The regional health gap between classes is mostly larger for women, except for London where it is comparable to men, and the South East where it is larger for men.
Islington had the largest health gap in rates of ‘Not Good’ health between occupational classes for both men and women; a difference of 33.3 and 31.4 percentage points respectively.
It is estimated that an additional 1.6 million men and 1.8 million women would be assessing their health as ‘Good’ if they had the same self-assessed health rates as those in the most advantaged occupations such as lawyers and medical doctors.
There is far more variation in the rates of ‘Not Good’ health in the most socio-economically disadvantaged classes across regions and local authorities; the size of the health gap within areas is mostly driven by the rates in these classes.
The local authorities with the largest health gaps are generally found in large population centres, such as Inner London.
This short story investigates the differences in age standardised rates of ‘Not Good’ health between groups of people based on their occupation and employment contracts (socio-economic class).These differences can be described as the health gap or inequality and can be compared between classes in the same geographical location. It is also valid to compare the rates of the same socio-economic class between areas and between men and women to show within class inequality and gender inequality.
Tackling health inequalities is an important policy issue as highlighted in the Marmot review. Census data provides an opportunity to measure health inequalities in ways that have benefits compared to using conventional survey data. The census has the advantage of high population coverage (the number and types of people surveyed), which increases the accuracy and precision of the rate estimates (greater certainty that the estimates reflect the true values that have been estimated).
The 2011 Census asked two questions related to health. The first question asked about an individual’s general health. An individual responding to Question 13 (Figure 1) can be categorised as having ‘Good’ health if they assessed their health as either Very good or Good, or as having ‘Not Good’ health if they assessed it as Fair, Bad or Very bad.
The second health question was asked to measure activity limitations (disability). A similar analysis to that reported here measuring socio-economic gaps in disability prevalence will be released in December 2013.
People1 aged 16 years and above who reported a current or (if not currently working) previous occupation and an employment status (whether a manager, a supervisor, an employee or self-employed) were placed into a socio-economic class on the basis of these personal details. The name for this grouping of occupations is the National Statistics Socio-economic Classification (NS-SEC). Those who had never worked or had been unemployed for six months or more or had no occupational information included in their census form were excluded from this analysis.
Unlike the health questions asked in the 2011 Census, the questions used to derive NS-SEC are numerous and the derivation procedure is complex. The 2011 Census questions used in this derivation of NS-SEC have been included in the background notes section of this article. More information about NS-SEC can be found on the ONS website.
|NS-SEC Class||Examples of Jobs|
|1. Higher managerial and professional||Lawyers, Architects, Medical doctors, Chief executives, Economists,|
|2. Lower managerial and professional|
|Social workers, Nurses, Journalists, Retail managers, Teachers|
|3. Intermediate||Armed forces up to sergeant, Paramedics, Nursery nurses, Police up to sergeant, Bank staff|
|4. Small employers and own account workers||Farmers, Shopkeepers, Taxi drivers, Driving instructors, Window cleaners|
|5. Lower supervisory and technical|
|Mechanics, Chefs, Train drivers, Plumbers, Electicians|
|6. Semi routine||Traffic wardens, Receptionists, Shelf stackers, Care workers, Telephone salespersons|
|7.Routine||Bar staff, Cleaners, Labourers, Bus drivers, Lorry drivers|
Health outcomes have been shown to vary markedly between people depending on their socio-economic position based on occupation. Socio-economic position is a good indicator of the general living conditions, access to goods and services, career development prospects, educational attainment, salary range, disposable income, wealth, assets and social standing: Such factors are important drivers of well-being and health.
Analysis of the 2001 Census found that there were significant health gaps between the least and most advantaged in terms of ‘Good’ health, life expectancy and healthy life expectancy (FOH 2006, ONS 2011, White and Edgar 2010). The statistics produced in this short story uses more up-to-date data from the 2011 Census to provide a picture of health inequality between occupations assigned to different socio-economic positions within and across geographical locations. Such information can be used as evidence to justify health policies aimed at tackling inequality in health.
In this analysis the age standardised rates of ‘Not Good’ health of those aged 16 and over and classifiable into the reduced NS-SEC scale have been calculated.
