Key Health Data for the West Midlands 2005 CHAPTER EIGHT: A COMPARISON OF DEPRIVATION INDICES: TOWNSEND 4 AND INDEX OF MULTIPLE DEPRIVATION 2004 |
|||||||||||||||||||
Main Body 7: Trend Data for Selected Local Authority Profile Indicators 8: A Comparison of Deprivation Indices: Townsend 4 and Index of Multiple Deprivation 2004 |
8.1 IntroductionThe last Census in 2001 offers a useful comparison of deprivation levels across the decades. There appears to be a widening gap in deprivation indices across the decades (McCarron 1994, McLoone1994, Phillimore 1994) The problem with addressing this issue, is that many articles use deprivation indices that were relevant to 1991 and 1981 census. The most often used deprivation indices are Townsend 4 and Jarman 8 scores. However the Office of National Statistics (ONS) and the Office of the Deputy Prime Minister have utilised a more wide ranging basket of indices which are known as the Index of Multiple Deprivation of which the current version is 2004 (IMD 2004). There has been considerable interest in comparing different forms of Deprivation Score. However a search on Pub Med, BMJ and Medline indicate that no papers have been published since the last census comparing these different indices (search criteria comparison deprivation score). The databases searched include the following Allied & Complementary Medicine - 1985 to date (AMED)
8.2 BackgroundThe Townsend Material Deprivation Index is made up of the following 4 Census 1991 variables:
It is important to note that the weight given to each of the 4 variables is equal – ie they account for 25% of the score. This is in sharp contrast to the Index of Multiple Deprivation and related domains which have unequal weights. IMD 2004 is a composite index. It contains seven domains that relate to Income deprivation, Employment deprivation, Health deprivation and disability, Education, skills and training deprivation, Barriers to Housing and Services, Living environment deprivation and Crime. The IMD 2004 index is a weighted index – ie the relative influence of each of these domains is not equal. The weights are as described below: Table 8.01: Domains and Weights used in the Calculation of IMD 2004
8.3 Domains used in the calculation of IMD 2004The domains are a composite of census and other variables derived from other government department statistics. These are updated at different time intervals and hence overcome one of the criticisms of the Townsend 4 scores as being a historical snap shot of the population on Census night. The domains are described below and their composite variables are described accordingly: 8.3.1 Income Deprivation DomainThe purpose of this Domain is to capture the proportion of the population experiencing income deprivation in an area.
In addition, an Income Deprivation Affecting Children Index and an Income Deprivation Affecting Older People Index were created. 8.3.2 Employment Deprivation DomainThis domain measures employment deprivation conceptualised as involuntary exclusion of the working age population from the world of work.
8.3.3 Health Deprivation and Disability DomainThis domain identifies areas with relatively high rates of people who die prematurely or whose quality of life is impaired by poor health or who are disabled, across the whole population.
8.3.4 Education, Skills and Training Deprivation DomainThis Domain captures the extent of deprivation in terms of education, skills and training in a local area. The indicators fall into two sub domains: one relating to education deprivation for children/young people in the area and one relating to lack of skills and qualifications among the working age adult population. The sub domain Children/young people comprises of the following;
The sub domain Skills comprises of the following:
8.3.5 Barriers to Housing and Services DomainThis domain is to measure barriers to housing and key local services. The indicators fall into two sub-domains: 'geographical barriers' and 'wider barriers' which also includes issues relating to access to housing, such as affordability. The sub domain wider barriers comprises of the following;
The sub domain geographical barriers comprises the following;
8.3.6 The Living Environment Deprivation DomainThis domain focuses on deprivation with respect to the characteristics of the living environment. It comprises two sub-domains: the 'indoors' living environment which measures the quality of housing and the 'outdoors' living environment which contains two measures about air quality and road traffic accidents. The sub-domain 'indoors' living environment comprises of the following
The sub-domain 'outdoors' living environment comprises the following;
8.3.7 Crime DomainThis domain measures the incidence of recorded crime for four major crime themes, representing the occurrence of personal and material victimisation at a small area level.
The methodological steps that were taken to create the IMD 2004 are described in the full report found on the Office of Deputy Prime Ministers (ODPM) www.odpm.gov.uk 8.4 Aims of this ChapterThe purpose of this chapter is to provide a comparison between Townsend 4 scores and the Index of Multiple Deprivation 2004 (IMD 2004) and its constituent domains. It has not been possible to compare the scores with health statistics, However this chapter will provide a correlation comparison between the scores to enable the reader to have a considered opinion as to which deprivation or domain score to use when. 8.5 MethodologyIMD 2004 data relating to super output areas was downloaded from the ODPM website. The data contained the IMD 2004 scores, constituent domain scores and ranks. The Townsend 4 scores were calculated at SOA level by the Department of Public Health and Epidemiology and the Department of Biostatistics at Leeds University. Details of the methodology can be obtained if required. The following 8 graphs compare Townsend 4 (T4) scores with the IMD 2004 composite score and the domains. 8.6 Results8.6.1 IMD and T4Figure 8.01: Comparison between IMD and T4
There appears to be a close correlation between the two variables. The exponential equation appears to fit the data better, despite both scores being transformed. 8.6.2 Income Domain & T4Figure 8.02: T4 and Income
This describes the same correlation, but the R2 is better than with the IMD 2004. 8.6.3 Employment Domain and T4Figure 8.03: T4 and Employment
There also appears to be an exponential relationship between the variables, but the regression coefficient R2 is not as large as the income domain or IMD 2004. 8.6.4 Disability Domain and T4Figure 8.04: T4 and Health Deprivation/Disability
There is a linear relationship between the two variables and MS Excel was unable to draw an exponential trend line because both scores became negative. 8.6.5 Education Domain and T4Figure 8:05: T4 and Education
The relationship between Education Domain and T4 (figure 8.4) is not as correlated as the above, however there appears to be more of an exponential relationship between the variables. 8.6.6 Housing Domain and T4Figure 8.06: T4 and Housing
There appears to be no correlation between these two variables 8.6.7 Environment Domain and T4Figure 8.07: T4 and Environment
There is a relationship between these variables, but it is weaker than most of the previous domains. 8.6.8 Crime Domain and T4Figure 8.08: T4 and Crime
There is a relationship between these variables, but it is weaker than most of the previous domains.
8.7 DiscussionThis chapter describes the association between T4, IMD 2004 and it’s constituent domains. Policy and policy makers have used various indices when comparing, contrasting and informing the debate that surrounds policy development. In addition, health and health service researches have used this data to shed a light on various aspects of health service provision. The production of the IMD scores not only allows a new model by which deprivation can be analysed but also provides a means by which historical trends can be made. However, for this to be done, a meaningful understanding between the different deprivation indices needs to be had. This chapter starts this comparison and it is for other researches to further the body of evidence relating to when, where and how these new census deprivation indices and domains are best used. The strong associations between income domain, IMD 2004 and T4 scores resonates with views on the census variables. In particular, the importance of particular variables, particularly car ownership has previously been explored (Christie et al 2003, Ebrahim-shah et al 2004) as probably the most important census variable associated with health. The indicator which shows the best correlation between T4 and IMD 2004 is the income domain (Figure 8.02). ReferencesChristie-S-M-L, Fone-D-L. “Does car ownership reflect socio-economic disadvantage in rural areas? : a cross-sectional geographical study in Wales, UK.” Public Health, March 2003, vol. 117, no. 2, p. 112-116. http://www.odpm.gov.uk/index.asp?id=1128440 |
For more information please contact Sarafina
Cotterill
© Department
of Public Health and Epidemiology, University of Birmingham