Ethnic disparities in COVID-19 mortality and cardiovascular disease in England and Wales between 2020-2022

Categories: Disease & Virus

July 2, 2025

Ethnic diversity in individuals diagnosed with COVID-19 in England and Wales

We identified 4,867,595 (60% women) individuals in England (Supplementary Fig. 1) and 451,077 (55% women) in Wales (Supplementary Fig. 2) who were registered in a General Practice for at least one year, were aged 30 years or older, and had a confirmed record of COVID-19 diagnosis. In England, the high-level ethnicity distribution was as follows: White (83.0%), Asian/Asian British (8.0%), Black/Black British (3.2%), Mixed (1.3%), and Other Ethnic Group (1.3%). In contrast, Wales was less diverse, where the distribution was: White (92.1%), Asian/Asian Welsh (2.5%), Black/Black Welsh (0.7%), Mixed (0.7%) and Other Ethnic Group (0.8%). Ethnicity was not reported for 3.3% of the individuals, in both England and Wales (Supplementary Table 3).

Ethnicity in the Welsh population was also available in 10 NER ethnic groups classification (Supplementary Table 4), whilst the larger size of the English population permitted us to study ethnicity more granularly, including the 19 NHS ethnicity codes and SNOMED-CT concepts. Baseline characteristics for the high-level ethnic groups and their corresponding 19 sub-groups in England are reported in Supplementary Table 5 and Supplementary Table 6, respectively.

Mean age at diagnosis of COVID-19 infection in England ranged from 43.67 years (standard deviation [SD]: 12.56) in the Bangladeshi to 55.70 years (SD: 16.80) in the Irish populations (Supplementary Table 6); whilst in Wales, it ranged from 43.55 (SD: 11.57) in the Mixed group to 54.56 (SD: 16.38) in the Unknown group (Supplementary Table 4).

In England, those with Pakistani ethnicity had the highest proportion of individuals living in the most deprived areas (46.8% in the lowest index of multiple deprivation [IMD] fifth), followed by Bangladeshi (41.2%), African (39.5%), Arab (36.8%), Caribbean (35.3%), Any other Black background (35.1%), White and Black Caribbean (32.9%), White and Black African (32.2%), which were well above White British (19.6%), Indian (17.5%) and Chinese (17.2%) populations (Supplementary Table 6). In Wales, the ethnic group with the highest proportion of individuals from the most deprived areas was Black African (46.9%), followed by Pakistani (38.0%), Black Caribbean (37.1%), Other Ethnic Group (33.7%), Bangladeshi (31.5%) and Mixed (31.4%) (Supplementary Table 4).

Incidence and hazard ratio differences in COVID-19 mortality and CVD between and within ethnicity groups

Due to the low number of individuals from non-White groups in the Welsh population and their lower number of outcome events observed, we report estimates for Wales using the 6-level ethnicity categorisation. More granular results for Wales are presented in Supplementary Table 7 and Supplementary Fig. 4.

Incidence rates [IR]

28-day COVID-19 mortality (age-standardised IR [95%CI] per 100,000 population/year)

In England, all non-White ethnic groups, except those with missing ethnicity, had higher incidence of mortality than White (Fig. 1A). Conversely, those with Unknown ethnicity had the highest mortality rates in Wales, and only Asian/Asian British group, and men self-identified as Black/Black British and Mixed show a significant incremented age-standardised IR compared to White population (Fig. 1C).

**Fig. 1: Age-standardised incidence rates (per 100,000 population/year) of 28-day mortality and 30-day CVD in England and Wales.**

A larger disparity in mortality rates is observed on the 19-level ethnicity group classification than using the broader categories in the English population, such as within the Asian/Asian British population, where mortality incidence in Bangladeshi (men: 116.8 [106.9 to 126.6], women: 65.5 [58.5 to 72.4]) were higher than Pakistani (men: 81.3 [77.0 to 85.6], women: 49.7 [46.4 to 52.9]) and Indian (men: 64.6 [61.5 to 67.7], women: 39.5 [37.1 to 42.0]).

At the most granular level of ethnicity classification available (SNOMED-CT classification, available only in England), we observed large differences between Central/South/Latin American (men: 178.5 [153.7 to 206.2], women: 55.2 [42.1 to 71.3]) and Iranian (men: 33.4 [23.4 to 46.2], women 10.2 [5.5 to 18.4]) within Other Ethnic Group; and between Nigerian and Somali men (137.0 [115.9 to 162.0] and 89.9 [72.8 to 110.0], respectively).

30-day CVD (age-standardised IR [95%CI] per 100,000 population/year)

Four ethnic groups of the high-level classification in England (i.e., Asian/Asian British, Mixed, Black/Black British, and Other Ethnic Group) were more likely to experience a CVD than the White group (Fig. 1B). In Wales, confidence intervals were wide for ethnic groups other than White, thus, only Asian/Asian British women show an incremented incidence (Fig. 1C).

