Research and analysis

Inequalities in emergency hospital admission rates for influenza and COVID-19, England: September 2022 to February 2023

Published 7 December 2023

Applies to England

What is known about this topic

Across multiple health hazards, there are differences between communities and population groups in terms of their risk of exposure to external health hazards and susceptibility to poor outcomes when exposed. The underlying causes of these inequalities are complex, with contributing factors including the impact of existing social and economic inequalities and pre-existing health conditions.

There are known inequalities in vaccination coverage between different population groups across multiple infectious diseases. Of relevance to this report, there are significant and persistent disparities in vaccine uptake for COVID-19 and influenza by ethnic group and deprivation level as seen in the COVID-19 vaccine quarterly surveillance reports and annual UK influenza surveillance reports.

What this report adds

This report describes inequalities in emergency hospital admission rates for influenza and COVID-19 in England, between deprivation levels and ethnic groups, between September 2022 and February 2023. The report forms part of the UK Health Security Agency’s (UKHSA) strategy to achieve more equitable health outcomes by identifying and monitoring existing inequalities in winter-associated infectious diseases that are potentially amenable to prevention.

For influenza, this report includes all emergency admissions with influenza, identified by the patient’s primary or secondary diagnosis. For COVID-19, this report includes all emergency admissions with a chief complaint corresponding to a respiratory symptom, and an associated positive COVID-19 test.

We report age-standardised admission rates for COVID-19 and influenza, for deprivation levels and ethnic group. For ethnic groups, we use categories based on Office for National Statistics (ONS) classifications, aggregated for consistency with UKHSA surveillance reporting. The white ethnic group is used as the reference group for comparisons as this group represents the largest population in England. The largest group is used because it has the most stable rate and therefore the comparison (rate ratios) will be subject to less fluctuation. For deprivation levels, we use Index of Multiple Deprivation (IMD) quintiles, the official measure of relative deprivation in England produced by the Department for Levelling Up, Housing and Communities (DLUHC).

For both influenza and COVID-19, people living in more deprived areas had higher admission rates across the period. The key findings are:

• influenza admission rates for people living in the most deprived areas were, on average, 2.6 times higher than the least deprived areas
• COVID-19 admission rates for people living in the most deprived areas were, on average, 2.1 times higher than the least deprived areas
• for both influenza and COVID-19, relative inequality between deprivation levels was higher in those aged 45 to 64 years than in children and the oldest age groups

Between ethnic groups, the key findings are:

• for COVID-19, between September 2022 and February 2023 on average, admission rates were similar for all ethnic groups
• for influenza, there were persistent differences between ethnic groups, between September 2022 and February 2023
• the Pakistani ethnic group had the highest admission rates for influenza, which were on average, 2.7 times higher than the white ethnic group across the period
• the black, African, Caribbean, or black British ethnic group had admission rates for influenza, on average, 1.6 times higher than the white ethnic group across the period

These results emphasise the need to improve vaccine coverage across ethnic groups and deprivation levels, for both influenza and COVID-19. There are significant and persistent disparities in vaccine uptake for COVID-19 and influenza by ethnic group and deprivation level. For example, during the winter season 2022 to 2023:

• individuals aged 65 years and over, 84% of the white British ethnic group received an influenza vaccine compared with 55% of the Pakistani ethnic group and 49% of the black Caribbean ethnic group
• influenza vaccine uptake in the least deprived group (IMD10) was 84% compared with 76% in the most deprived group (IMD1)

This analysis also highlights the importance of routine incorporation of variables such as ethnicity and deprivation into our surveillance data. This can help to support the development of policies and actions to reduce these inequalities at national and local level. These may include:

• informing the planning of local health and care systems to prepare for and prevent surges in winter respiratory infections
• informing local and national prioritisation decisions regarding allocation of resource

Limitations

This analysis on hospital admission rates does not control for underlying factors that may drive differences between groups such as differential vaccination uptake, the prevalence of pre-existing health conditions, or other contributing factors. Previous analysis (for instance, ONS analysis on rates of death) has shown that once contributing factors are accounted for, observed differences in health outcomes between ethnic groups and deprivation levels may reduce.

Using the white ethnic group as the reference group will not highlight all inequalities between and within ethnic groups. This report has not looked at the intersection between deprivation and ethnic group explicitly. Within ethnic groups, there will be variation by deprivation level in admission rates.

The analysis presented here should be read alongside the annual UK influenza surveillance reports, the COVID-19 vaccine quarterly surveillance reports, and the weekly national flu and COVID-19 surveillance reports. These reports present data on the observed inequalities in vaccine uptake by ethnic group and deprivation levels. Vaccination is a key preventative intervention for both influenza and COVID-19.

