Appendix 1 : Details of the main Deprivation Indices used in this report

The information in this appendix was taken from the Compendium of Clinical and Health Indicators 2001 published by the Department of Health and produced by the Health Outcomes and Developments Agency and the London School of Hygiene and Tropical Medicine.

In general, deprivation indices measure the proportion of households in a defined small geographical unit with a combination of circumstances indicating low living standards or a high need for services, or both. An important note to be made about all ecological measures of deprivation – that is measures based on geographic areas, rather than individual circumstances – is that not all deprived people live in deprived wards, just as not everybody in a ward ranked as deprived are themselves deprived.

In interpreting deprivation scores it is important to remember that many deprivation scores are relative measures, that is the score for any one area is standardised by reference to the mean for the total of all areas included in the calculation.

Department of the Environment, Transport and the Regions' Indices of Deprivation 2000

In December 1998, the Department of the Environment, Transport and the Regions (DETR) (now the Department for Transport, Local Government and the Regions) commissioned the University of Oxford to review and update the 1998 Index of Local Deprivation. There were criticisms of the 1998 Index of Local Deprivation (ILD) and the 1991 Index of Local Conditions (ILC) that it updated – the sub-district level indicators were out of date and the chi-squared method needed to be reviewed (the 1998 ILD and 1991 ILC are described briefly below). Also, better small area data at the sub district level was about to come on stream for the first time.

In August 2000, following extensive consultation, DETR published the Indices of Deprivation 2000 (ID 2000). The ID 2000 is made up of:

The new IMD 2000 is an innovative and detailed ward level Index with local authority level presentations. It is based on six separate 'domains' of deprivation:

These each contain information essential to local authorities and others about their areas and the nation-wide picture. The IMD 2000 uses up-to-date information from 33 indicators to describe deprivation at ward level. This includes information from previously untapped data sources, such as Department of Social Security (DSS) benefits data and University and Colleges Admissions Service (UCAS) data. Most of the indicators can be updated regularly, and so form the basis for a dynamic Index. In addition to the Domain Indices, the overall ward level Index of Multiple Deprivation brings this substantial amount of knowledge and information together for the first time.

Drawing together these indicators for the first time gives the IMD 2000 a major advantage over previous indices; the range of indicators at ward level enables a focus on deprivation at a small geographical level that was not possible before. This is an improvement on the 1998 ILD which was able to present very little information at ward level, and the information that was included was based on the 1991 census and therefore was increasingly out of date. In addition, the ward-level information allows the new Index to be presented in six ways to represent overall deprivation and pockets of deprivation at local authority level.

The information from the Indices of Deprivation 2000 has been aggregated to enable local authority districts to be ranked according to how deprived they are relative to other districts using six measures. All of the six measures are equally valid and they should not be used in isolation from each other. There is not one overall set of rankings. Patterns of deprivation are complex – in some places the entire district may be generally deprived – but with no very severe areas. Elsewhere deprivation may be concentrated in very severe pockets that co-exist alongside generally affluent areas. The indices have to attempt to reflect these different patterns through six different measures that reflect the differing mosaic of area deprivation in different areas.

The six measures are:

As all the rankings are based on the same data, there is not one measure that can be used as an overall ranking. Some Authorities may be ranked poorly on some measures, but less so on others.

Department of the Environment's Index of Local Conditions (1991) – superseded by later DETR Indices of Deprivation

The Index of Local Conditions (ILC) comprised thirteen variables, seven census variables and six non-census variables (all 1991 except where stated):

Census variables

1. Unemployment

2. Children in low-earner households

3. Overcrowding

4. Housing lacking basic amenities

5. No car

6. Children in unsuitable accommodation

7. Educational participation

Non-census variables (sources and dates)

1. Long-term unemployment

2. Income support

3. Low educational attainment

4. Standardised mortality ratios

5. Derelict land

6. Home insurance weightings

The index of local conditions is an unweighted summation of the selected indicators using their log-transformed signed chi-square values. ²⁴ The actual number of persons which have each selected variable are compared to the numbers that would be expected if average English rates applied. The difference between the actual and expected numbers is squared and then divided by the expected number after which the value of 1 is added. A log transformation is then applied and those scores where the actual rate was below the expected rate are given negative signs. Summed scores greater than zero indicate greater levels of material deprivation.

This index differs from those previously described in using actual numbers rather than percentage rates as the input into the calculations. This has the effect of giving lower weights to those areas where the actual counts are small – and hence statistically less reliable (i.e. an area where 3 out of 10 persons are unemployed will have a lower score than one where unemployment is 30 out of 100).

