Quantifying health inequalities in England

Introduction

Non-communicable illness is responsible for 88% of the burden of disease in England, with the majority falling most heavily on the poorest in society. The COVID-19 pandemic has highlighted pervasive socioeconomic, ethnic and geographical health inequalities in our society. But quantifying health inequalities can be difficult due to the complexities of comparing people with multiple different long-term conditions (multimorbidity) and assessing the implications for their health care needs.

In this analysis we use a novel tool known as the Cambridge Multimorbidity Score (see Box 1) to assess the relative impact of different patterns of diagnosed illness on people and their use of the health care system. This score has been developed by doctors, academics and primary care experts, and depends on administrative patient-level data linking primary care and hospital records. Given that our analysis relies on health care data, we are observing diagnosed health inequalities, as opposed to underlying health needs. Inequalities in access to care imply that there may be greater undiagnosed health problems in different population groups – it is not possible to take this into account in this analysis. We briefly outline more about the data and the Cambridge Multimorbidity Score in Box 1 below. More detail on data cleaning and disease definitions can be found in the appendix.

Diagnosed health inequalities by deprivation of local area

We analyse socioeconomic health inequalities by dividing local areas in England into 10 groups, or deciles, according to the Index of Multiple Deprivation (IMD). We then calculate the average Cambridge Multimorbidity Score (CMS) for people living in local areas within each decile. Decile 1 is those living in the most deprived 10% of local areas, while decile 10 is those living in the least deprived 10% of local areas.

Figure 1 shows the average comorbidity score (CMS-weighted prevalence) for diagnosed illness in each decile, broken down by contributing condition. Higher levels of local area deprivation are associated with a higher average comorbidity score. The gradient is steepest for the most deprived deciles, with greater disparities in diagnosed illness for those living in more deprived areas than for people living in the wealthiest half of the country.

Figure 1

Chronic pain, COPD, alcohol problems, anxiety and depression, cardiovascular disease and diabetes contribute to 64% of the burden of disease in the most deprived decile, compared with 49% in the least deprived decile. All of these conditions are at least 50% more prevalent in the most deprived areas of England than the least, and COPD is over three times more common once we standardise for age.

In contrast, cancer is one condition for which the diagnosed prevalence is higher in richer areas. This is a surprising result, as we know from other research that there are higher levels of diagnosed incidence of cancer in more deprived areas. Inequalities in cancer are complex and multifaceted: we see higher levels of cancer incidence by age in poorer areas, but there are further differences in terms of tumour locations, stage at diagnosis and survival rates. People in less deprived areas are more likely to live for longer with a cancer diagnosis, therefore the share of people living with the disease is greater.

People in poorer areas become ill earlier in life

Looking at diagnosed long-term conditions by age, gender and deprivation we can see how much earlier the people living in poorer areas reach a given level of ill health.

Figure 2

Figure 2a plots the average comorbidity score by age group and local area IMD decile. Inequalities in diagnosed long-term conditions are widest in absolute terms for people in their 70s, although in relative terms people living in the poorest areas have an average CMS of more than double those in the least deprived areas in their 50s and 60s.

We can compare how the diagnosed illness (average CMS) of those in the most deprived decile stand in relation to the older people living in wealthier areas. The analysis shows how the inequalities observed in Figure 1 are the result of diagnosed disease emerging earlier in more deprived areas. For example, people aged 30–49 living in the most deprived 10% of local areas have the same average comorbidity score as those aged 50–59 living in the least deprived local areas.

If we compare the extremes, those in the most and least deprived areas, we can see just how different lived experience of health is across the country. Figure 2b shows the levels of diagnosed illness (average CMS) by age and gender for the most and least deprived deciles: on average 60-year-old women in the most deprived decile have around the same level of diagnosed morbidity as 76-year-old women in the least deprived areas. Men have lower levels of socioeconomic inequality in diagnosed ill health at the same age, with a 10-year disparity in diagnosed illness compared with 16 years for women.

For diagnosed illness, there is a clear transition across the life cycle

Figure 3 disaggregates the gap in diagnosed long-term illness between the most and the least deprived areas, showing which conditions contribute to the inequalities in each age group. It is striking that over 50% of inequality in diagnosed illness for all age groups older than 30 years is caused by only three conditions: chronic pain, COPD and diabetes. These conditions, along with alcohol problems, heart failure and dementia, account for at least three-quarters of the absolute inequality and most of the burden of disease that exists within the population aged 50 years and older.

Figure 3

In Figure 3, we show the contribution of each condition to health disparities by age group. Absolute levels of inequality are very small for younger people, but this analysis shows that there is a clear transition of inequality across the life cycle. Children in more deprived areas are much more likely to be living with asthma, epilepsy and to experience alcohol problems and are less likely to be diagnosed with anxiety or depression. From the age of 20 we see greater differences in chronic pain, alcohol problems and anxiety and depression. From the age of 30 we begin to see health inequalities manifesting in different kinds of health conditions. Disparities in diabetes, COPD and cardiovascular disease rates grow and overtake anxiety and depression, although we still observe growing inequality in chronic pain and alcohol problems. Among people aged 50 and older, inequality in morbidity is dominated by chronic pain, COPD, diabetes and dementia. Contrary to other diseases, we see a higher prevalence of diagnosed cancer in the least deprived areas among older patients.

These findings suggest that health inequalities start at an early age and accumulate over time. Chronic stress and adverse conditions have been shown to contribute directly and indirectly to poorer health outcomes through their cumulative impact. This phenomenon, known as allostatic load, is linked to earlier death and is greater for those from poorer backgrounds. The stages of inequality in ill health shown here strengthens the case for a life course approach to public health policy.

