Unfortunately, your browser is too old to work on this website. Please upgrade your browser
Analysis

Geographic inequalities in premature mortality
Exploring the role of socioeconomic factors and ethnicity

Published 8 July 2025
Time to read clock icon About 13 mins
Authors
Topics

Key points

  • There are significant differences in how likely people are to die early (before the age of 75) between different parts of England and Wales. After accounting for age and sex, people are almost twice as likely to die early in the 10% of local authority areas with the highest rates of premature mortality as those in the 10% of local authority areas with the lowest rates.
  • Premature mortality rates are highest in local authorities in the North East, the North West and Yorkshire and the Humber, and lowest in London.
  • Our research explores why premature mortality is more common in some areas than others. We look at the role of socioeconomic factors as well as ethnicity and being born outside the UK. We also look at how much inequality is left unexplained, pointing to other potential factors at play.
  • Our analysis finds:
    • Socioeconomic factors, ethnicity and being born outside the UK account for 61% of the variation in premature mortality rates. On their own, ethnicity and being born outside the UK account for 7% of the variation.
    • This means socioeconomic factors (highest qualification achieved, occupation and area-level deprivation) play a major role in geographic inequalities in premature mortality rates.
    • The remainder – almost 40% – is unexplained by the factors we looked at, but could result from differences in environmental exposures, other socioeconomic factors and building blocks of health (such as social connection) and access to quality health and care services.
  • The analysis also looks at four of the leading causes of death – cancer, cardiovascular disease, diabetes and respiratory conditions. The gap in premature mortality rates is widest for deaths associated with diabetes and deaths due to respiratory conditions.
  • Addressing the scale of health inequalities will require long-term action for economic recovery in areas with longstanding industrial decline; a strong focus on prevention; investment in public health services; and action to address risk factors, such as smoking, alcohol and poor diet. These actions should be brought together in a clear strategy for tackling health inequalities.
  • Our analysis also shows that the impact of accounting for ethnicity, being born outside the UK and socioeconomic factors on premature mortality rates differs across local authorities. For example, in Blackpool, these factors almost entirely explain its higher premature mortality rate. This points to the urgent need for action to address the socioeconomic factors behind this. However, in Wrexham, these factors do not entirely explain higher premature mortality rates, highlighting the need for further understanding of other drivers of these disparities.
 

Introduction

There is a longstanding and well-documented gap in health outcomes between different local areas in England and Wales. For example, in 2021 to 2023, male life expectancy in Blackpool – the local area with the lowest male life expectancy – was 10.3 years lower than in Hart – the local area with the highest. Female life expectancy in Blackpool – the local area with the lowest female life expectancy – was 7.6 years lower than in Kensington and Chelsea – the local area with the highest. 

Differences in premature mortality rates contribute to geographic inequalities in life expectancy and are a helpful indicator of working-age health. Premature mortality rates capture someone’s chances of dying earlier than expected, defined as before the age of 75 years old. They are a more nuanced indicator of inequality compared with all-age mortality rates, as premature mortality rates highlight disparities in preventable ill health and health outcomes that lead to early deaths. 

Given the wide and persistent geographic inequalities across England and Wales, it is important to better understand what explains them. Office for National Statistics data provide a unique opportunity to explore mortality patterns at a local authority level by linking census data to mortality records.

Existing evidence suggests mortality rates vary depending on ethnicity, being born outside the UK, socioeconomic status and an area’s deprivation level. This research builds on this existing evidence by focusing on premature mortality rates and assessing how their variation across local authorities is influenced by an area's deprivation level and population make-up.

To highlight where the chances of dying prematurely are highest, we estimated differences in premature mortality rates across upper tier local authorities in England and unitary authorities Wales. We then estimated – through a series of statistical models – how much these differences can be explained by the population make-ups of these areas (in terms of ethnic composition, proportion of people born outside the UK, education and occupation) and the areas’ level of deprivation. We considered both all-cause premature mortality and cause-specific premature mortality. 

The statistical models used assume the population of every local authority has the same age and sex profile. We estimated how much the relative gap in premature mortality rates between the local authorities with the 10% highest and lowest rates would change if there were no differences in the ethnic compositions or migration backgrounds of their populations, nor in individual-level socioeconomic characteristics and area-level deprivation. Lastly, we estimated the relative gap in premature mortality rates that remains unexplained by differences in population and area characteristics, which is likely due to other factors.  

Data and approach

The data underpinning the analytical work presented in this analysis was produced by the Office for National Statistics on behalf of the Health Foundation.