Consistent with mortality incidences, the 19-level ethnic group classification in England showed different CVD incidence within the high-level ethnic group classification, such as higher CVD incidence in Pakistani (men: 85.03 [80.98 to 89.08], women: 39.2 [36.5 to 41.9]) and Bangladeshi (men: 88.74 [81.12 to 96.37], women: 38.3 [33.5 to 43.1]) sub-groups vs the other Asian sub-groups.

At the most granular level, SNOMED-CT ethnicity concepts revealed a larger incidence among Turkish/Turkish Cypriot (men: 93.2 [75.3 to 111.1], women: 44.2 [32.5 to 58.7]) and among “Middle Eastern” women (excluding Israeli, Iranian and Arab, age-standardised IR: 73.5 [58.1 to 91.8]), compared with their corresponding ethnic group in high-level classification (i.e., Other Ethnic Group).

Supplementary Table 8 summarises all age-standardised IR estimates for 28-day mortality and 30-day CVD for England, whilst Supplementary Table 7 summarises it for Wales.

Hazard Ratios [HR] (with White as reference group)

Survival analyses for England showed that the differences in mortality (Fig. 2A) and CVD (Fig. 2B) observed in age-standardised IR were maintained even when adjusted by age, IMD, vaccination status, pregnancy, geographical location in England, time of diagnosis, comorbidities, and medication/s use. In Wales, increased risk of mortality was confirmed in Asian/Asian British and Unknown, and increased CVD risk in Asian women and men with unknown ethnicity (Fig. 3).

**Fig. 2: Adjusted hazard ratios of 28-day mortality and 30-day CVD in England.**

**Fig. 3: Adjusted hazard ratios of 28-day mortality and 30-day CVD in Wales.**

Supplementary Table 9 and Supplementary Table 10 summarises all HR estimates for 28-day mortality and 30-day CVD, and their adjustments, respectively for England. Supplementary Table 11 and Supplementary Table 12 have the HR estimates for 28-day mortality and 30-day CVD, and their adjustments, respectively for Wales.

Incidence and hazard ratio of COVID-19 mortality and CVD across and within ethnicity groups: trend over time

Reduced number of outcomes in Wales impairs their reliability when stratified over time. Thus, the following estimates reporting incidence and survival over time are focused on England population. Wales IR estimates over time are included in Supplementary Table 13 and Supplementary Fig. 5.

Incidence rates in England

Supplementary Table 14 summarises England’s age-standardised IR estimates for 28-day mortality and 30-day CVD over time. Population size for ethnic groups observed through the SNOMED-CT concepts were too small to be explored across time.

28-day mortality (age-standardised IR [95%CI] per 100,000 population/year)

There was an overall decrease in mortality incidence from January 2020 to June 2022. At the beginning (from January to June 2020), the age-standardised IR for non-White ethnic groups (except those with Unknown ethnicity) were higher than the White group. This difference (with respect to the White group) fluctuated in magnitude but remained during the subsequent 18 months, disappearing only in the final 6 months until April 2022 (Fig. 4).

**Fig. 4: Age-standardised incidence rates of 28-day mortality (per 100,000 population/year) by period of recorded COVID-19 diagnosis in England.**

The 19 sub-groups display better the fluctuations in the mortality rates across the different ethnic sub-groups, where Bangladeshi and Pakistani populations stand out by always having higher mortality rates than White British during the first two years of the pandemic (i.e., from 23^rd January 2020 until 31^st December 2021). Despite the overall decrease and the closed mortality gap for most of ethnic groups in the last study period (1^st January to 1^st April 2022), Pakistani men still presented an incremented incidence (14.97 [8.74 to 24.08]) compared to with British (5.83 [5.42 to 6.23]).

30-day CVD (age-standardised IR [95%CI] per 100,000 population/year)

Age-standardised IR of CVD were generally higher during the first 6 months of the pandemic (i.e., 01st January to 30^th June 2020), similar from July 2020 to June 2021, and slightly lower after July 2021. Likewise, with mortality, inequities in CVD incidence (with respect to the White group) varied over the studied period. However, the gap between distinct non-White groups compared to the White British was maintained in the final 6 months until April 2022 (Fig. 5).