Calculating the crude rates of health for a population is relatively straightforward. The problem is that using crude health rates can lead to unreliable comparisons of populations with different age structures. This is because there is a strong relationship between health and age, with health worsening with increasing age. Therefore, it would be expected that a population with an older age structure would have a higher crude rate of ‘Not Good’ health than a population with a younger age structure, all else being equal, which could bias any comparisons. This age bias can mask other important differences between populations such as socio-economic class, gender or geographical location.
The method of age standardisation applies an adjustment through weighting which enables different populations to be compared on an equal footing in terms of age structure. Age standardisation gives a health rate for the whole population of interest, by removing age as a factor and thus allowing meaningful comparisons between genders, socio-economic classes or geographical locations. The European Standard Population (ESP) 2013 has been used in this short story. Please see the background notes section for more detail.
The distribution of men and women in each socio-economic class was not equal. More men and women worked in 2011 (or had worked) in lower managerial and professional occupations (Class 2) than in any other socio-economic class at 22.1 per cent and 25.7 per cent respectively. The smallest class for men was the intermediate occupations (Class 3) at 7.7 per cent and lower supervisory and technical occupations (Class 5) was the smallest for women at 4.8 per cent.
In the most socio-economically advantaged higher managerial and professional occupations (Class 1), there was a larger proportion of men (15.6 per cent) than women (7.7 per cent). In the most disadvantaged routine occupations (Class 7) the proportions of men and women were more comparable at 15.0 per cent for men and 11.9 per cent for women.
It is interesting to note the larger proportions of men in small employers and own account workers (Class 4) and lower supervisory and technical occupations (Class 5). Far greater proportions of women were working or had worked in the intermediate occupations (Class 3).
The rates of ‘Not Good’ health presented and discussed in this analysis are the age standardised rates. Please see background notes for more detail.
For men and women at the England and Wales level there is a general pattern of increasing rates of ‘Not Good ‘health in line with increasing disadvantage associated with the socio-economic position of the occupation (i.e. Class 7 had worse health than Class 1).
This general pattern is also observed in each English region and Wales. However, although the overall pattern between classes is similar in each region, there is considerable variation between the health rates within each class between regions. This suggests it is not only where you are in the socio-economic scale which influences your health but also your geographical location. For example, the rate of ‘Not Good’ health for women in routine occupations in Wales was 38.5 per cent; 8.5 percentage points higher than for women in routine occupations in the South East at 30.0 per cent.
The regions that generally had the lowest rates of ’Not Good’ health for men and women in all NS-SEC classes were the East of England and South East respectively, whereas the North East and Wales generally had the highest rates for men and women. To illustrate the effect of region of residence on the social differences in health rates, men working in the East of England and in Class 7 had lower rates of ‘Not Good’ health than those in Class 5 in the North East at 27.0 per cent and 27.9 per cent respectively (Figure 3). For women, a similar pattern was observed with workers in Class 7 in the South East having better health rates than workers in Class 5 in Wales at 30.0 per cent and 30.9 per cent respectively (Figure 4).
A comparison of regions shows there was more variability in self-assessed health within the less advantaged NS-SEC classes, particularly in Class 7, but less variability in the more advantaged classes, particularly Class 1. For example when considering the regions for men there is only a 2.4 percentage point difference in ‘Not Good’ health rates in Class 1 between the East of England and the North East, while for Class 7 the difference was greater at 7.2 percentage points. This means the health of those in Class 1 seems to be less affected by the region they live in, whereas the health of those in Class 7 were more affected. A similar pattern was also observed for women.
Generally, the range is greater in those regions where rates of ‘Not Good’ health are highest. For both genders the regions with low rates of ‘Not Good’ health and low ranges are found in the South; specifically, the East, South East and South West. The regions with the highest rates of ‘Not Good’ health and highest ranges, thus having a greater health gap, are the North West, and the North East, plus Wales. As the rates of ‘Not Good’ health are more variable between regions among Class 7, it is the less favourable health of the men and women in Class 7 in the North of England and in Wales that is contributing most to the large health gaps found there.