The analysis only presents data at the national level and is not able to make inferences regarding regional or local variation. Further work will be needed to enable analysis at a regional level or by Integrated Care Boards (ICBs).

Background

UKHSA’s mission is to prepare for, prevent and respond to health threats, save lives and protect livelihoods. The threats we protect against range in type, scale and intensity, covering infectious diseases – from novel pathogens with pandemic potential to everyday infections such as colds and flu – and environmental threats including radiation, chemical, nuclear and extreme weather events. To do this, we need to identify and monitor the burden of external health threats in populations most at risk, to inform prioritisation and action.

Across multiple health hazards, there are differences between communities and population groups in terms of their risk of exposure to external health hazards and susceptibility to poor outcomes when exposed. The underlying causes of these inequalities are complex, with influencing factors including the impact of existing social and economic inequalities and pre-existing health conditions. As part of our winter response and surveillance activities, and as part of our aims to achieve more equitable outcomes, UKHSA aims to regularly monitor inequalities in key infectious diseases and related winter health hazards.

More broadly, UKHSA is exploring what can be done to address these inequalities across 4 areas of priority including:

• building scientific and data knowledge and capability
• taking a ‘people and place approach’
• advancing health equity through partnership approaches
• creating a culture where our workforce has the capacity and capability to achieve more equitable health security outcomes

Understanding the groups and communities at highest risk of respiratory hospital admissions enables UKHSA, and wider health partners, to effectively target resource and activities such as vaccination campaigns, health protection response and communications to the groups and settings who will benefit most, contributing to our public sector equalities duties. Regular reporting can help us monitor progress towards reducing these inequalities. 

This report looks at inequalities in emergency hospital admission rates for influenza and COVID-19 in England, between September 2022 and February 2023. Inequalities are considered in the contexts of deprivation level (Index of Multiple Deprivation quintiles) and ethnic group, adjusting for age differences between groups. This report is based on surveillance work from winter 2022 to 2023. This analysis does not control for other underlying factors that may drive differences between groups, such as vaccination, the prevalence of pre-existing health conditions, or other contributing factors. Previous analysis has shown that once these factors are accounted for, observed differences in health outcomes between ethnic groups and deprivation levels may reduce. For instance, ONS analysis on rates of death involving COVID-19 shows that “a large proportion of the difference in the risk of COVID-19 mortality by ethnic group can be explained by demographic, geographical and socioeconomic factors”. The same research has not been carried out for influenza, but other research has evidenced worse outcomes for influenza for some ethnic groups, even after adjusting for deprivation and region.

Differential vaccination uptake is another factor that can drive differences in health outcomes between population groups. For COVID-19, there have been significant disparities in vaccine uptake between ethnic groups since the start of the COVID-19 vaccine programme and these have been persistent through each of the booster campaigns. At the end of the winter season 2022 to 2023 (March 2023), 60% of the white British ethnic group aged 50 years or over had been vaccinated against COVID-19 in the 6 months prior, compared with 19% of the Pakistani ethnic group and 23% of the black Caribbean ethnic group as shown by the national flu and COVID-19 surveillance data report: 30 March 2023 (week 13). Similarly, COVID-19 vaccine uptake is consistently lower in the most deprived groups when compared with the least deprived groups. At the end of winter 2022 to 2023, 64% of those aged 50 years or over from IMD decile 10 (the least deprived) had been vaccinated against COVID-19 in the 6 months prior, as compared with 40% of those aged 50 years or over from IMD decile 1 (the most deprived).

For influenza, across the vaccination cohorts where ethnicity data are collected (at risk groups, pregnant women, individuals aged 65 years and over), there are differences in vaccine uptake between ethnic groups, as presented in the seasonal influenza vaccine uptake in GP patients in England: winter season 2022 to 2023 report. For individuals aged 65 years and over, during the winter season 2022 to 2023, influenza vaccine uptake was highest in the white British ethnic group (84%) and lowest in the black Caribbean ethnic group (49%). Uptake in the Pakistani ethnic group was 55%. For at risk groups, uptake was highest in the white British ethnic group (54%) and lowest in the black Caribbean ethnic group (28%). For pregnant women, uptake was highest in the Chinese ethnic group (46%) and lowest in the black Caribbean ethnic group (14%). These patterns vary by region and age group. Influenza vaccine uptake is also consistently lower in the most deprived groups compared with the least deprived groups.

