Does the variation in the socioeconomic characteristics of an area affect mortality?

BMJ 1996; 312 doi: (Published 20 April 1996) Cite this as: BMJ 1996;312:1013
  1. Yoav Ben-Shlomo, lecturer in clinical epidemiologya,
  2. Ian R White, lecturer in medical statisticsb,
  3. Michael Marmot, professor of epidemiology and public healtha
  1. a International Centre for Health and Society, Department of Epidemiology and Public Health, University College London Medical School, London WC1E 6BT
  2. b Medical Statistics Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT
  1. Correspondence to: Dr Ben-Shlomo.
  • Accepted 23 November 1995

Our research in England has shown that the more deprived an area the greater its incidence of premature mortality.1 Wilkinson has argued that in the developed world income distribution is a more important predictor of life expectancy between countries than simply mean income.2 We aimed to determine whether the risk of mortality in a geographical area was related to the degree of socioeconomic variation within that area as well as the average level of deprivation.

Methods and results

For each of the 8464 wards in England we obtained the Townsend deprivation index from the 1981 census1 and directly standardised all cause mortality for 1981-5. Mortality under the age of 65 was used as an indicator of premature mortality. Male and female mortality rates were averaged for each ward. Twenty four wards were excluded because we could not compute the mortality rate, and two local authorities were excluded because each contained only one ward. The remaining 369 local authorities contained an average of 23 wards (6-47).

For each local authority we computed the median of the ward Townsend scores as a measure of overall deprivation and their interquartile range as a measure of variation in deprivation (correlation between the two measures 0.33). We also computed the average of the ward mortality rates. The local authorities were divided according to their quartile of deprivation and variation and the mean mortality for each group computed. We then constructed models in which mortality was regressed on quartile of variation within each quartile of deprivation. Because deprivation still varied between wards within a quartile of deprivation “fully adjusted” analyses also controlled for deprivation as a continuous variable in each model.

Mortality was strongly positively associated with average deprivation (table). The trend for mortality was 26 per 100000 per quartile of deprivation (95% confidence interval 23 to 28, P<0.001). Mortality was also positively associated with variation: the average fully adjusted trend was 7 per 100000 per quartile of variation (4 to 9, P<0.001). Although this effect appeared to be stronger in the middle quartiles of deprivation, the trends did not differ significantly (P=0.09 for heterogeneity).

Mean mortality of local authorities by average deprivation and variation of deprivation. Values are mean mortality per 100000 (and number of local authorities) for each quartile of variation and deprivation and trend in mortality

View this table:

Results were similar using mean and standard deviation, all age mortality, and male and female mortality separately; after ward mortality had been transformed by taking either the square root or the square; and after we had adjusted for the number of wards both as a continuous and a quartile variable.


Our results confirm a strong gradient in mortality related to deprivation, together with a positive association between degree of variation within an area and increased mortality (P<0.001). These results support the hypothesis that variations in income contribute an additional effect on mortality over the effect of deprivation alone.

This analysis cannot show which wards in an area of greater inequality suffer higher mortality: all might, or only the most deprived. Alternatively, increased mortality for poor wards might not be balanced by decreased mortality for rich wards in the same area—that is, the relation may not be linear, but analyses on transformed data did not alter the effect of variation on mortality.

The association between variation and mortality appears to be least in the most affluent and most deprived areas, although the result of a heterogeneity test was not significant. These findings deserve further investigation as some evidence exists that community solidarity may have a beneficial effect on all residents.3

Studies have produced contradictory results on whether area characteristics have a truly independent effect on mortality.4 5 Although our analysis is based on areas, not individuals, it suggests that the characteristics of individuals are insufficient to account fully for differences between areas, as individuals in more variable areas appear to have worse mortality than their counterparts in more homogeneous areas.


  • Funding None.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract