Deprivation indices: their interpretation and use in relation to health.

STUDY OBJECTIVE: To examine the use of deprivation indices in relation to health. DESIGN: This paper reviews selected publications which illustrate the diversity of use of deprivation indices in the past decade. Most of this work is based in the major routine databases which exist in this country: the census, population, mortality, cancer register, and health service records all now incorporate a postcode identifier which permits the derivation of data at small area level, and thus the examination of health events in relation to the characteristics of that area - usually ward or postcode sector. The small area approach provides a valuable tool both in deprivation and in other epidemiological studies which examine the influence of the environment on health. SETTING: The setting is various journals and official publications. MAIN RESULTS: The link between deprivation and health has been clearly demonstrated in a number of studies, with populations living in deprived areas exhibiting levels of mortality, particularly below the age of 65, which vastly exceed those in affluent areas. In the decade 1981-91, these differentials increased in Scotland and the Northern Health Region and inequalities in health are shown to have widened. Analysis shows that particular causes of death and sites of cancer are more likely to reflect the influence of socio-economic factors. The work so far mostly shows the associations between these factors and health measures and more investigation is required into the determinants of health, which are likely to reside as much in past as in current circumstances. A measure of deprivation has proved of value in excluding the likely variation in the incidence of disease in studies directed towards determining the influence of the physical environment on populations living in the vicinity of possible harmful industrial processes. A deprivation measure has been adopted by the Department of Health as a basis for making enhanced payments to general practitioners for patients living in these areas, but the resource allocation formula for allocating funds to regional authorities has failed to incorporate such a measure in the formula. CONCLUSIONS: An area measure of deprivation has proved a valuable tool in examining differentials in health and death and is likely to prove of continuing value to health authorities in planning the delivery of health care. Future work should strive to examine the determinants of health as well as the associations, although this is unlikely to be possible through the routine databases which have provided the main basis for analysis so far.

Avoidable deaths in countries of the European community and in Scotland.

A second edition of the European Community Atlas of Avoidable Death, 1980-84, has recently been published. The data in relation to Scotland are reviewed in detail.

Which deprivation? A comparison of selected deprivation indexes.

A number of indexes of deprivation which have been devised or adopted for use by the health services are examined in relation to their performance in explaining the variation observed in a range of health measures, using data for postcode sectors in Scotland. The Scottish deprivation score and the Townsend index are found to explain most variation, and to adhere most closely to the concept of material disadvantage. The Jarman score is less effective as a result of the inclusion of individual variables which are seen to correlate very weakly, even negatively, with the health indicators. The use of these measures in relation to resource allocation and the new GP contract is discussed.

Association of age and social class with suicide among men in Great Britain.

STUDY OBJECTIVE: The aim was to investigate suicide and "undetermined" deaths by age, economic activity status, and social class in Great Britain among males of working age. DESIGN: The study was a cross sectional analysis of Registrar General's data for England and Wales around 1981, repeated for around 1971, and for Scotland around 1971 and 1981. MEASUREMENTS AND MAIN RESULTS: For England and Wales around 1971, suicide and undetermined death rates showed a progressive increase with age and a markedly higher rate in the lower social classes. A significant interaction effect was identified in the central age groups of the lower occupational categories. This interaction was confirmed in the remaining three data sets, notwithstanding some differences in the profile of age specific mortality. Other findings included a higher standardised mortality ratio for the economically inactive, who also showed an earlier peak in age specific mortality, and a relative concentration of undetermined as compared to suicide deaths in the lower social classes, but not all these further results were fully replicated. CONCLUSIONS: There is a concentration of suicide and undetermined deaths in the middle age groups of the lower social classes. Plausible explanations include both the social drift and the social genesis hypotheses, the latter including the effects of long term unemployment.

Deprivation and health in Scotland.