If the South East, South West and East of England in this context are defined as the South, and the North East, North West, and Yorkshire and the Humber are defined as the North, then a clear North-South divide exists, both in terms of the difference in class specific age standardised health rates and the size of the gaps between Class 1 and Class 7. London is not included in the South due to its differing characteristics which include its transient population, access to services and limited rural population.
In the context of current health legislation in England, a duty is placed on health organisations to have regard to reducing health inequalities when formulating policies and making decisions about funding health services. For the purposes of the government policy, a large health gap is undesirable.
The range in health rates can be viewed as a simplified version of the health gap. In order to measure the health gap more reliably, a statistic named the Slope Index of Inequality (SII) is used. The SII measures the health gap by accounting for the inequality existing between the other classes (classes 2 to 6) and also accounting for the varying distribution and numbers working in the differing socio-economic classes between regions and local authorities.
In this analysis the SII represents the absolute difference in health rates between the least and most disadvantaged socio-economic classes, taking account of the health rates in all the classes. Smaller SII values represent narrower health gaps, while larger values represent larger health gaps and greater inequality. The SII is reported here as the percentage of people with ‘Not Good’ health. For example, an SII of 10 per cent means that the difference between the rates of ‘Not Good’ health for the most advantaged men and women in Class 1 and the least advantaged in Class 7 is 10 per cent. Therefore if the most advantaged in Class 1 have a ‘Not Good’ health rate of 15 per cent then the least advantaged in Class 7 would have a ‘Not Good’ health rate of 25 per cent.
Using the SII, the health gap is larger for women in every region except the South East, where the health gap is larger for men, and in London where the health gap is identical1 for both genders (see table 2).
For men the largest health gap was in the North East at 21.6 per cent followed by the North West at 21.5 per cent and Wales at 21.0 per cent. The smallest health gap for men was in the East of England at 16.7 per cent followed by the South West at 17.2 per cent and the South East at 17.7 per cent.
For women the largest health gap was in Wales at 23.4 per cent followed by the North East at 23.1 per cent and the North West at 22.4 per cent. The smallest health gap was in the East of England at 17.2 per cent followed by the South West and South East both at 17.4 per cent.
Between regions the health rates of those in the least advantaged socio-economic classes vary more than the health rates of the most advantaged classes. In addition, regions with higher rates of ‘Not Good’ health across all NS-SEC classes also have the largest health gaps. The main driver for a small health gap in a region was the comparatively low rates of ‘Not Good’ health in classes 5, 6 and 7. The regions with a large inequality had comparatively high rates in these most disadvantaged socio-economic classes.
|Region/Country||Health gap men (SII)||Heath gap women (SII)||Gender difference in health gap||Rank men Health gap||Rank women Health Gap|
|Yorkshire and The Humber||19.9||21.4||1.5||5||6|
The fact that the size of the health gap varies between regions suggests that socio-economic position alone does not fully account for inequality in health. Regional factors such as varying job certainty, housing quality, access to services and differing concentrations of hazardous occupations (within the occupational groupings) can contribute to variations in the size of the health gaps between regions.
It is possible to demonstrate the health benefits of removing inequality between the most advantaged class and the other classes. If all working people (or those that have worked) in England and Wales were to have the same rate of ‘Not Good’ health as someone of comparable age and sex in the higher managerial and professional grouping (Class 1) then 1.8 million more women and 1.6 million more men would have assessed their health as ‘Good’ rather than ‘Not Good’.
As mentioned previously, in the context of health policy, reducing inequality is desirable. Tables 3 and 4 show the 10 local authorities with the smallest health gaps between socio-economic classes and the 10 local authorities with the largest gaps for men and women respectively. Again, the statistic used to measure the health gap is the Slope Index of Inequality (SII), where a small SII value indicates a small health gap or inequality and a large SII value indicates greater inequality.
Unsurprisingly, the 10 local authorities that have the smallest health gaps are generally found in those regions with small health gaps. South Holland, in the East Midlands, was the most equal for men with an SII of 9.8 per cent, while Rochford, in the East of England, was the most equal for women with an SII of 11.1 per cent. South Holland, Blaby and Broadland all feature in the top 10 most equal local authorities for both men and women.
