There are several area-based tools which characterise the socio-demographic make-up of local populations, which can be useful in understanding health needs and inequalities, even if they were originally developed for other purposes. All of these characterise areas and cannot be validly used to target individuals as any individual may not share the characteristics of the area (the ecological fallacy).
Index of Multiple Deprivation (IMD)
This is produced by the Department for Communities and Local Government (DCLG), and was originally issued in 2000, and showed relative levels of social and economic deprivation across all the counties of England at a ward level. In its original form it considered these seven criteria:
Income
Employment
Health
Education
Housing
Access
Child Poverty
Subsequent debate led to the decision to issue further versions at smaller geographical units, and to some revision of the criteria used.
In the next release, IMD 2004, the geographical unit chosen was the Lower Super Output Area (LSOA) an area grouping that had been developed for the Census 2001 (see the Census section above). Aggregated results were available for larger areas such as Local Authority districts. The criteria were referred to as Domains, which could be separately reported, if desired.
In the 2004 release:
Health was replaced by Health and disability
Education was replaced by Education, skills and training
Housing was replaced by Barriers to Housing and Services
Access was replaced by Living environment
Child Poverty was replaced by Crime
Each Domain contained a number of indicators, 37 in all. Two supplementary indices were created as a subset of the Income domain. These relate to income deprivation affecting children and income deprivation affecting older people.
The IMD 2007 and 2010 releases followed the same structure as IMD 2004.
Full details are available from: https://www.gov.uk/government/statistics/english-indices-of-deprivation-2010 [accessed 11 January 2016]
The IMD 2015 release used the redrawn LSOA boundaries from the 2011 Census,
but in other ways followed a similar approach to previous releases. Full details are available from: https://www.gov.uk/government/statistics/english-indices-of-deprivation-2015 [accessed 11 January 2016]
Strengths
Consistency: the same data sets are used across England, enabling valid comparisons over the whole country. There is no cost to the user.
Studies have repeatedly shown that IMD values are strongly correlated with many health outcomes, from mortality to childhood obesity. This holds even when the (relatively small) health and disability domain is excluded.
The IMD is widely used across central government to focus programmes on the most deprived areas. Locally, it is often used as evidence in the development of strategies, to target interventions, for understanding differences between areas, e.g. within Local Authorities and in bids for funding. The voluntary and community sector also uses the Index, for example, to identify areas where people may benefit from the services they provide.
Weaknesses
Each country in the UK produces its own version of the Index of Multiple Deprivation using similar methodologies. However, differences in the indicators used, the time periods covered and the sizes of their small areas, mean that it is not possible to make direct comparisons between these indices.
The IMD shows relative levels of deprivation, but not magnitudes. You can say area A is more deprived than area B, but not by how much.
The indicators use specific components that should not be generalised beyond their meaning. The largest component domain, Income Deprivation, uses only data on people in receipt of benefits and tax credits, and asylum seekers. It is therefore no indication of wealth or affluence as data on actual incomes are not included.
Despite apparent similarities of method, the IMD is not designed to provide ‘backwards’ comparability with previous versions so these should not be used as a time series. Changes in definitions between releases (for example, changes to benefit entitlements) mean that apparent similarities can be misleading.
Census Data: Area Classifications
Following the 2001 Census, an Output Area Classification was developed at the University of Leeds and published by ONS. Output Areas are the smallest geographical unit for publication of Census data.
Area classifications group together geographic areas according to key characteristics common to the population in that grouping, using statistical software to perform a multi-variate segmentation.
The process was repeated for the 2011 Census. The first of the 2011 area classifications to be published was the 2011 Area Classification for Output Areas (2011 OAC), which was produced by an external partner, University College London. Following on from this, a 2011 Area Classification for Local Authorities has been published.
Output Area groupings are called clusters and are derived using census data. The area classifications are hierarchical classifications, consisting of three tiers: Supergroups, Groups and Subgroups. These relate to degree of similarity between areas as determined by statistical analysis of 59 census variables. This makes the OA 2011 useful in identifying suitable comparator areas for use in benchmarking. Full details of the variable and analysis performed can be found at: http://www.ons.gov.uk/ons/guide-method/geography/products/area-classifications/ns-area-classifications/ns-2011-area-classifications/methodology-and-variables/index.html [accessed 11 January 2016).
Data is presented for local authorities by their position in the hierarchy as "pen portraits". For example:
Supergroup: 3 – London Cosmopolitan
20 local authorities – 7.7% of UK population, median age 32
The population of this Group is located within 19 of the 33 London boroughs, with all Inner London boroughs represented. The areas covered by this Supergroup are characterised by a very high population density, and a high proportion of young to middle aged adults (aged 25 to 44). All non-white ethnic groups have a higher representation than the UK average especially people of Black or Asian ethnicity, with an above average number of residents born in other EU countries. Residents are more likely to live in flats and more likely to rent. A higher proportion of people use
public transport to get to work, with lower car ownership, and higher unemployment. Those in employment are more likely to work in the accommodation, information and communication, financial, and administrative related industries.
Group: 3a – London – Cosmopolitan North London
8 local authorities – 3.5% of UK population, median age 32
When compared with the parent Supergroup, this Group has a higher proportion of children aged 5 to 14, adults are more likely to be married, and residents more likely to have been born in the new EU. Households are more likely to live in semi-detached properties, and to own their own home. Households are more likely to own 2 or more cars, and to use private transport for travelling to work. Resident workers are more likely to work in the following industries: (1) mining, quarrying and construction, and (2) wholesale and retail trade.
Subgroup: 3a1 – Cosmopolitan North London
5 local authorities – 2.3% of UK population, median age 32
This Subgroup has a higher proportion of people with Indian/Pakistani/Bangladeshi ethnicity when compared with the Group. The population is more likely to have been born in the new EU, and more likely to have a main language other than English. Households are more likely to live in privately rented accommodation. Employment by occupation is similar to that of the parent Group.
Strengths
Good for identifying similar areas for benchmarking and other comparisons.
Methodology is open and published.
Weaknesses
Limited direct relevance to health issues.
Census data becomes less meaningful the further in time from the census date.
Other geodemographic classification systems
Several large market research organisations have developed variations on their basic systems in attempts to persuade health organisations to use their services. The best known of these (in the UK) are:
Mosaic, from Experian, essentially census based, with input from a range of other data sources (e.g. marketing databases). Originally developed at postcode level, more recent versions operate at the individual household level. Mosaic UK 2009, classifies the UK population into 15 main socio-economic groups and, within this, 67 different types. Mosaic’s naming proved to be controversial leading Experian to introduce Mosaic Public Sector with more politically correct segment names.
Acorn from CACI. Originally also census-based, the current version (issued in 2013) eschews census data for other UK government Open data along with other sources. Acorn segments households, postcodes and neighbourhoods into 6 categories,18 groups and 62 types.
P2 People and Places, from Beacon Dodsworth. This uses data from public and private sector sources (including census data) to categorise people. Information about UK consumers is then presented in the form of demographic types. P2 geodemographic classification is structured as 16 Trees and 44 Branches.
Strengths
Proprietary geodemographic systems have all been developed for marketing purposes. They can sometimes be useful in social marketing, where the purpose is to identify target groups which might benefit from public health interventions.
Weaknesses
Their prime function is financial: to increase their clients' sales. Health-associated information is a side-benefit.
They are generally expensive.
© M Goodyear 2016 and S Seager 2018