This research uses data from the 2021 census to look at differences in premature mortality rates (that is, deaths occurring below the age of 75 years old) across 172 local authorities – upper tier local authorities in England and unitary local authorities in Wales (referred to throughout as local authorities). We used 2022 local authority boundaries, combining the City of London and Hackney, and Cornwall and Isles of Scilly, due to small sample sizes. The analysis includes people living in England and Wales who are aged 18 to 74 years and included in the 2021 census. Our study period included deaths from March 2021 to December 2023. Any adult who died from any cause before their 75th birthday within this time period was classified as a premature death. Premature mortality is also examined for selected causes (cardiovascular disease, cancer, respiratory conditions and diabetes).

Premature mortality rates were estimated as the number of people per 100,000 who die prematurely in one year. Premature mortality rates at the local authority level were estimated using three statistical models to adjust for different population and area characteristics across local authorities: 

  1. A baseline model adjusting for differences in age and sex
  2. A second model adjusting for age, sex, ethnicity and migration background (using the 10-group ethnic classification and migration background, where a migrant is defined as someone born outside the UK)
  3. A third model adjusting for age, sex, ethnicity, migration background and socioeconomic factors (area-level deprivation based on the Index of Multiple Deprivation, occupation based on National Statistics Socio-economic classification and highest qualification achieved) and the interaction between ethnicity and socioeconomic factors.

Geographic inequality in premature mortality rates is defined as the relative difference between average premature mortality rates in the 10% of local authorities with the highest and lowest rates. These three models enabled us to estimate what happens to geographic inequality in premature mortality rates when only taking into account differences in population ethnicity and migration background compared with when taking into account differences in population ethnicity, migration background and socioeconomic factors (occupation, education level and area-level deprivation) together. 

This approach helps us explore the reasons premature mortality is more common in some areas than others. People from different ethnic and migration backgrounds have different risks of premature death, and they are also more likely to live in certain areas. This means that some geographic inequality will be explained by the population make-up of an area in terms of ethnicity and migration background. 

Socioeconomic factors help explain how and why the risk of dying early varies between different areas. We consider these factors alongside ethnicity and migration background because some of the geographic inequality in premature mortality rates reflects the combined effects of socioeconomic disadvantage and the ethnic and migration backgrounds of the population. If we only adjusted for socioeconomic factors without considering ethnicity and migration background, we might misattribute some of the variation in mortality to socio-demographic factors alone — when in fact it may also reflect other differences across ethnic groups that affect health, such as exposure to discrimination, differences in health behaviours or access to services.

Full details on the data and methods used can be found in the Office for National Statistics publication. 

Limitations

Although the data used to conduct this study included all people living in England and Wales who were counted in the 2021 census, some groups are still more likely to be underrepresented, such as those experiencing homelessness. People with incomplete records were also excluded from this analysis, and some deaths may have been missed due to delays in death registrations.

Our estimates might have been affected if we took into account other factors known shape health outcomes. Some – including health behaviours such as smoking or alcohol consumption – were harder to include because of data availability, while others – such as access to and provision of health care services and environmental factors – can be considered in further work. These are mentioned in the conclusions.

We did not examine COVID-19 as a cause of premature mortality during the study period (2021 to 2023), nor did we compare rates of premature mortality between pre-pandemic and pandemic years. Therefore, our study findings cannot help draw conclusions about the direct and indirect impacts of COVID-19 on geographic variation in premature mortality rates.

 

What is the extent of geographic inequalities in premature mortality from all causes of death?

There are considerable inequalities in premature mortality rates between local authorities in England and Wales. Because people of different ages and sexes have different chances of dying prematurely, some differences in premature mortality rates across areas might just reflect their population make-up (for example, a slightly older population). To ensure rates were comparable, we looked at differences in premature mortality rates after taking into account population make-up in terms of age and sex.

After accounting for population differences in age and sex, premature mortality rates varied from 324 people per 100,000 in the 10% of local authorities with the lowest premature mortality rate to almost double – at 596 people per 100,000 – in the 10% of local authorities with the highest premature mortality rate. 

Figure 1 includes premature mortality rates for all local authorities in England and Wales, grouped by region. In 2021 to 2023, chances of dying prematurely were highest in local authorities in the North East, North West and Yorkshire and the Humber and lowest in London. In the North East, all but one local authority had premature mortality rates above the overall premature mortality rate for England and Wales (Figure 1). On the other hand, in London, only one local authority had a rate above the England and Wales rate. 

Figure 1

 

To what extent do differences in ethnicity, migration background and socioeconomic factors explain geographic inequalities in premature mortality from all causes of death?

Ethnicity and migration background

Demographic characteristics are often linked to someone’s chances of dying prematurely. Because of this, differences in premature mortality rates across areas might simply reflect the demographic characteristics of their populations. To ensure rates are comparable across local authorities, it is important to take into account differences in population make-up.