A common feature of the 10 local authorities with the smallest health gap is the absence of large population centres. By examining the health rates of the most disadvantaged Class 7 in these authorities, they all had rates of ‘Not Good’ health much lower than the England and Wales average for routine workers, for both men and women. This demonstrates that the relatively better health of those in the more disadvantaged classes leads the inequality to be smaller.
When considering the 10 local authorities with the largest health gaps for men and women, seven were in London for men and five for women. Islington had the highest health gap for both men and women with an SII of 33.3 per cent and 31.4 per cent respectively. Tower Hamlets, Cardiff, Camden and Newcastle upon Tyne also featured in the 10 with the largest inequality for both men and women.
The local authorities with large health gaps have the common feature of being large urban areas with more pronounced differences in the levels of deprivation their resident populations are exposed to. When examining self-assessed health for those in the more disadvantaged classes in these local authorities, it showed they all had higher rates of ‘Not Good’ health than the England and Wales average for their specific class.
The percentage point difference in the health gap between the local authorities with the highest and lowest gaps was 23.5 per cent for men and 20.3 per cent for women. These differing inequalities mean the health gap for men in Islington is three times greater than that of men in South Holland.
|Smallest 10||Local Authority Name||Region||Health Gap (SII)||LA Rank|
|South Holland||East Midlands||9.8||1|
|Largest 10||Local Authority Name||Region||Health Gap (SII)||LA Rank|
|Newcastle upon Tyne||North East||25.2||338|
|Hammersmith and Fulham||London||27.9||341|
|Kensington and Chelsea||London||28.6||342|
|City of London, Westminster||London||30.7||343|
|Smallest 10||Local Authority Name||Region||Health Gap (SII)||LA Rank|
|East Dorset||South West||11.8||3|
|South Holland||East Midlands||12.6||6|
|Largest 10||Local Authority Name||Region||Health Gap (SII)||LA Rank|
|Blackburn with Darwen||North West||25.2||337|
|Neath Port Talbot||Wales||25.4||338|
|Newcastle upon Tyne||North East||28.2||345|
In order to reduce the health gaps it is important to understand what is driving the difference between local authorities. Analysing all local authorities there is a strong correlation between the self-assessed ‘Not Good’ health rate of those in Class 7 and the size of the health gap (correlation coefficient1 0.75 for men and 0.86 for women).The correlation between the ‘Not Good’ health rates of those in Class 1 and the health gap is either non existent or very weak (correlation coefficient 0.06 for men, 0.32 for women). Figure 7 shows that there is an increase in the correlation with decreasing socio-economic advantage for men and women. This suggests the health gap observed in all local authorities is determined to a greater extent by the health rates of those in the more disadvantaged socio-economic occupational classes; specifically the Lower supervisory and technical, Semi-routine and Routine classes.
The differing distributions of particular occupations or working practices that are hazardous to health, between regions could explain why the rates of ‘Not good’ health vary more among the most disadvantaged socio-economic classes. For example, night time working may be more prevalent for routine workers in large urban centres in the North than in more rural settings in the South.
In addition, the geographic variations in the rates of ‘Not Good’ health among the most disadvantaged socio-economic classes may also be affected to some extent by the deprivation they experience in their area of residence (White and Edgar 2010).In the context of health policy, these results suggest greater reductions in health inequalities could be made nationally by improving the health of the semi-routine and routine classes in those authorities with the widest health gaps.
Focus On Health (FOH) 2006 (2.85 Mb Pdf) , available on the ONS website.
ONS (2011) ‘Trends in life expectancy by the national statistics socio-economic classification 1982-2006’ (129.1 Kb Pdf) , available on the ONS website
White, C. and Edgar, G. ‘Inequality in healthy life expectancy by social class and area type England: 2001-03, Health Statistics Quarterly, 45, 2010 (749 Kb Pdf)
pp 28-56, available on the ONS website
Please explore the data yourself using our interactive maps.
Questions 26-38 in the 2011 Census were used to derive each individuals NS-SEC and can be seen in box 1 below
Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: firstname.lastname@example.org