**Fig. 5: Age-standardised incidence rates of 30-day CVD (per 100,000 population/year) by period of recorded COVID-19 diagnosis in England.**

Fluctuations in CVD rates can be better represented through the 19 sub-groups. Within Arian/Asian British, Bangladeshi and Pakistani populations constantly emerged as presenting incremented CVD rates when compared to White British over time, whilst the Chinese population was not significant incremented in the final 6 months (i.e., 1^st January to 1^st April 2022). Within Black/ Black British, Caribbean women presented continuous incremented rates during the full study period, and Caribbean men after the initial 12 months (i.e., 1^st January 2021 to 1^st April 2022), whilst rates in African women were only incremented at the initial 6 months and rates in African men were only incremented at the final 6 months.

Hazard Ratios (with White as reference group) in England

28-day mortality

During the first 6 months of the pandemic, women from Other Ethnic Group (HR [95%CI]: 1.31 [1.11 to 1.54]), and individuals from Asian/Asian British (HR [95%CI]: 1.19 [1.12 to 1.27] in men, 1.30 [1.19 to 1.42] in women) and Mixed (HR [95%CI]: 1.24 [1.05 to 1.45] in men, 1.24 [1.03 to 1.51] in women) had an increased risk of mortality post COVID-19 as compared to White population. Whilst those with an increased mortality in the last 6 months of the study period were women from Other Ethnic Group (HR [95%CI]: 2.06 [1.09 to 3.88] in women), and Asian/Asian British (HR [95%CI]: 1.40 [1.08 to 1.82] in men, 1.52 [1.08 to 2.12] in women) (Supplementary Fig. 6).

Considering the 19 ethnic sub-groups, mortality risk (Fig. 6) was increased in Bangladeshi and Pakistani from the onset of the pandemic until end of December 2021, where the lower HR [95%CI] of Bangladeshi men and women were 1.55 [1.33 to 1.81] and 1.47 [1.22 to 1.77], respectively, and risk estimates in Pakistani ranged from 1.15 [1.04 to 1.28] to 1.33 [1.15 to 1.55] in men and 1.16 [1.02 to 1.32] to 1.54 [1.29 to 1.84] in women. We observed other differences, such as an increased mortality during the first 12 months of the pandemic (from 23rd January to 31st December 2020) in Any Other Black background, and in men self-identified as Indian or White and Black Caribbean; in Indian women during the first 6 months, or in White and Black Caribbean women during the first 6 months and from 1st July to 31st December 2021.

**Fig. 6: Adjusted hazard ratios of 28-day mortality of the 19 NHS ethnicity groups by months of recorded COVID-19 diagnosis, using White ethnicity as the reference group, in England.**

No significant differences in 28-day mortality risk were detected after January 2022 (i.e., 1st January to 1st April 2022) (Supplementary Table 15).

30-day CVD

During the first 6 months, only women from Mixed (HR [95%CI]: 1.89 [1.43 to 2.49]) and Asian/Asian British (HR [95%CI]: 1.20 [1.04 to 1.38]) high-level groups had a significant increased risk of CVD compared to the White group. Whilst in the last 6 months increased CVD risk was observed in women from Other Ethnic Group (HR [95%CI]: 1.83 [1.34 to 2.52]) and Black/Black British (HR [95%CI]: 1.39 [1.14 to 1.71]). Conversely, men from Asian/Asian British and Other Ethnic Groups show an increased CVD risk from July 2020 to December 2021 (Supplementary Fig. 7).

When observing the initial 6 months of the study using the 19 NHS ethnicity codes (Fig. 7), we could detect an increased CVD risk in Arab women (HR [95%CI]: 3.80 [1.42 to 10.18]), and we observed that only Bangladeshi women (HR [95%CI]: 1.86 [1.23 to 2.80]) within the Asian/Asian British groups, and women from the White and Black Caribbean (HR [95%CI]: 2.46 [1.63 to 3.73]) and White and Black African (HR [95%CI]: 2.30 [1.30 to 4.06]) within the Mixed groups had an incremented CVD. When observing the male sub-groups during the period of July 2020 to December 2021, only Pakistani and Bangladeshi were increased within the Asian/Asian British, and Any other ethnic group within Other Ethnic Group.

**Fig. 7: Adjusted hazard ratios of 30-day CVD of the 19 NHS ethnicity groups by months of recorded COVID-19 diagnosis, using White British ethnicity as the reference group, in England.**

As a highlight, in the last 6 months of the study, the following sub-groups had a remaining increased CVD risk: men from Pakistani (HR [95%CI]: 1.58 [1.32 to 1.90]), White and Asian (HR [95%CI]: 2.20 [1.28 to 3.80]) or Any other White background (HR [95%CI]: 1.17 [1.04 to 1.31]), and women from Bangladeshi (HR [95%CI]: 1.75 [1.13 to 2.72]), Caribbean (HR [95%CI]: 1.55 [1.19 to 2.02]), or Any Other Ethnic Group (HR [95%CI]: 1.82 [1.31 to 2.53]) (Supplementary Table 16).

Source link