This report focuses on providing clear, quantifiable, and comparable metrics to describe the inequality in admission rates between groups. Data presented in this report for influenza and COVID-19 are new analyses and provide novel insights into the disproportionate impact that these respiratory conditions had on certain groups for winter 2022 to 2023. This provides a foundation for further investigation, including on the contribution of underlying factors, the reasons for inequalities and the operational delivery of mitigations such as immunisation to reduce the impact of these infections. This analysis also prompts local health care systems, including ICBs and Directors of Public Health, to consider the relevance of this at local level. The analysis presented here should be read alongside the annual UK influenza surveillance reports, the COVID-19 vaccine quarterly surveillance reports, the monthly Influenza vaccine uptake reports and the weekly national flu and COVID-19 surveillance reports. These reports present data on the observed inequalities in vaccine uptake by ethnic group and deprivation levels.

Inequalities in this report are described using two metrics that compare admission rates for different groups: rate ratio and rate difference. The rate ratio captures relative inequality and is useful for comparing inequalities in different outcomes or across time. The rate difference captures absolute inequality and is a more direct assessment of the health impact of inequalities between groups. It is important to consider both inequalities metrics together. This is because when population hospital admission rates are low, even small rate differences between groups can drive large rate ratios. Conversely, when population hospital admission rates are high, even large differences between groups can lead to small rate ratios. We highlight inequality rate ratio differences above a threshold of more than 1.25, following guidance from the Race Disparity Unit (RDU). Admissions for influenza and COVID-19 are identified using different data sources, therefore admission rate differences are not directly comparable.

For additional information including data supplementary to this report, please refer to the accompanying data file.

Inequalities by deprivation level

Influenza admissions by deprivation level

Emergency hospital admissions for influenza were identified using International Statistical Classification of Diseases and Related Health Problems 10th Revision codes (ICD-10) which are assigned to patients on discharge from hospital. This includes all emergency admissions with influenza as a primary or secondary diagnosis; approximately 80% of these admissions were identified using primary diagnosis codes, with 20% identified from secondary diagnosis codes. The deprivation level of an area (Index of Multiple Deprivation quintile) was mapped to each admission using patient home postcode.

Figure 1. Directly age-standardised emergency hospital admission rates for influenza by deprivation level (30-day rolling total), England: September 2022 to February 2023

Across the period, admission rates for influenza increased more quickly, and were higher, for people living in more deprived areas compared with those in less deprived areas (Figure 1). Admission rates were, on average, 2.6 times higher for people living in the most deprived areas, as compared with the least deprived areas (95% confidence interval (CI): 2.5, 2.6). In terms of rate difference, an average of 6 more people per 100,000 from the most deprived areas were admitted to hospital each month (95% CI: 5, 7) than those from least deprived areas.

Emergency hospital admissions for COVID-19 by deprivation level

Admissions with COVID-19 were defined as an admission with a positive COVID-19 test within 14 days before and 1 day after the admission date, and with a chief complaint corresponding to a respiratory symptom (for complaint codes – see definitions). This means that some admissions will be included where COVID-19 was not the primary reason for the admission and instead the infection was identified through routine testing. This is unlikely to be a large proportion as approximately 70% of admissions were assigned a primary or secondary diagnosis of COVID-19, according to clinical classification standards.

Figure 2: Directly age-standardised emergency hospital admission rates for COVID-19 by deprivation (30-day rolling total), England: September 2022 to February 2023

Similar to influenza, people living in more deprived areas had consistently higher admission rates for COVID-19 than people in less deprived areas during winter 2022 to 2023 (Figure 2). Admission rates for people living in the most deprived areas were, on average, 2.1 times higher than for people living in the least deprived areas (95% CI: 2.1, 2.1). In terms of rate difference, there was an average of 4 more admissions per 100,000 per month in the most deprived areas (95% CI: 3, 4) compared with least deprived areas.

Table 1. Directly age-standardised median rate ratios and rate differences between the most (IMD1) and least (IMD5) deprived areas for COVID-19 and influenza, England: September 2022 to February 2023 [note 1]

Metric (95% confidence intervals)	Influenza	COVID-19
Rate ratio	2.6 (2.5, 2.6)	2.1 (2.1, 2.1)
Rate difference	6 (5, 7)	4 (3, 4)

[note 1] These metrics are not adjusted for differential vaccine uptake, the prevalence of pre-existing health conditions or other contributing factors

Differences between deprivation levels, by age group and sex

For all age groups, people were more likely to be admitted to hospital for influenza and COVID-19 if they were living in a more deprived area, across the period. However, the difference in admission rates between the most and least deprived areas varied by age group, with similar trends for influenza and COVID-19.

