South West Public Health Observatory

Previous Page

 

 

 

 

 

 

PolicyAction Team 18: Better information

WorkingPaper: Measuring deprivation: a reviewof indices in common use

Thispaper has been reproduced withpermission from the Policy Action Team(PAT) 18 Better information website at:
http://www.cabinet-office.gov.uk/seu/2000/pat18/Depindices.htm

ThisWorking Paper was produced to inform,and support the work of, Policy ActionTeam (PAT) 18 on ‘Better Information’.It reviews the most commonly useddeprivation measures and highlights someof the issues surrounding their use. Thepaper presents evidence considered bythe PAT in reaching its conclusions. Itdoes not form part of the mainPAT 18 report on Better Information andis not a statement of Government policy.

Introduction

Thereare a variety of deprivation indicescurrently in use which have beendeveloped to meet a range of differentobjectives. Despite the conceptualdifferences, there has been muchdiscussion of their relative merits.This paper outlines the general conceptof deprivation indices. It then presentsbrief summaries of the most commonlyused measures and highlights some of theissues surrounding their use.

DeprivationIndices

Ingeneral, deprivation indices "measurethe proportion of households in adefined small geographical unit with acombination of circumstances indicatinglow living standards or a high need forservices, or both" (Bartley andBlane 1994) 1 . An importantnote to be made about all ecologicalmeasures of deprivation – that is,measures based on geographic areas,rather than individual circumstances –is that "notall deprived people live in deprivedwards, just as not everybody in a wardranked as deprived are themselvesdeprived" (Townsend et al 1988) 2 .

Thispoint is reiterated by Sloggett andJoshi who note "deprivation indices may begainfully used to identify areas ofrelative concentration of disadvantage,in the absence of data at the personallevel, or where the fact of geographicconcentration is pertinent … butdisadvantaged people also live elsewhereand could be excluded in large numbersif interventions were planned purely onthe basis of a local, census-based,deprivation score" (Sloggettand Joshi, 1994) 3 They go onto note that "formaximum effectiveness, health policyneeds to target people as well asplaces" (Sloggett and Joshi,1994) 4 .

Ininterpreting deprivation scores it isimportant to remember that manydeprivation scores are relative measures - that is, the score for anyone area is standardised by reference tothe mean for the total of all areasincluded in the calculation. Forexample, scores derived for all thewards in one Local Authority cannot becompared to scores derived separatelyfor all the wards in another LocalAuthority, because the scores for eachset of wards are relative to the meanfor the respective Local Authority.

Measuresof Deprivation in Common Use

Thereare many different measures ofdeprivation in common use. In a studycomparing how the use of differentmeasures of deprivation may influenceresource allocation decisions Mackenzieet al (1998) 5 highlightedthat different organisations havepreferences for different measures. Forexample, they state that the TownsendMaterial Deprivation Score is favouredby Health Authorities, whereas LocalAuthorities tend to use the Departmentof the Environment, Transport and theRegions’ Index of Local Deprivation.The Jarman Underprivileged Area Score isused by the Department of Health formaking additional payments to generalpractitioners. The Breadline BritainScore has been used by the media toestimate the percentage of poorhouseholds in a particular area.

TheTownsend Score and the Carstairs Score(which is a very similar measure thatwas developed specifically for theanalysis of Scottish health data) wereboth developed as measures of materialdeprivation 6 . They tend to becommonly used in epidemiologicalanalyses, for example in the recentpublication by the Office for NationalStatistics on Trends in Cancer Survival(Coleman et al, 1999) 7 . TheJarman Underprivileged Area Score wasnot originally constructed to measuredeprivation but as a measure of GeneralPractice workload. The Jarman Score wasderived to take account of geographicvariations in the demand for primarycare based on a survey of GPs’subjective expressions of the socialfactors among their patients that mostaffected their workload. The series ofindices that have been produced by the(former) Department of the Environment(DoE), and (now) the Department of theEnvironment, Transport and the Regions(DETR), were designed as generalmeasures used to identify areas ofgreatest need in order to assist centralgovernment when allocating resources toLocal Authorities.