Jarman Underprivileged Area Score

The Jarman Underprivileged Area Score was not originally constructed to measure deprivation but as a measure of General Practice workload. The Jarman Score was derived to take account of geographic variations in the demand for primary care based on a survey of GPs subjective expressions of the social factors among their patients that most affected their workload. The variant of the score in most common use –  the UPA8 score – comprises eight variables which were individually weighted by a sample of London GPs.

Unemployment

(3.34) % of residents unemployed as a percentage of economically active

Overcrowding

(2.88) % of residents in overcrowded households (more than one person per room).

Lone parents

(3.01) % of residents in 'lone parent' households.

Under 5s

(4.64) % of residents aged under 5 years.

Elderly living alone

(6.62) % of elderly persons living alone.

Ethnicity

(2.50) % of households headed by a person born outside the United Kingdom.

Low social class

(3.74 ) % of residents where household head is unskilled (social class V).

Residential mobility

(2.68) % of residents who changed address in the previous year.

Each variable is based on the percentage of all residents in households, with the exceptions of unemployment, which is based on the percentage of the economically active population which is unemployed, and residential mobility where the denominator is the total resident population. Each variable is firstly expressed as a proportion (between 0 and 1). The proportions are then transformed by first calculating the square root and then finding the equivalent arc sine (asin). The variables are expressed as z scores and multiplied by their respective weighting factors. The final score is obtained by summing the variables (after statistically reworking). Higher scores indicate greater levels of GP workload.

Townsend Material Deprivation Score

The Townsend Score is based on four variables, originally taken from the 1981 census, that were selected to represent material deprivation: unemployment (lack of material resources and insecurity), overcrowding (material living conditions), lack of owner occupied accommodation (a proxy indicator of wealth) and lack of car ownership (a proxy indicator of income). The Townsend Score is a summation of the standardised scores (z scores) for each variable (scores greater than zero indicate greater levels of material deprivation). Two of the variables – those relating to unemployment and overcrowding – are firstly transformed using the log transformation y = ln (x + 1) to produce more normal distributions. The Townsend score was considered the best indicator of material deprivation available (until the release of the Indices of Deprivation 2000 (ID 2000) in August 2000, although the ID 2000 only covers England, not Wales – unlike Townsend scores). The four variables that make up the Townsend Score are combined together in an overall deprivation index, with each variable being given an equal weight. The units of measurement of the four indicators are:

Unemployment – % of economically active residents aged 16-59/64 who are unemployed.

Car ownership – % of private households who do not posses a car.

Home ownership – % of private households not owner occupied.

Overcrowding – % of private households with more than one person per room.

The variables selected are direct indicators of deprivation, that is, they represent the condition or state of deprivation.

Carstairs and Morris Scottish Deprivation Score

The index was constructed by Carstairs and Morris for the analysis of Scottish health data. Like Townsend, the index is based on four variables originally taken from the 1981 census which were judged to represent, or be determinants of, material disadvantage. Three of the indicators are the same as those used in Townsend, the fourth, social class, is used in place of housing tenure. The units of measurement of the four indicators are:

Overcrowding – persons in private households living at a density of more than one person per room as a proportion of all persons in private households.

Male unemployment – proportion of economically active males who are seeking work.

Social class IV or V – proportion of all persons in private households with head of household in social class IV or V.

No car – proportion of all persons in private households with no car.

The deprivation measure is an unweighted combination of the four standardised variables.

Breadline Britain Score

The Breadline Britain Score is the result of two surveys carried out by MORI for London Weekend Television and the Joseph Rowntree Foundation in 1983 and 1990.

In the 1990 Survey respondents were presented with a set of 44 cards onto each of which was written the name of a different item covering a range of possessions and activities that relate to standards of living. For example, a television, a night out once a fortnight and a warm waterproof coat. Respondents were asked to place the 44 cards into one of two boxes. Box A was for items which they considered necessary; those items which all adults should be able to afford and which they should not have to do without. Box B was for items which they considered to be desirable but not necessary. They were also asked if they felt differently about any of the items in the case of families with children. An item was deemed to be a socially perceived necessity if more than 50 per cent of respondents put it into Box A. Later in the interview the respondents were asked to assign one of the following five options to each of the 44 items:

1.       Have and couldn't do without

2.       Have and could do without

3.       Don't have and don't want

4.       Don't have and can't afford

5.       Not applicable/don't know

Respondents (and their households) were assigned a deprivation index score each time they answered 'don't have and can't afford' to an item that was considered to be a necessity by more than 50 per cent of respondents.