Figure 3. Crude age-specific emergency hospital admissions for influenza by deprivation level (30-day rolling total), England: September 2022 to February 2023 [note 2]

[note 2] The figure uses different scales for admission rates for each age group to allow the rate ratio for deprivation level to be compared by eye

Relative inequality between areas of the lowest and highest deprivation (measured as rate ratios) for influenza was lowest in children and the oldest age groups, and highest in those aged 45 to 64 years (Figure 3). For instance, for the youngest age group, aged 0 to 4 years, the admission rates were 1.6 (95% CI: 1.5, 1.7) times higher in the most deprived areas for influenza (Figure 3). By comparison, for people aged 45 to 64 years, admission rates were 4.3 (95% CI: 4.1, 4.3) times higher in the most deprived areas for influenza. Relative inequality for COVID-19 had a similar pattern to influenza between age groups; admission rates in 0 to 4 years were 1.4 (95% CI: 1.3, 1.4) times higher in the most deprived areas, whilst in the 45 to 64 age class they were 3.1 (95% CI: 3.0, 3.2) times higher.

In terms of rate difference, a measure of absolute inequality, the greatest gap was for the oldest age groups (over 64 years) for both influenza and COVID-19 admissions. The average difference in the admission rates between the most and least deprived areas for people aged 85 years and over was 16 and 23 admissions per 100,000, per month for influenza and COVID-19, respectively (95% CI: 13, 17; 95% CI: 21, 24).

Table 2. Crude age-specific median rate ratios and rate differences by age group between the most (IMD1) and least (IMD5) deprived areas for COVID-19 and influenza, England: September 2022 to February 2023 [note 1]

Metric (95% confidence intervals)	Age group	Influenza	COVID-19
Rate ratio	0 to 4 years	1.6 (1.5, 1.7)	1.4 (1.3, 1.4)
Rate ratio	5 to 14 years	2.3 (2.2, 2.5)	1.6 (1.4, 1.6)
Rate ratio	15 to 44 years	2.3 (2.2, 2.3)	1.9 (1.4, 1.9)
Rate ratio	45 to 64 years	4.3 (4.1, 4.3)	3.1 (3.0, 3.2)
Rate ratio	65 to 84 years	4.0 (3.8, 4.3)	2.4 (2.4, 2.4)
Rate ratio	85+ years	2.0 (1.9, 2.2)	1.4 (1.3, 1.4)
Rate difference	0 to 4 years	5 (1, 7)	1 (1, 1)
Rate difference	5 to 14 years	3 (3, 3)	0 (0, 0)
Rate difference	15 to 44 years	2 (1, 3)	1 (1, 1)
Rate difference	45 to 64 years	5 (3, 6)	3 (3, 3)
Rate difference	65 to 84 years	13 (8, 16)	13 (12, 13)
Rate difference	85+ years	16 (13, 17)	23 (21, 24)

See note 1 as above.

For the whole cohort, admission rates for males were generally higher than for females for COVID-19, between September 2022 to February 2023. There was no difference in influenza admission rates between the sexes. Relative inequality between areas of deprivation was similar for both males and females across influenza and COVID-19. For instance, influenza admission rates for females living in the most deprived areas were 2.7 times higher than for those in the least deprived areas (95% CI: 2.5, 2.7), compared with 2.4 times higher for males (95% CI: 2.4, 2.5). The rate difference was also similar between sexes, with 7 more females admitted for influenza in the most deprived areas compared with the least deprived areas, and 5 more males, per 100,000, per month (95% CI females: 5, 8; 95% CI males: 4, 6).

Table 3. Directly age-standardised median rate ratios and rate difference by sex between the most (IMD1) and least (IMD5) deprived areas for COVID-19 and influenza, England: September 2022 to February 2023 [note 1]

Metric (95% confidence intervals)	Sex	Influenza	COVID-19
Rate ratio	Female	2.7 (2.5, 2.7)	2.2 (2.2, 2.3)
Rate ratio	Male	2.4 (2.4, 2.5)	2.0 (2.0, 2.0)
Rate difference	Female	7 (5, 8)	4 (3, 4)
Rate difference	Male	5 (4, 6)	3 (3, 4)

See note 1 as above.