Detaileddescriptions of measures of deprivationand area classifications in common use

JarmanUnderprivileged Area Score

TheJarman Underprivileged Area (UPA) Scorewas not originally constructed tomeasure deprivation but to measureGeneral Practice workload. The JarmanScore was designed to take account ofgeographic variations in the demand forprimary care based on a survey of GPs’subjective expressions of the socialfactors among their patients that mostaffected their workload. The variant ofthe score in most common use – theUPA8 score – comprises eight variableswhich were individually weighted by asample of London GPs (Carr-Hill, 1988) 8 .

  1. Unemployment – (3.34) residents unemployed as a percentage of economically active;

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

  3. Lone parents – (3.01) % of residents in ‘lone parent’ households;

  4. Under 5s – (4.64) % of residents aged under 5 years;

  5. Elderly living alone – (6.62) % of elderly persons living alone;

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

  7. Low social class – (3.74 ) % of residents where household head is unskilled (social class V);

  8. Residential mobility – (2.68) % of residents who changed address in the previous year.

Eachvariable is based on the percentage ofall residents in households, with theexceptions of unemployment, which isbased on the percentage of theeconomically active population which isunemployed, and residential mobility,where the denominator is the totalresident population. Each variable isfirstly expressed as a proportion(between 0 and 1). The proportions arethen transformed by calculating thesquare root and then finding theequivalent arc sine (asin). Thevariables are expressed as Z scores 9 and multiplied by their respectiveweighting factors. The final score isobtained by summing the variables (afterstatistically reworking). Higher scoresindicate greater levels of GP workload.

Theindex has been criticised as "beingbetter at defining inner-citydeprivation because it includes factorslike overcrowding and ethnicity" (Davies, 1998) 10 . Talbot(1991) 11 has extended thiscriticism by stating that "inparticular, there is a strong biastowards London in the proportion of thepopulation classified as deprived" .He goes on to state that "theindex fails to recognise the nature ofdeprivation in the north ofEngland…benefit[ing] the Thamesregions at the expense of peripheralregions" (Talbot, 1991) 12 .

TownsendMaterial Deprivation Score

TheTownsend Score is based on fourvariables originally taken from the 1981Census that were selected to representmaterial deprivation: unemployment (lackof material resources and insecurity),overcrowding (material livingconditions), lack of owner occupiedaccommodation (a proxy indicator ofwealth) and lack of car ownership (aproxy indicator of income). It is asummation of the standardised scores (Zscores) for each variable (scoresgreater than zero indicate greaterlevels of material deprivation). Two ofthe variables – those relating tounemployment and overcrowding – arefirstly transformed using the logtransformation y = ln (x + 1) to producemore normal distributions. The Townsendscore is considered the best indicatorof material deprivation currentlyavailable. The four variables arecombined together in an overalldeprivation index, with each variablebeing given an equal weight.

Theunits of measurement of the fourindicators are:

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

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

  3. Home ownership – % of private households not owner occupied;

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

Thevariables selected are direct indicators of deprivation - that is,they represent the condition or state ofdeprivation. In contrast, indirect indicators of deprivation represent thevictims of those conditions or states,for example, the elderly, ethnicminorities and single parents. Townsendet al (1988) 13 highlighted that "evenif many among these minorities aredeprived, some are not, and the point isto find out how many are deprived ratherthan operate as is all are in thatcondition. It is the form which theirdeprivation takes and not their statuswhich has to be measured" .

Townsendscores can be recalculated using theequivalent variables extracted for areasfrom the 1991 Census. However, it shouldbe noted that the change in variablescores for any one area between 1981 and1991 cannot be taken as indicative ofreducing or increasing relativedeprivation - primarily because ofexogenous changes in the socialcharacteristics of car and homeownership. Hence, explanatory modelscalibrated using 1981 Townsend scoresshould not be applied to 1991 scoreswithout recalibration.

Carstairsand Morris Scottish Deprivation Score

Thisindex was constructed by Carstairs andMorris for the analysis of Scottishhealth data. Like the Townsend Score, itis based on four variables originallytaken from the 1981 Census which werejudged to represent, or be determinantsof, material disadvantage. Three of theindicators are the same as those used inTownsend, the fourth - social class - isused in place of housing tenure. Theauthors state "wedo not find Townsend’s reasons forexcluding social class convincing sincewe believe that being in a low socialclass, equally with being unemployed,places families in a position of pooraccess to material resources…" (Carstairs and Morris, 1989) 14 .The authors considered housing tenure tobe "lessrelevant in Scotland as a much higherproportion of housing stock is in thepublic sector and the variable would nothave acted as a discriminator betweenlarge sections of the population" (Morris and Carstairs, 1991) 15 .