In addition to age and sex, another demographic characteristic relevant in this context is ethnicity. This is for two reasons. First, evidence shows that mortality rates differ depending on ethnicity, with – as of 2012 to 2014 – the white British population being the ethnic group with the highest overall mortality. Second, different areas are likely to have smaller or larger populations of people from certain ethnic backgrounds. For example, 85% of the population in Newham belongs to an ethnic minority compared with 3% of the population in Redcar and Cleveland. Additionally, while the same proportion of the populations of Tower Hamlets and Redbridge belong to an ethnic minority (77%), their compositions are very different. 35% of Tower Hamlets’ population is Bangladeshi, compared with only 10% of Redbridge’s. 

Being born outside the UK is also observed to be associated with differences in mortality rates. People who move to the UK tend to have lower mortality rates than those born in England and Wales who belong to the same ethnic group. This is often referred to as the ‘healthy migrant effect’. For example, the mortality rate for UK-born Bangladeshi females is 1.9 times higher than that of Bangladeshi females born outside the UK and 1.4 times higher than that of white British females – although the premature mortality rates for people of other UK-born minority ethnic groups still remains lower than that of white British people. Similar to ethnicity, people born outside the UK are more likely to live in specific areas. 

For all these reasons, it is important to look at how much the ethnic and migration composition of different areas contributes to variations in premature mortality rates. Our analysis shows that the relative gap in premature mortality rates between the local authorities with the 10% highest and lowest rates would be 7% lower if all local authorities had the same population make-up in terms of age, sex, ethnicity and migration background compared with only accounting for population differences in age and sex.

Ethnicity, migration background and socioeconomic factors (level of education, occupation, area-level deprivation)

Existing evidence shows that mortality is higher for people working in low-paid occupations, with lower qualifications or living in more deprived areas. As such, these factors likely play a large role in variations in geographic inequality in premature mortality. 

Our analysis shows that the relative gap in premature mortality rates between the 10% of local authorities with the highest and lowest rates would be 61% lower if all local authorities had the same population make-up in terms of ethnicity, migration background, individual-level socioeconomic status and area-level deprivation compared with only accounting for population differences in age and sex. This indicates that socioeconomic factors play a major role in explaining inequality in premature mortality.

Unexplained variation

39% of the relative gap remains unexplained. This indicates that other factors are associated with differences in premature mortality rates between areas. These might include access to and the provision of health care services as well as environmental factors such as air quality. As many of these factors often vary depending on someone’s income, employment status and the level of deprivation in the area where they live, their impact might be partially captured by the socioeconomic factors we included. 

 

To what extent can ethnicity, migration background and socioeconomic factors explain differences between individual local authorities?

As explained, inequality between the 10% of local authorities with the highest and lowest premature mortality rates reduces by 61% when all factors – ethnicity, migration background, individual-level socioeconomic status, area-level deprivation, age and sex – are considered. However, the role of those factors changes when looking at different local areas. 

Figure 2 illustrates this by comparing four selected local authorities (Blackpool, Knowsley, Southwark, and Wrexham) to Richmond upon Thames – chosen as a reference point because it is the local authority with the lowest premature mortality rate. The comparison was carried out by assessing how much premature mortality rates in these four selected local authorities would change if we assumed there were no differences in ethnicity, migration backgrounds and socioeconomic factors between the selected local authorities and Richmond upon Thames.

Figure 2

For all the selected local authorities, adjusting for these factors brings their premature mortality rate closer to that of Richmond upon Thames. However, the effect differs depending on the local authority considered:

  • Blackpool is the local authority with the highest premature mortality rate. Once we accounted for the factors considered in this analysis, we found that they explain most of the difference in Blackpool’s premature mortality relative to Richmond upon Thames. This suggests that addressing Blackpool’s disadvantage in terms of socioeconomic factors could go a long way towards improving its health outcomes.
  • Knowsley and Southwark are two local authorities whose premature mortality rates end up being better than those of Richmond upon Thames after adjusting for the factors considered in this analysis. This suggests that there might be specific assets in these areas that would allow them to have lower premature mortality rates than Richmond upon Thames if they had the same demographic and socioeconomic characteristics accounted for in this analysis.
  • Wrexham is the local authority with the highest premature mortality rate after accounting for these factors, meaning they explain only some of the difference in its premature mortality rate relative to Richmond upon Thames. This suggests that to improve health outcomes in line with Richmond upon Thames, action should be taken to address both socioeconomic factors and the additional drivers of higher premature mortality which we have not accounted for in this analysis. 

The map in Figure 3 highlights how these different effects play out across England and Wales. 77 of the 176 local authorities analysed still have higher premature mortality rates than Richmond upon Thames after adjusting for ethnicity, migration background and socioeconomic factors (although only some significantly higher). It is notable that this includes all the local authorities in Wales. On the other hand, 80 of the 176 local authorities have lower premature mortality rates than Richmond upon Thames once adjusted for ethnicity, migration background and socioeconomic factors (although only some significantly lower), pointing to the existence of other possible assets in these areas. More than half of the local authorities in London have lower premature mortality rates than Richmond upon Thames after adjustment.