Inequalities between ethnic groups

This analysis describes differences in emergency admission rates for COVID-19 and influenza between ethnic groups, adjusted throughout for age differences between groups and, when noted, for differences in deprivation levels between groups. It does not control for other underlying factors that may explain differences between ethnic groups, such as vaccination, the prevalence of pre-existing health conditions, or other factors. For instance, it does not adjust for any geographic concentration of COVID-19 and influenza transmission. Further analysis is required to assess how the characteristics of areas of high disease prevalence could contribute to ethnic group differences, and how differences in pre-existing health conditions or vaccination uptake affect inequalities.

Admission rates are presented by ethnic group. The categories are:

• black, African, Caribbean, or black British
• Indian (Asian or Asian British)
• mixed or multiple ethnic groups
• other Asian or Asian British
• other ethnic groups
• Pakistani (Asian or Asian British)
• white

Other published analyses disaggregate between groups further, for instance presenting rates for the Bangladeshi ethnic group separate from the other Asian or Asian British ethnic group. The categories presented here were chosen for consistency and comparability with the UKHSA weekly national influenza and COVID-19 report (see Figure 9). The white group was used as the reference group for comparisons as this group represents the largest population in England. The largest group is used because it has the most stable rate and therefore the comparison (rate ratios) will be subject to less fluctuation. Data available in the supplementary data file allow additional comparisons between groups.

Emergency hospital admissions for influenza by ethnic group

Figure 4. Directly age-standardised emergency hospital admissions for influenza, by ethnic group (30-day rolling total), England: September 2022 to February 2023

Across the period, the Pakistani ethnic group had admission rates substantially higher than any other ethnic group and the white ethnic group had a lower average admission rate than any other ethnic group (Figure 4). Admission rates were on average, 2.7 times higher for the Pakistani ethnic group compared with the white ethnic group (95% CI: 2.4, 2.9). In terms of rate difference, the Pakistani ethnic group had an average of 8 more emergency admissions per 100,000, per month than the white ethnic group (95% CI: 6, 11).

The black ethnic group had admission rates that were, on average, 1.6 times higher than the white ethnic group (95% CI: 1.5, 1.7). In terms of rate difference, the black ethnic group had 1 more admission per 100,000 per month (95% CI: 0, 1) compared with the white ethnic group.

In this report, admission rates by ethnic group are age-standardised. To examine whether deprivation accounts for some of the differences between ethnic group, we also adjusted for both age and deprivation level – these rate ratios and rate differences based on age- and deprivation-standardised rates are included in the supporting data file. After adjusting for both age and deprivation level, the average rate ratio and rate difference for the Pakistani ethnic group is 2.5 (95% CI: 2.4, 2.7) and 7 (95% CI: 6, 9), per 100,000 per month, respectively, compared with the white ethnic group. After adjusting for both age and deprivation level, the black ethnic group had admission rates that were, on average, 1.3 times higher than the white ethnic group (95% CI: 1.3, 1.4). However, the rate difference is less than 1 per 100,000, per month (95% CI: 0, 0).  For both the black and Pakistani ethnic group, this is a smaller difference than when only adjusting for age, but the difference remains disproportionate as they are above the threshold suggested by RDU guidance. All the other ethnic groups had similar admission rates to the white ethnic group when adjusting for either age or both age and deprivation level.  

It is important to note that it is also likely that there are differences between deprivation levels within ethnic groups. For instance, people from the most deprived areas within the white ethnic group are likely to have experienced high admission rates, even though the white ethnic group on average had lower admission rates.

Emergency hospital admissions for COVID-19 by ethnic group

Figure 5. Directly age-standardised emergency hospital admissions for COVID-19 by ethnic group (30-day rolling total), England: September 2022 to February 2023

For COVID-19, differences in admission rates among ethnic groups were less consistent during winter 2022 to 2023, when compared with influenza (Figure 5). On average, the rate ratio was highest for the Pakistani ethnic group, indicating higher admission rates. However, the rate ratio was below 1.25, with admission rates being, on average, 1.1 times higher in the Pakistani ethnic group compared with the white ethnic group (95% CI: 1.1, 1.2). Rate differences were also small; the Pakistani ethnic group had an average of 1 more emergency admission per 100,000, per month, than the white ethnic group (95% CI: 0, 1). This contrasts with earlier stages of the COVID-19 pandemic where there were much greater differences in COVID-19 admission rates among ethnic groups (see COVID-19 Health Inequalities Monitoring for England tool).

When compared with the group with the lowest average admission rates over the period (the black ethnic group), admission rates were on average 1.5 times higher for Pakistani ethnic group (95% CI: 1.4, 1.6) and 1.2 times higher for the white ethnic group (95% CI: 1.2, 1.3).