Theunits of measurement of the fourindicators are:

  1. 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;

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

  3. Social Class IV or V – proportion of all persons in private households with head of household in Social Class IV or V;

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

Thedeprivation measure is an unweightedcombination of the four standardisedvariables.

MATDEPand SOCDEP

MATDEP(a material deprivation index) andSOCDEP (a social deprivation index) areboth indices of deprivation that weredeveloped by Forrest and Gordon (1993) 16 following the 1991 Census. Thedistinction between material and socialdeprivation has been explicitly statedby Townsend - "Material deprivationentails the lack of goods, services,resources, amenities and physicalenvironment which are customary, or aleast widely approved in the societyunder consideration. Social deprivation,on the other hand, is non-participationin the roles, relationships, customs,functions, rights and responsibilitiesimplied by member of a society and itssub-groups. Such deprivation may beattributed to the affects of racism,sexism and ageism …" (Townsendet al, 1998) 17 .

Thedistinction between material and socialdeprivation has two conceptual forms - "theargument between the use of direct andindirect measures and the differentdimensions of deprivation when taking asocial (roles and relationships) and amaterial (goods and services)perspective" (Lee et al, 1995) 18 .

Indicatorsused in MATDEP:

  1. Overcrowding : % households with more than 1 person per room;

  2. Lack amenity : % households lacking or sharing use of a bath/shower and/or inside WC;

  3. No central heating : % households with no central heating;

  4. No car : % households with no access to a car.

Indicatorsused in SOCDEP:

  1. Unemployment : % economically active population unemployed;

  2. Youth unemployed : % economically active 16-24 year olds unemployed;

  3. Lone parents : lone parent households as a proportion of all households;

  4. Elderly : % households containing a single pensioner;

  5. Long-term illness : % households containing a person with limiting long-term illness;

  6. Dependent only : households containing dependants only (e.g. single pensioners with long-term illness) as a percentage of all households.

MATDEPand SOCDEP scores are the summation ofthe unweighted standardised scores foreach variable. Each variable isstandardised by dividing the percentagevalue for each indicator in a particulargeographic area by the maximum value foreach indicator in all areas to give avalue between 0 and 1. This means thatthe maximum score for SOCDEP is 6 andthe maximum score for MATDEP is 4 (theminimum score for both indices is 0).Higher scores indicate greater levels ofdeprivation.

Departmentof the Environment’s Index of LocalConditions (1991)

TheIndex of Local Conditions (ILC)comprises 13 variables - seven of whichare Census variables and six of whichare non-Census variables:

Censusvariables:

  1. Unemployment – residents aged 16 and over who are unemployed compared to residents aged 16 and over who are economically active;

  2. Children in low-earner households – resident dependent children (aged under 16) living in households with no one in employment or with a single parent in part-time work compared to all dependent children (aged under 16);

  3. Overcrowding – households with above one person per room compared to all households;

  4. Housing lacking basic amenities – residents in households lacking amenities compared to all residents in households in permanent and non-permanent accommodation;

  5. No car – households without a car compared to all households;

  6. Children in unsuitable accommodation – dependent children (aged under 16) living accommodation that is either purpose-built flats, non-permanent, or unshared dwellings either converted or not self-contained compared to all dependent children (aged under 16);

  7. Educational participation – residents aged 17 years not in full-time compared to all residents aged 17 years.

Non-censusvariables (Sources and dates) 19

  1. Long-term unemployment – the ratio of long-term unemployment (more than one year) to total unemployment (DfEE 1991);

  2. Income support – persons or households in receipt of income support expressed in relation to total adult population (DSS 1991);

  3. Low educational attainment – passes in GCSE exams at grade D or below in relation to the total number of passes (DfEE 1991);

  4. Standardised mortality ratios – all cause SMRs expressed in relation to a value of 100 for Great Britain (1991);

  5. Derelict land – area of land defined as derelict in DoE survey in relation to total land area (DoE 1988);

  6. Home insurance weightings – weightings for premiums on house contents insurance used by three national insurance companies (1991).

Theindex of local conditions is anunweighted summation of the selectedindicators using their log-transformedsigned chi-square values 20 .The actual number of persons having eachselected variable are compared to thenumbers that would be expected ifaverage English rates applied. Thedifference between the actual andexpected numbers is squared and thendivided by the expected number afterwhich the value of 1 is added. A logtransformation is then applied and thosescores where the actual rate was belowthe expected rate are given negativesigns. Summed scores greater than zeroindicate greater levels of materialdeprivation.

Thisindex differs from those previouslydescribed in using actual numbers ratherthan percentage rates as the input intothe calculations. This has the effect ofgiving lower weights to those areaswhere the actual counts are small - andhence statistically less reliable (i.e.an area where 3 out of 10 persons areunemployed will have a lower score thanone where unemployment is 30 out of100).

Departmentof the Environment, Transport and theRegions’ Index of Local Deprivation(1998)

InJune 1998, following consultation, theDepartment of the Environment, Transportand the Regions (DETR) published anupdated version of the 1991 ILC. The1998 Index of Local Deprivation (ILD),based mainly on data for 1996, wascalculated for all 354 Local AuthorityDistricts as they stood at April 1998.The ward and ED level indexes are basedon the 1991 Census Area definitions.There are 12 indicators in the districtlevel ILD which relate to differentdimensions of deprivation – income,health, education, environment, crimeand housing. The indicators, theirmeasures and their sources are:

  1. Unemployment – persons unemployed compared persons economically active (ONS claimant count 1997);

  2. Dependent children of income support recipients – compared to persons aged 16 and under (DSS 1996);

  3. Overcrowding – households above one person per room compared to all households (1991 Census);

  4. Housing lacking basic amenities – residents in households lacking amenities compared to all residents in households in permanent and non-permanent accommodation (1991 Census);

  5. Non income support recipients in receipt of council tax benefit 22 – compared to total population aged 18 and over (DSS 1996);

  6. Educational participation – 17 year-olds not in full-time education compared to all 17 year-olds (1991 Census);

  1. Long-term unemployment – the ratio of long-term unemployment (more than one year) to total unemployment (ONS claimant count 1997);

  2. Income support – persons or households in receipt of income support expressed in relation to total adult population (DSS 1996);

  3. Low educational attainment –fifteen year olds with no GCSE passes or gaining GCSE passes at grades D-G only compared to all 15 year olds (DfEE 1996);

  4. Standardised mortality ratios – all cause SMRs for under 75 year olds (ONS 1996);

  5. Derelict land – area of land defined as derelict in DoE survey in relation to total land area (DoE 1993);

  6. Home insurance weightings – weightings for premiums on house contents insurance used by three national insurance companies (1996).

Thereare two main differences in themethodology between the 1991 and 1998Indexes. Firstly, in the 1991 Index thevalues for the indicators were simplyadded together, whereas in the 1998Index only the positive values (thosewhere the actual count exceeded thatexpected) have been added together toproduce the overall index score.Secondly, in the 1991 Index noweightings were attached to any of theindicators. However, in the 1998 Indexthe values for the standardisedmortality ratio and insurance premiumindicators

"havebeen multiplied by two to give them asimilar level of influence in theoverall index" (DETR, 1998) 23 .

1999Review of the Department of theEnvironment, Transport and theRegions’ Index of Local Deprivation

Althoughthe 1998 update of the Index of LocalDeprivation was not intended to revieweither the methodology or thecombination of indicators used, theconsultation process highlighted anumber of issues relating to theseaspects. Moreover, since the last reviewwas undertaken many new sources ofsub-district level data have becomeavailable. In the light of this, DETRdecided to undertake this further reviewof the Index to look at:

·          the conceptual basis of the ILDas a mechanism for identifying the mostdeprived areas in England;

·          the current indicators, to assesswhether they represent the best,up-to-date measure of relative generaldeprivation and, if not, test andrecommend robust alternatives;

·          the methodology used to combinethe individual indicators into singleindexes of general deprivation at thedifferent spatial levels and, ifappropriate, suggest improvements basedon assessment of alternative methods.

TheIndex will be revised and updated on thebasis of the recommendations from thesereviews, in agreement with the DETR andfollowing extensive consultation withusers in central and local governmentand other relevant organisations.

Atthe end of the review users will be senta short summary of the revised andupdated Index and guidelines on how touse and interpret it. The full Indexwill also be available, and DETR willpublish a full report of the processincorporating a description of patternsof deprivation and how these differ fromthe 1998 ILD (Noble et al . 1999) 24 .