Figure 3

 

What do inequalities in premature mortality look like for specific causes of death – and what is the role of ethnicity, migration background and socioeconomic factors?

We now look at the extent of geographic inequalities in premature mortality rates for selected causes of death and inspect whether differences between the local authorities with the 10% highest and lowest premature mortality rates for specific causes of death can be explained by the ethnic and migration make-up of their populations, individual-level socioeconomic factors and area-level deprivation in the same way as all-cause premature mortality rates. We focus on some of the leading causes of death (respiratory conditions, cardiovascular diseases, cancers and diabetes). Over the period 2021 to 2023, 25% of deaths were due to cancer, 24% due to cardiovascular disease, 11% due to respiratory conditions and 12% involved diabetes. Some causes of death (for example, deaths involving alcohol-specific causes, drug-related causes and suicide) were excluded from the analysis because the sample size was not large enough to allow for a robust estimate. 

As shown in Figure 4, for most causes of death, regional inequalities follow a similar pattern to all-cause mortality, with local authorities in the North East and North West having higher premature mortality rates and local authorities in London having lower premature mortality rates. The one exception is deaths involving diabetes, where local authorities in the North East and North West have average premature mortality rates and rates are highest in the East Midlands and West Midlands. This suggests different factors might play a role in premature mortality rates for deaths that involve diabetes compared with the other causes of death. 

Figure 4

The extent of geographic inequality varies by cause of death. As shown in Figure 5, inequality is highest for diabetes and deaths due to respiratory conditions, where the premature mortality rate in the 10% of local authorities with the highest rates is respectively 228% and 175% higher than in the 10% of local authorities with the lowest rates. Figure 6 also shows that for all causes of death, the relative difference in premature mortality rates between the 10% of local authorities with the highest and lowest rates reduces by a large proportion when differences in ethnicity, migration background and socioeconomic factors are accounted for. The effect of adjustment differs across causes of death, ranging from a 43% reduction in variation for cardiovascular diseases to a 65% reduction for respiratory conditions. This suggests that for some causes of death, factors other than the ones considered in this analysis might be more important in explaining geographic variation in premature mortality rates. 

Figure 5

 

Conclusions

As is the case for other health outcomes, there are large differences in premature mortality rates across local authorities in England and Wales. People living in the 10% of local authorities with the highest premature mortality rates have almost twice the chance of dying before 75 years of age than those living in the 10% of local authorities with the lowest rates. Many of these deaths are avoidable.

Using restricted data from the 2021 census, this research shows that the relative difference in premature mortality between the 10% of local authorities with the highest premature mortality rates and the 10% with the lowest reduces by 7% once differences in population ethnicity and migration have been accounted for. This difference in premature mortality rates reduces by 61% once also socioeconomic factors are also accounted for. 

This research has also shown that the role of ethnicity, migration background and socioeconomic factors in these differences depends on the specific local authority considered. This highlights areas where addressing socioeconomic disadvantage is likely to close the gap in premature mortality, areas where this should be combined with other interventions and areas that appear to have assets that allow them to do better than expected in terms of premature mortality.

Geographic health inequalities in the UK are wide and longstanding, and reducing them remains a significant challenge. This research shows the significant role socioeconomic factors – the building blocks of health – play in differences in premature mortality across local authorities. 

The government’s health mission aims to reduce health inequalities in local areas. This cannot be achieved by any single department or policy alone. Action is needed on the building blocks of health – including ensuring equitable access to education and good quality jobs across England and Wales – to mitigate socioeconomic disparities. As this analysis also highlights, local challenges differ significantly. It is important that solutions are developed or adapted to suit local population needs.  

39% of geographic inequalities in premature mortality remain unexplained by the population and area characteristics considered in this analysis. This means there are other factors that could account for the remaining variation, including differences in exposure to environmental hazards, other socioeconomic characteristics not accounted for in this analysis (for example, social connection within local communities) and access to and provision of health care services. Historical factors, such as people’s socioeconomic position throughout their lifetime, as well as internal migration between local areas might also contribute to the variation. Further work could be done to understand the extent to which these different factors account for geographic inequalities in premature mortality.

The data underpinning this analytical work was produced by the ONS Health Research Group on behalf of the Health Foundation. All interpretation of those data reported here is the work of the Health Foundation. 

Many thanks to the External Advisory group who helped shape and interpret this analysis. A special thanks to Greg Fell, President of the Association of Directors of Public Health, and Veena Raleigh, Senior Policy Fellow at The King’s Fund, who reviewed a draft of this publication, providing us with valuable and timely feedback.

Further reading

Related content

View all related content