After adjusting for both age and deprivation level, the rate ratio and rate difference for the Pakistani ethnic group were reduced to 1.0 (95% CI: 0.9, 1.0) and less than 1 per 100,000 per month (95% CI: 0, 0) compared with the white ethnic group, demonstrating the intersectional inequalities in this ethnic group. All other ethnic groups also had rate ratios below 1.25, when compared with the white ethnic group. As for influenza, it is also likely that there are notable differences within ethnic groups between deprivation levels. Full results can be found in the supplementary data file.

Table 4. Directly age-standardised median rate ratios and rate differences by ethnic group relative to the white ethnic group, England: September 2022 to February 2023 [note 1]

Metric (95% confidence intervals)	Ethnic group	Influenza	COVID-19
Rate ratio	black/African/Caribbean/black British	1.6 (1.5, 1.7)	0.8 (0.8, 0.8)
Rate ratio	Indian (Asian or Asian British)	1.3 (1.2, 1.4)	0.9 (0.9, 0.9)
Rate ratio	mixed/multiple ethnic groups	1.3 (1.2, 1.4)	0.8 (0.7, 0.8)
Rate ratio	other Asian/Asian British	1.2 (1.2, 1.3)	0.9 (0.9, 1.0)
Rate ratio	other ethnic groups	1.2 (1.2, 1.3)	0.9 (0.9, 1.0)
Rate ratio	Pakistani (Asian or Asian British)	2.7 (2.4, 2.9)	1.1 (1.1, 1.2)
Rate difference	black/African/Caribbean/black British	1 (0, 1)	-1 (-1, -1)
Rate difference	Indian (Asian or Asian British)	1 (0, 1)	-1 (-1, 0)
Rate difference	mixed/multiple ethnic groups	1 (1, 1)	-1 (-2, -1)
Rate difference	other Asian/Asian British	1 (1, 1)	0 (-1, 0)
Rate difference	other ethnic groups	1 (0, 1)	0 (0, 0)
Rate difference	Pakistani (Asian or Asian British)	8 (6, 11)	1 (0, 1)

See note 1 as above.

Table 5. Directly age and IMD-standardised median rate ratios and rate differences by ethnic group relative to the white ethnic group, England: September 2022 to February 2023 [note 1]

Metric (95% confidence intervals)	Ethnic group	Influenza	COVID-19
Rate ratio	black/African/Caribbean/black British	1.3 (1.3, 1.4)	0.8 (0.8, 0.8)
Rate ratio	Indian (Asian or Asian British)	1.2 (1.2, 1.3)	0.8 (0.8, 0.8)
Rate ratio	mixed/multiple ethnic groups	1.2 (1.1, 1.2)	0.7 (0.7, 0.8)
Rate ratio	other Asian/Asian British	1.2 (1.1, 1.2)	0.8 (0.8, 0.9)
Rate ratio	other ethnic groups	1.1 (1.1, 1.2)	0.9 (0.9, 1.0)
Rate ratio	Pakistani (Asian or Asian British)	2.5 (2.4, 2.7)	1.0 (0.9, 1.0)
Rate difference	black/African/Caribbean/black British	0 (0, 0)	-1 (-1, -1)
Rate difference	Indian (Asian or Asian British)	0 (0, 0)	-1 (-1, -1)
Rate difference	mixed/multiple ethnic groups	0 (0, 1)	-1 (-2, -1)
Rate difference	other Asian/Asian British	0 (0, 0)	-1 (-1, -1)
Rate difference	other ethnic groups	0 (0, 1)	-1 (-1, 0)
Rate difference	Pakistani (Asian or Asian British)	7 (6, 9)	0 (0, 0)

See note 1 as above.

Data sources

Admitted Patient Care (APC)

Hospital Episode Statistics Admitted Patient Care

This data set records all admissions to secondary care, including patient demographics and clinical information and is maintained by NHS England. Clinical information is recorded according to international classification standards (ICD-10 codes). These codes are assigned after discharge and include detailed disease diagnoses. These analyses are based on hospitalisation data derived from the Secondary Uses Service (SUS) data feed which has an 8 to 12 week lag in inclusion of full details of admissions. Future updates to the data source will result in low-impact changes to the admission rates stated for February in this report.

Emergency Care Data Set (ECDS)

Emergency Care Data Set (ECDS)

This data set records Accident and Emergency activity including patient demographics and clinical information and is maintained by NHS England. Clinical information is recorded as a chief complaint code. These codes are symptom based and assigned by a clinician when a patient presents at hospital and before investigations have taken place. There is a 3 week lag in inclusion of full details of admissions. As chief complaints are assigned upon presentation to hospital, rather than at discharge, clinical information in ECDS has a shorter lag than the APC data set.