The disparity in diagnosed ill health then reduces for people aged 80 or older. People living to this age have time to develop long-term illness even if they live in the least deprived areas. This is a study of people living with diagnosed illness and so does not reflect the stark inequalities in mortality rates. It is therefore worth thinking about the amount of time spent in ill health in the context of life expectancy.

Life expectancy and living with illness

Differences in the prevalence of diagnosed long-term conditions can be explained by inequalities in the wider determinants of health. Here, we explore the age at which we see disparities in health starting to develop.

We have produced alternate estimates of life expectancy to national statistics using our primary care data linked to mortality records, along with an additional measure of disease-free life expectancy (see Table 1). Disease-free life expectancy provides an estimate of the average number of years a person would live without being diagnosed with any of the illnesses included in the CMS.

Inequalities in life expectancy are wider for men, but women spend longer in ill health

Our estimates reinforce the large inequalities in life expectancy indicated by official national estimates (Table 1). Both sets of estimates show that inequalities in life expectancy are greater for men than for women. They also both estimate that on average men living in the poorest areas live 9 fewer years  than their counterparts in the wealthiest areas of England. The socioeconomic inequality in life expectancy for men is such that, despite spending a greater share of their lives with diagnosed illness (46% compared with 44%, Table 1), men in the most deprived areas spend less time living with diagnosed illness (35.7 years compared with 37.9 years in the least deprived areas).

But women overall face more time spent living with illness. Women in the poorest areas are expected to be diagnosed with a long-term illness at 40 years of age and are expected to spend over half their lives in ill health (52%). Women in the least deprived areas, on the other hand, on average develop their first illness 8 years later. They are expected to live an additional 5 years than those in the most deprived areas according to our estimates (the equivalent figure is an additional 8 years from official national estimates).

There are inequalities by region and by ethnic background

This novel method of analysis clearly shows inequalities in diagnosed illness both by the patient’s ethnicity and where in England they live. We observe more diagnosed ill health in some people from minority ethnic backgrounds (Figure 4) and in the north of England (Figure 5). As evidenced in previous research, health inequalities are heavily linked to ethnic and geographic socioeconomic inequalities. For example, a recent study showed that once you consider the differences in socioeconomics and other important factors between ethnic groups, the differences in rates of obesity between ethnic groups are either severely diminished or removed entirely. There is much we still do not know about the interactions between ethnicity, region, the wider determinants of health, and how these impact access and need for health care.

To directly compare average rates of illness for people across ethnicities is to ignore their demographic differences. The average age of the white population is higher than those from other ethnic backgrounds. The white population in this analysis includes those from minority white backgrounds. As shown in Figure 4, once we standardise for different age distributions across ethnicities, we find higher levels of diagnosed ill health in Pakistani, Bangladeshi, and black Caribbean people than among the white population in England.

Figure 4

People from Pakistani and Bangladeshi ethnic backgrounds in England have the highest age-standardised rates of chronic pain, diabetes, dementia and cardiovascular disease. Despite not having the highest prevalence in other cardiovascular disease, the white population has the highest rate of diagnosed atrial fibrillation, an important risk factor for cardiovascular disease that can lead to preventive intervention in primary care. The white population also have the lowest rate of diabetes but are more likely to be living with diagnosed anxiety or depression, alcohol problems and cancer.

As previously discussed, these data reflect disparities in diagnosed health conditions and therefore they only show inequalities in needs that are identified by the health system. Evidence suggests that access and engagement with care is limited for some population groups, often on the basis of ethnicity. For instance this year, a review from NHS Race and Health Observatory and the Race Evidence Review found evidence to suggest that there are ‘clear barriers’ for people from minority ethnic backgrounds to seeking help for mental health problems. Other research has also found lower access to cancer screening and lower levels of patient satisfaction. The results presented here should therefore be considered in this context and further research is required to accurately represent underlying health inequalities.

Figure 5 shows the age-standardised average morbidity score by region, adding to the evidence of a north-south health divide. People living in the North East and North West regions of England have the highest health care needs, in part due to higher levels of chronic pain, alcohol problems, COPD and cardiovascular disease. These four conditions account for 83% of the inequality in diagnosed illness between the North East (highest) and the East of England (lowest). The north-south divide seen here is reflected in regional disparities in average incomes, wealth, economic opportunity and educational attainment, and speaks to the strong relationship between inequalities in the wider determinants of health and inequalities in diagnosed ill health.

Figure 5

Conclusions

There is a vital need to quantify health inequalities to better focus health and wider policy to address them. Equality is at the centre of NHS England’s strategy, but insights on the extent of health inequalities like those presented here are few and far between. Our analysis adds to a growing body of evidence that health outcomes for disadvantaged groups are dramatically worse across a number of measures, with poorer people living shorter lives in greater discomfort due to ill health. These socioeconomic inequalities are borne out in greater levels of diagnosed ill health among minority ethnic populations and in different regions of England.

Our results show health inequalities starting at a very early age and continuing to develop through adulthood. The early ages and changing structure of health inequalities reinforces the notion that nothing short of a joined-up policy approach can address the wide and complex health inequalities we see in England. We believe this new presentation of health inequalities shows powerfully that investing in the circumstances in which people live will help people stay healthier for longer.

These are some of the first results from analysis conducted by the Health Foundation’s REAL Centre in the development of a health and care funding projections model, in partnership with the University of Liverpool’s Institute of Population Health. The analysis was conducted using the Clinical Practice Research Datalink (CPRD). The data are provided by patients and collected by the NHS as part of their care and support. Regulatory approvals to use CPRD data for this analysis were granted by the CPRD Independent Scientific Advisory Committee (ISAC protocol number 20-000096).