4.5 Deprivation indicators

Measuring Deprivation

Repeated studies have shown a strong relationship between material deprivation, poor health and the need for health care services. In general terms more socially deprived populations experience more ill-health and also underutilize the health care system. There is therefore a role for population-based indicators of need to include some measure of social deprivation as this will capture an element of need for health care which is not necessarily apparent from observing the actual pattern of use. The construction of appropriate indicators of material deprivation in the population can therefore play an important role in the health care planning process. This learning object considers methods for the construction of deprivation indicators, focusing particularly on recent indicators used in the UK.

In the UK context, factors such as low car ownership, unemployment and low status employment, high percentages of social housing, single parents, children and the elderly have all been identified as contributing to need for primary health care. The Jarman index, also known as the Underprivileged Area Score (UPA) was used during the 1980s and 90s to allocate additional resources to general practitioners working in deprived areas. It was based on eight census variables identified from a large survey of GPs as adding to their workload. These were standardized and weighted according to the importance of each in the survey results. The Townsend and Carstairs scores have been used extensively as standard measures of social deprivation in health care analysis and planning studies. They use fewer variables and the selection was based on the expert consideration of their creators rather than a specific empirical study. Similar approaches for the construction of multivariate indicators could be employed in any national setting where baseline small area data are available for a suitable range of variables. The construction of all these indicators involves essentially the same steps, which are illustrated here with reference to the Townsend index.

Pollock and Vickers (1998) provide one example from the very many studies that use the Townsend index to establish a relationship between health and deprivation, concluding that earlier diagnostic and referral procedures for a range of cancers need to be targeted to primary care centres in deprived areas. Pampalon and Raymond (2000) make extensive reference to the Townsend index in developing a deprivation for health care planning in the Canadian context. McIntyre et al. (2000) consider the more challenging transfer of deprivation indicators from the developed world setting to the needs of health care planning in southern Africa.

For many purposes, the UK has now moved away from deprivation indicators based solely on census data and much resource allocation from central to local government is based on a multi-domain Index of Deprivation. The 2004 release of this index is described in Noble et al. (2004) and draws on census data, administrative information from government departments and factors such as distance to services. Because these measures are being used as proxies for a very generalized model of need for health care, there are numerous alternative ways of constructing them, none of which is inherently ‘right’ or ‘wrong’, although it is recognized that some will be more appropriate in specific contexts. Beale et al. (2001) consider the extent to which a measure of housing value such at the UK’s council tax band could be used as an alternative to the Jarman score for health care purposes. Neil et al (2007) consider another approach to the construction of deprivation indices based around multiple criteria evaluation techniques.

Geodemographics

Since the late 1980s, geodemographic classifications have emerged as an alternative means of identifying areas of demand for healthcare services. Geodemographics is the grouping of people in geographic areas according to their socio-economic characteristics, usually for market research purposes. There are now many different geodemographic classifications available, but all share certain characteristics:

• Typically, they draw on a wide census and governmental data, such as details of the models of vehicles registered at particular addresses. Often, such data may be supplemented by information drawn from the private sector, such as large-scale market research surveys or information gathered through store card loyalty schemes.
• Households or neighbourhoods are classified into different types according to their socio-economic characteristics. Detailed household types will be nested within broad classes of households, so that different levels of household differentiation are available, depending on the specific application.
• Typically, geodemographic data are referenced to small areas, a small area being an administrative unit typically with a population of less than 10,000. Many geodemographic data sets provide counts of the numbers of households of a particular type resident within a given small area. The small areas generally follow either census or postal geographies.

One of the earliest and best known geodemographic classifications was Caci Limited’s ACORN (A Classification of Residential Neighbourhoods). However, many different types of geodemographic classification now exist: some provide general descriptions of household consumption patterns (e.g. ACORN; Personicx Geo); others provide information about use of services provided by government (e.g. see Mosaic Public Sector below); whilst others yet provide information about household ethnic origins or levels of engagement with digital technologies. Although the use of geodemographic classifications is increasingly common in the developed world, they rely on the availability of georeferenced digital data on consumption patterns from both the public and private sectors. Such classifications seldom exist in a developing world context because of a lack of such data to form the basis of the classification.

Such geodemographic classifications can help identify the demand for healthcare (Petersen et al, 2010), particularly in private insurance or market-based systems. Geodemographic classifications may help identify those households with private medical insurance, or those groups who have the financial resources to access more expensive, private forms of healthcare.

---

 

---