National Immunisation Management System (NIMS)

The National Immunisation Management System (NIMS) was developed by a range of health and digital government agencies as the system of record for the NHS COVID-19 vaccination programme in England. Vaccinations administered are entered on an application which is verified by individual National Health Service number in a centralised system. They are linked to demographic data obtained from the National Health Service (NHS) (for example, gender, date of birth), using the individual’s unique NHS number. UKHSA receive a feed of this data to use for monitoring vaccine coverage, effectiveness and safety.

Second-Generation Surveillance System (SGSS)

This is the national laboratory reporting system, recording all notified positive laboratory reports of SARS-CoV-2. It is maintained by NHS England.

Methodology and definitions

Influenza admissions

Emergency hospital admissions for influenza were identified using the APC data set. They were defined as an emergency admission with a primary or secondary ICD-10 diagnosis code for influenza; approximately 80% of admissions were identified from primary diagnosis codes, with 20% identified using secondary diagnosis codes. The APC data set was used instead of ECDS because testing data for influenza cases was not available in ECDS at the time.

ICD-10 codes used are listed below:

• influenza due to identified zoonotic or pandemic influenza virus (ICD-10 code: J09)
• influenza due to identified seasonal influenza virus (ICD-10 code: J10)
• influenza with pneumonia, seasonal influenza virus identified (ICD-10 code: J100)
• influenza with other respiratory manifestations, seasonal influenza virus identified (ICD-10 code: J101)
• influenza with other manifestations, seasonal influenza virus identified (ICD-10 code: J108)
• influenza, virus not identified (ICD-10 code: J11)
• influenza with pneumonia, virus not identified (ICD-10 code: J110)
• influenza with other respiratory manifestations, virus not identified (ICD-10 code: J111)
• influenza with other manifestations, virus not identified (ICD-10 code: J118)

Admissions with COVID-19

Emergency hospital admissions with COVID-19 were identified using the ECDS data set. They were defined as an admission with a chief complaint corresponding to one of the respiratory symptoms listed below and a positive COVID-19 test within 14 days before and 1 day after the admission date (linked via SGSS). This means that some admissions will be included where COVID-19 was not the primary reason for the admission and instead the infection was identified through routine testing. This is unlikely to be a large proportion as approximately 70% of admissions were assigned a primary or secondary diagnosis of COVID-19, according to clinical classification standards.

They may also exclude admissions which were due to COVID-19 but were without a positive COVID-19 test. Therefore, they may not reflect the true burden of COVID-19 on hospitals, but do provide a consistent measure on which to base inequalities measures.

Admission rates

Admission rates were calculated as directly age-standardised emergency hospital admission rates per 100,000 population for influenza and COVID-19.

Directly age-standardised rates were used to control for differences in age structure between population groups. These were calculated for each day as the sum of hospital admissions over the preceding 30 days (rolling 30-day total admissions), divided by the population, standardised to the relevant standard population (see below), and multiplied by 100,000.

Admission rates for each ethnic group were also calculated as age and IMD-standardised emergency hospital admission rates per 100,000 population for influenza and COVID-19. Directly age and IMD-standardised rates were used to control for differences in age and deprivation structure between the ethnic groups. This methodology followed that of the directly age-standardised rates but separated the admissions and population denominators into age and deprivation level stratified groups.

For breakdowns by deprivation level, population denominators were taken from ONS mid-year population estimates (2019). These population estimates were also used to generate a standard population for calculating the directly age-standardised admission rates.

For breakdowns by ethnic group, population denominators were calculated using the NIMS data set, using patient records as at 28 February 2023. NIMS population denominators were used for breakdowns by ethnicity because the ONS 2019 mid-year population estimates for England were not available by both ethnic group and age. These population estimates were also used to generate a standard population for calculating the directly age-standardised admission rates.

For breakdowns by age group, crude age-specific rates are presented (rolling 30-day total admissions divided by the population, multiplied by 100,000).

Deprivation level

The Index of Multiple Deprivation (IMD) is the official measure of relative deprivation in England produced by DLUHC. It describes the relative level of deprivation in a small area (lower layer super output area (LSOA)), but not necessarily the individual people living in the area. Many non-deprived people live in deprived areas and many deprived people live in non-deprived areas. Areas are ordered by IMD score and grouped into 5 categories (IMD quintiles). IMD 1 refers to the most deprived areas and IMD 5 refers to the least deprived areas.

