaths

Risk factors

Long COVID

Vaccinations

Vaccine sentiment

Vaccine uptake

Outcomes following vaccination

Progress with the first batch of NIHR-funded research projects

Is this page useful?

Help us improve GOV.UK

Help us improve GOV.UK

Cookies on GOV.UK

Differences between the first, second and third waves

Hazard ratios

Annualised age-adjusted mortality rates

Cumulative age-adjusted mortality rates

Age-adjusted case rates

Excess deaths

Risk factors

Long COVID

Vaccinations

Vaccine sentiment

Vaccine uptake

Outcomes following vaccination

Progress with the first batch of NIHR-funded research projects

Is this page useful?

Help us improve GOV.UK

Help us improve GOV.UK