A measure of deprivation has been devised using data from the 1981 Census about area characteristics which may be considered to reflect material deprivation, and a score is calculated for each postcode sector in Scotland. The routine data sources provide information for a range of health measures which are shown to be strongly linked to levels of deprivation, with specific causes--respiratory conditions, cancer of the lung and cancer of the cervix--exhibiting very considerable differentials in respect of both mortality and cancer morbidity. Nevertheless not all causes exhibit this positive association. This area-based measure provides a valuable tool for the analysis of health data in relation to socio-economic circumstances which overcomes many of the limitations which arise in the use of the more traditional measure of social class. Like social class however it depends on data from the census which are available only at ten-yearly intervals. Other existing data sources are suggested which could possibly provide relevant information on area characteristics and populations for inter-censal years.

Deprivation: explaining differences in mortality between Scotland and England and Wales.

To detect reasons for the difference in mortality between Scotland and England and Wales a measure of deprivation was studied, comprising overcrowding, unemployment of men, low social class, and not having a car. Data for Scotland for 1980-2 showed this measure to be strongly associated with mortality, with gradients being particularly steep in young adults. Deprivation was much severe in Scotland than in England and Wales. These findings suggest that much excess mortality may be ascribed to more adverse conditions. Standardising the mortality ratios to take account of the relative affluence and deprivation of the two populations led to the differentials observed being radically adjusted, while standardising for social class had little effect. Deprivation measures based on areas overcome many of the limitations associated with social class analysis and also show much greater discrimination between populations. Measures of deprivation apparently provide a powerful basis for explanation of health differences. Such measures should therefore form part of the 1991 census output to facilitate their use on a consistent basis.

Deprivation and mortality: an alternative to social class?

Mortality rates for males aged 20 to 64 in Scotland (1980-1982) display an increasing gradient both by social class (from class 1 to 5) and by the deprivation category of the area (from affluent to deprived) in which the event occurred. Social classes also exhibit gradients in mortality across deprivation categories, the corresponding rates in the most deprived category being around twice those in the most affluent areas. The gradients by deprivation category remain after standardizing for differences in social class composition. Within health boards in Scotland populations show only minor variability in terms of social class, while their composition on the dimension of deprivation is markedly diverse, and the classification of populations by the deprivation category of their area of residence appears to offer a superior basis for the explanation of differences in mortality between health boards than does social class. This evidence of area effects supports a proposal for area socioeconomic characteristics to be adopted as key variables in epidemiological analysis, and for an area classification to be designed and incorporated in the 1991 census output for use on a consistent basis.

[Geographical variations in mortality due to diseases amenable to medical intervention in Europe--Commission of European Communities: atlas of avoidable deaths]. / Variations géographiques de la mortalité due à des maladies justiciables d'une intervention médicale en Europe--Commission des Communautés Européennes: atlas des décès évitables.

The urgent need to develop measures of the outcome of health-care services has led to the collaboration of 10 countries of the European Community in the production of the European Community atlas of avoidable death (1974-1978). Seventeen disease groups were chosen for which it was considered that death within specified age groups should be either wholly or substantially avoidable when appropriate medical care is sought and provided in good time. Mortality from these causes was compared across 360 health-service administrative areas in the participating countries. For all diseases there was considerable variation in mortality both within and between the countries of the European Community and it is suggested that high levels of mortality from these causes should be viewed by health authorities as warning signals of potential failures of health-care services. Work is in progress on a further edition of the Atlas for the years 1980-1984. Changes in avoidable mortality over time could indicate which health authorities have persistent problems and which authorities are succeeding in reducing avoidable mortality.

Spina bifida and anencephaly in Scotland.

Data obtained from routine sources showed that from 1971 to 1982 the birth prevalences of spina bifida and anencephaly in Scotland fell. When known terminations after routine alpha fetoprotein screening were added to total births the adjusted birth prevalence could be calculated. In 1974-82 this fell by 40% for spina bifida (3 X 0-1 X 8) and 36% for anencephaly (2 X 2-1 X 4). These findings were compared with data on birth prevalences in England and Wales, Northern Ireland, and Glasgow. The fall in birth prevalences of spina bifida and anencephaly over the past decade appears to have been due both to a true fall in incidence as well as to increased screening and termination for these conditions.