Endnotes

1        For education 'sparsity' is the percentage of an authority's population which lives in wards where the density of population is between 0.5 and 4 people per hectare and 'super-sparsity' is the percentage of an authority's population which lives in wards where the density of population is 0.5 or fewer people per hectare (DETR 1997a) with super-sparsity given twice the weight of sparsity. The definitions for other services are the same but measured at ED level.

2        Where the economy is considered buoyant enough to offer opportunities for all those who wish to participate in the labour market.

3        Using principal components analysis on 15 indicator variables, Pacione (1995) argues that the four indicators of most significance to rural deprivation are self-employment, agricultural employment, car-less households and households lacking basic amenities.

4        Defined as income lower than 140% of supplementary benefit entitlement.

5        The findings of this study are documented in a recent Scottish Executive report (Shucksmith et al, 2000), which also provides a review of other research on rural deprivation and exclusion in England.

6        The rural bus funding, provided by the DETR includes three types of financial support, the rural bus service grant, a challenge fund (covering, for example information and new facilities) and a rural transport partnership to provide community transport.

7        In 1991 the Department of the Environment published an Index of Local Conditions (ILC) for all local authorities, wards and enumeration districts, based primarily on variables from the population census – with chi2 values summed to produce a single index (DoE, 1994 and Bradford et al. 1995). At ED level there are three variables measuring income deprivation (resources domain) unemployed persons, children in low earning households, households with no car; and three measuring housing stress (environment domain) overcrowded households, residents in households without basic amenities and children living in unsuitable accommodation (Robson et al. 1997). Some modest revisions were made to produce the ILD in 1998 (DETR 1998 Index of Local Deprivation: a summary of results) with variables for children in unsuitable accommodation and households lacking a car dropped and a variable for non-income recipients in receipt of council tax benefit added.

8        Now the Countryside Agency.

9   Interestingly, the notions of economic vulnerability or social isolation were not pursued.

10      Initially they also considered the distance from urban settlements as well.

11      Health deprivation and more general deprivation measures were tested.

12      People under the retirement age of 65.

13      We are aware of no similar indicator of relevance to Scotland. Some consideration might be given to exploring similar structures in due course.

14      For instance, UK academics have access to MIMAS (http://www.mimas.ac.uk), which provides pre-calculated ward-level indices for the DoE Index of Local Conditions, Carstairs' Index of Deprivation, Jarman's Under-privileged Area Index and Townsend's Index of Material Deprivation. With precise definitions of many other indices readily available, even basic desktop PCs have the capacity to generate from 1991 census data ward-level measures of deprivation.

15      The ONS have classified all 1991 EDs as either urban or rural on the basis of land use data provided by the Ordnance Survey. Depending on the proportion of their constituent EDs classified as rural or urban, the ONS then placed each of the 8,481 English wards into one of six categories from wholly urban to wholly rural.

16      Morbidity data are derived from the Limiting Long Term Illness (LLTI) question in the 1991 Census. Health status is thus defined in terms of the proportion of people resident in households, under the age of 65, who report an LLTI in the census. The data are age standardised.

17      Mortality data were obtained from the ONS for the number of deaths from all causes to those in the 0–74 age group in the period 1991–1996. These data are age-standardised.

18      The Office for National Statistics (ONS) has generated a geodemographic Ward Classification that classifies all of the wards in England and Wales into one of fourteen categories. The categories 'Rural Fringe' and 'Rural Areas' are used, with the remaining twelve categories aggregated to form the urban group.

19      For instance, a strong relationship between LLTI and the Townsend deprivation index in urban areas (p=0.72) becomes a lot weaker in the rural fringe (p=0.27) and weaker still in rural areas (p=0.18). A similar pattern also emerges from the relationship between mortality and the Townsend index, with the strong relationship in urban areas (p=0.61) again becoming weaker in the fringe (p=0.14) and rural areas (p=0.22).

20      The Standardised Illness Ratio (SIR) is derived from the Limiting Long Term Illness question on the 1991 census.

21      Multi-skilling is likely to prove relatively expensive since higher pay is likely to be required for dual capability and it may actually prove difficult for certain professionals (e.g. midwives) to practice at a sufficient level to sustain their professional status.

22   Undertaken by the NHS National Co-ordinating Centre for Service Delivery and Organization, Research and Development.

23      Our Countryside: the Future: A Fair Deal for Rural England, CM 4909.

24      The standardisation and transformation ‘has the merits of: taking account of the small size of the denominatiors of many of the observations; using an interpretable value of zero; and using values which approximate the normal curve’ (DoE 1994, p.86).