BreadlineBritain Score

TheBreadline Britain Score is the result oftwo surveys carried out by MORI forLondon Weekend Television and the JosephRowntree Foundation in 1983 and 1990.The 1983 study pioneered the use of the‘consensual’ or ‘perceived’deprivation approach to measuringpoverty. The approach set out todetermine whether there are some peoplewhose standard of living is below theminimum acceptable to society. Theminimum standard of living wasdetermined by interviewing a quotasample (based on age, sex and workingstatus) of 1,174 adults in 1983 and1,831 adults in 1990. Aggregated datawere weighted by age, household type,household tenure and ACORN housing type(see below) to be representative of thepopulation of Great Britain. In order toensure a large sample of people livingin deprived areas over-sampling wasconducted in ACORN areas known tocontain poor households.

Inthe 1990 Survey respondents werepresented with a set of 44 cards ontoeach of which was written the name of adifferent item covering a range ofpossessions and activities that relateto standards of living. For example, atelevision, a night out once a fortnightand a warm waterproof coat. Respondentswere asked to place the 44 cards intoone of two boxes. Box A was for itemswhich they considered necessary; thoseitems which all adults should be able toafford and which they should not have todo without. Box B was for items whichthey considered to be desirable but notnecessary. They were also asked if theyfelt differently about any of the itemsin the case of families with children.An item was deemed to be a sociallyperceived necessity if more than 50 percent of respondents put it into Box A.Later in the interview the respondentswere asked to assign one of thefollowing 5 options to each of the 44items:

  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 adeprivation index score each time theyanswered ‘don’t have and can’tafford’ to an item that was consideredto be a necessity by more than 50 percent of respondents (Gordon andPantazis, 1997) 25 .

LowIncome Scheme Index (LISI - a deprivation scale based onprescribing in general practice)

Theindices listed so far are calculated forarea populations; their values may notbe reliable when attributed toregistered lists of GeneralPractitioners, if the personsregistering with a particular practicedo not represent an unbiased sample ofthe population of their area ofresidence. An alternative direct measureof practice list deprivation has beenderived from prescription data. Morethan 80 per cent of items dispensed fromprescriptions issued by NHS GPs inEngland are exempt from prescriptioncharges. Most exemption (54.6 per cent)is based on age. An additional 6.6 percent of items are accounted for byfamily health service authority (FHSA)exemptions – mainly related topregnancy. However, 12.1 per cent ofitems are exempt from payment under thelow income scheme (by ticking boxes H,I, J, K, L on the back of the FP10Cform). The low income scheme coversrecipients (and their dependants) ofFamily Credit and Income Support andothers who qualify on the grounds of lowincome. Two types of Low Income SchemeIndex (LISI) have been calculated, onebased on the number of prescriptionitems and the other on the netingredient cost (Lloyd et al, 1995) 26 .

Exemptioninformation is not, however, routinelycollected where patients have theirdrugs supplied by a practice dispensary- and hence LISI scores are notcalculated for practice lists where morethan 30 per cent of the list haveapplied for practice dispensing. Thenational file of LISI scores wascirculated to all Health Authorities aspart of the Attribution Data Set (NHSE,1998) 27 .

AreaClassifications

Inaddition to the measures of deprivationthat have been discussed above there arevarious area classifications which "cluster[geographic distance] places together onthe basis of various socio-economiccommonalities" (Burrows andRhodes, 1998) 28 .

TheONS classification of local and healthauthorities of Great Britain (1991)

Thisclassification, often referred to as theOffice for National Statistics (ONS)Area Classification, provides a generalpurpose summary indicator of thecharacteristics of each Local AuthorityDistrict and Health Authority in GreatBritain. Using information collected atthe 1991 Census it classifies eachauthority into one of 6 Families, 12Groups and 34 Clusters (HealthAuthorities are only categorised intoFamilies and Groups) on the basis of 37separate socio-demographic variables.The general choice of variables for theclassification was determined byinclusion of:

·          variables to represent the maindimensions of the 1991 Census data(demographic; employment andsocio-economic; household compositionand housing) bearing in mind thegeneral-purpose use of theclassification;

·          variables used in one or both ofthe 1971 or 1981 Office of PopulationCensuses and Surveys (OPCS) areaclassifications, or nearest equivalent variables, or a replacement which improves theclassification process;

·          variables available for the firsttime from a Census, such as ethnic groupor limiting long-term illness, withpotential significance for areaclassification.