Deprivation level was assigned to each hospital admission based on the area where the patient lived, using their home postcode at the date of admission.

Ethnic group

Ethnic group categories were based on ONS classifications, aggregated for consistency with other UKHSA surveillance reporting. Data on ethnicity for emergency hospital admissions were preferentially assigned from NIMS (linking via NHS number), then from ECDS or APC (if missing from NIMS), then from SGSS (for COVID-19 admissions, if missing from NIMS and ECDS or APC). Ethnicity recorded in administrative data for hospital admissions can be of poor quality for some ethnic groups; whereas ethnicity from NIMS is derived from both secondary and GP records and is more complete than ethnicity data available from secondary care only (SUS). Approximately 1% of admissions data was excluded from analysis due to missing ethnicity data.

​Rate ratios

Rate ratios by deprivation level were calculated by dividing the directly age-standardised admission rate for the most deprived areas (IMD 1) by the directly age-standardised admission rate for least deprived areas (IMD 5)​.

Rate ratios by ethnic group were calculated by dividing the directly age-standardised admission rate for each ethnic group by the directly age-standardised admission rate for the white ethnic group. The white group was chosen as the reference group as they represent the largest relative population in England.

The interpretation of rate ratios depends on their size and their statistical significance, with guidance on this published by the Race Disparity Unit (RDU). In this report, we have used the following thresholds:

Rate ratio is less than 0.8 or more than 1.25

A rate ratio of less than 0.8 or more than 1.25 would be highlighted as disproportionate. The RDU guidance recommends that rate ratios of this size that are also statistically significant are deemed high priority for policy action. If rate ratios are greater than this threshold, but not statistically significant, they merit ongoing monitoring. If rate ratios are greater than 0.8 and below 1.25, they are deemed below the threshold for disproportionality.

Comparing rate ratios

When comparing rate ratios between groups or conditions, they are highlighted as different to one another if the difference between rate ratios is greater than 0.2.

Confidence intervals

95% confidence intervals were used as measure of certainty around the estimates of median rate ratio to determine whether they were statistically different from parity (a rate ratio equal to 1). For example, the deprivation rate ratio for influenza was 2.6 (95% CI: 2.5, 2.6). These confidence intervals do not include 1 so the rate ratio is significantly different from parity.

As noted by the RDU guidance, while significance and size of the rate ratio are important pieces of information, analysts and policymakers may still want to consider more granular breakdowns of different groups to reveal hidden inequalities. For instance, there may be intersectional or cumulative impacts of disparities in outcomes, or some population groups may be persistently or multiply disadvantaged.

Confidence intervals (95% CI) for the rate ratios and rate differences (described below) presented are provided in the results and the accompanying data file. Average rate ratios and rate differences were calculated as the median value for the period (September 2022 to February 2023). Bootstrapped 95% confidence intervals for these estimates were generated using 10,000 bootstrap replicates with R Package DescTools (version 0.99.48) and are provided in the results and the accompanying data file.

Rate differences

Rate differences by deprivation level were calculated by subtracting the directly age-standardised admission rate for least deprived areas (IMD 5) from the directly age-standardised admission rate for most deprived (IMD 1).

Rate differences by ethnic group were calculated by subtracting the directly age-standardised admission rate for white ethnic group from the directly age-standardised admission rate for each other ethnic group.

Median rate ratios and median rate differences are calculated separately for the period.

Further information

Related statistics

National Flu and COVID-19 surveillance reports 2022 to 2023 season

Annual UK influenza surveillance reports

COVID-19 vaccine quarterly surveillance reports

Monthly influenza vaccine uptake reports

 Acknowledgements

We are grateful to NHS England for provision of the data used in this report and for feedback from the UKHSA COVID-19 and winter Incident Management Teams, as well as colleagues in the Office for Health Improvement and Disparities, the Office for National Statistics and NHS England.

This report was produced by the All Hazards Intelligence Directorate, UKHSA.

Please direct any queries or comments to [email protected]

Authors and contributors

Catherine Falconer – Deputy Director (Health Equity and Inclusion Health)

Flora Stevens – Senior Epidemiologist

James Guilder – Senior Health Analyst

Lucy Browning – Senior Health Analyst

Megan Scott – Senior Health Analyst

Kerry Cella – Public Health Analysis Lead

Sarah Deeny – Deputy Director (National Hazards Assessment and Analysis)

UKHSA All Hazards Intelligence Directorate

UKHSA Health Equity and Inclusion Health Division

UKHSA Immunisation and Vaccine-Preventable Diseases Division

UKHSA COVID-19 Vaccines and Epidemiology Division