Theclassification is based on LocalAuthorities as they existed on CensusDay in 1991 and on Health Authorities at1 April 1994.

TheONS Area Classification is not actuallyan index of deprivation but an indicatorof "socio-economicsimilarity and difference betweenareas" (Wallace and Denham,1996) 29 . However, the term‘deprivation’ is used as a generaldescriptive term to refer to the valuesfor a group of five variables which formpart of the classification and whichcharacterise generally poorsocio-economic circumstances. Thesevariables, and their units ofmeasurement, are: the standardised rateof limiting long-term illness; thepercentage of children with a singleparent; the percentage of dependantswith a lone carer; the unemploymentrate; and the percentage of householdswithout a car. Authorities with highvalues for these variables are deemed tobe the more socio-economically deprived.

TheONS Area Classification is usedextensively for resource allocation andperformance management purposes by theNHS Executive.

TheONS classification of local and healthauthorities of Great Britain: revisedfor authorities in 1999

ONShave recalculated the classification toLocal Authority boundaries as at 1 April1998 and Health Authority boundaries asat 1 April 1999 reflecting thepost-reorganisation areas for both Localand Health Authorities 30 .(The original 1991 classification wasalso released at ward level using thesame methodology and census variables.However, this was a separateclassification and has not beenrevised).

Therevised version is not a newclassification scheme. It has beencompiled using the same methodology andapproach as the original 1991 version,and the same 37 component variables fromthe 1991 Census. None of the data havebeen updated. ONS have simply attemptedto recast the classification to reflectthe new Local and Health Authorityareas. The revision has, however,produced a slightly differentclassification solution because of theeffect of the boundary changes. The‘clusters’ of similar authoritiesand the average characteristics forthese clusters have necessarily changedas the Local and Health Authorities havechanged. There is therefore littledirect comparability with the results ofthe original classification. However,ONS have, where possible and whereappropriate, retained some of the samenames for the Families, Groups andClusters within the classification.

Geo-DemographicClassifications

Geo-demographicclassifications are not measures ofdeprivation but they are usedextensively for marketing purposes totarget customer groups. The assumptionis made that those households within adefined neighbourhood are likely to havesimilar lifestyles and buying habits.Like area classifications, they groupgeographically disparate places togetheron the basis of certain characteristicsthat distinguish customer groups.Traditional forms of socialclassification used for market researchwere based on the occupation of the headof the household in which they werecategorised. Geo-demographics takes thisconcept further, and can be expressedas, the classification of people by theneighbourhood in which they live and byhaving a similar pattern in terms oftheir ability to consume, behave orpurchase.

ACORN– A Classification of ResidentialNeighbourhoods

ACORNis a geo-demographic classificationdeveloped by the market analysis companyCACI. The classification is builtentirely using Census data and includesinformation on age, sex, marital status,occupation, economic position,education, home ownership and carownership. Like the ONS AreaClassification, ACORN is based on athree-tier system which classifiespeople living in Great Britain into oneof 6 Categories, 17 Groups or 54 Types(plus one ‘unclassified’ in eachcase). Although not a true measure ofdeprivation there are Groups and Typeswhich relate to areas which have suchcharacteristics as high unemployment, ahigh percentage of elderly with healthproblems, a high percentage of loneparents and severe overcrowding.

SuperProfiles

SuperProfiles is another geo-demographicsystem, similar to ACORN, available fromthe market analysis company CLARITAS. Itcategorises households in Great Britainaccording to the characteristics of theneighbourhood in which they are located.The classification is derived fromnumerous data sources including the 1991Census, the electoral roll, creditinformation and market research data.The classification is again based on athree-tier system. At the most detailedlevel there are 160 different SuperProfiles Clusters which are ranked inorder of affluence. Each Cluster relatesto one of 40 Market Groups, which inturn relate to one of 10 Lifestyles.

Informationprovided by: Julia Bunting, Departmentof Health

 

Notesand References

 

1

Bartley, M. and Blane, D. (1994) ‘Appropriateness of deprivation indices must be ensured’ British Medical Journal 309: 1479.

2

Townsend, P., Phillimore, P. and Beattie, A. (1988) Health and Deprivation: Inequality and the North . Routledge, London.

3

Sloggett, A. and Joshi, H. (1994) ‘Higher mortality in deprived areas: community or personal disadvantage?’ British Medical Journal 309: 1470-4.

4

See note 3.

5

Mackenzie, I.F., Nelder, R. Maconachie, M. and Radford, G. (1998) ‘My ward is more deprived than yours’. Journal of Public Health Medicine 20: 186-90.

6

Material deprivation is the lack of goods, services, resources, amenities and physical environment. This is in contrast to social deprivation which is non-participation in the roles, relationships, customs, functions, rights and responsibilities implied by member of a society and its sub-groups. Such deprivation may be attributed to the affects of racism, sexism and ageism etc (see Note 2).

7

Coleman, M. et al (1999) Cancer Survival Trends in England and Wales 1971-1995: deprivation and NHS Region . The Stationery Office. London.

8

Carr-Hill, R. (1988) ‘Revising the RAWP formula: Indexing deprivation and modelling demand’ Centre for Health Economics, University of York. Discussion Paper 41.

9

Z scores express each variable in terms of its mean value in the population and its standard error. If this were not done then variables with longer scales would have more weight than variables with shorter scales in the overall score. For example, the number of children in a household could vary from 0 to 10, while the number of cars could range from 0 to 3. Simply adding these together would given children more weight than cars – standardisation is intended to avoid this problem (Bartley and Blane, 1994).

10

Davies, J. (1998) ‘Healthy Living Centres’ Health Services Journal 5 November pp1-5.

11

Talbot, R.J. (1991) ‘Underprivileged areas and health care planning: implications of use of Jarman indicators of urban deprivation’. British Medical Journal 302: 383-6.

12

See Note 11.

13

See Note 2.

14

Carstairs, V. and Morris, R. (1989) ‘Deprivation and mortality: an alternative to social class?’ Community Medicine 11: 210-9.

15

Morris, R. and Carstairs, V. (1991) ‘Which deprivation? A comparison of deprivation indexes’ Journal of Public Health Medicine 13: 318-26.

16

Forrest, R. and Gordon, D. (1993) People and Places: a 1991 Census atlas of England . SAUS. University of Bristol.

17

See Note 2.

18

Lee, P., Murie, A. and Gordon, D. (1995) Area measures of deprivation: a study of current methods and best practices in the identification of poor areas in Great Britain . Centre for Urban and Regional Studies. University of Birmingham.

19

At the electoral ward level only the census-based variables are used in the calculation of the index.

20

The standardisation and transformation ‘has the merits of: taking account of the small size of the denominators of many of the observations; using an interpretable value of zero; and using values which approximate the normal curve’ (see Note 21).

21

Department of the Environment (1995) 1991 Deprivation index: a review of approaches and a matrix of results . HMSO. London.

22

The 1998 indicator on ‘non income support recipients in receipt of council tax benefit’ replaces the 1991 indicator ‘no car’ which was used as an income proxy. The 1991 indicator ‘children in unsuitable accommodation’ was dropped from the 1998 classification.

23

Department of the Environment, Transport and the Regions (1998) 1998 Index of Local Deprivation: A summary of results . DETR. London.

24

Noble, M. et al (1999) Index of Deprivation 1999 Review . Newsletter 1 – January 1999.

25

Gordon, D. and Pantazis, C. (eds) (1997) Breadline Britain in the 1990s . Ashgate Publishing Limited. England.

26

Lloyd, D., Harris, C. and Clucas, D. (1995) ‘Low income scheme index: a new deprivation scale based on prescribing in general practice’. British Medical Journal 310: 165-70.

27

NHSE (1998) Using the Attribution Data Set.

28

Burrows, R. and Rhodes, D. (1998) Unpopular places?: Area disadvantage and the geography of misery in England . The Policy Press. England.

29

Wallace, M. and Denham, C. (1996) The ONS classification of local and health authorities of Great Britain . Studies in Medical and Population Subjects No 59. HMSO, London.

30

Office for National Statistics (1999) The ONS classification of local and health authorities of Great Britain: revised for authorities in 1999 . Studies in Medical and Population Subjects No 63. ONS, London.

The Public Health Observatory is part of the South West Observatory, a wider Regional intelligence function, currently supported by the South West Regional Assembly the Department of Health, Government Office South West, the South West of England Regional Development Agency and the Environment Agency.