Close window to return

Issue 2, July 1998, pp.8-9.

Explaining geographical inequalities in health
Mary Shaw, Danny Dorling and Nichola Brimblecombe

The social and spatial polarisation of life chances
There have been dramatic reductions in death rates in Britain over the past century, with an accompanying rise in life expectancy. However, this gain has not been equally distributed throughout the population. Not only are there persisting differences in mortality between social classes, but the gap has been widening. For example, for men, the difference in life expectancy between those in social classes I and II and those in classes IV and V is now over 5 years, and for women this gap is over 3 years.

Alongside this social polarisation of life chances, there has been a spatial polarisation of life chances. Regional inequalities in mortality have been reported since data were first available in the last century. It is consistently found that mortality rates are highest in the north and in Scotland and lower in the south: Britain can effectively be divided into two zones of high and low mortality. As well as these regional differences in mortality, there are differences between places at smaller geographical levels: at local government district and, particularly, at electoral ward level. Moreover, these differences between places are becoming greater. In the most recent period for which data are available (1990-92), the ten percent of the population living in areas of Britain with the highest death rates have the worst ever recorded relative mortality rates for a single set of districts in Britain, with a standardised mortality rate (SMR) of 142.3 for people aged under 65. Since 1981, the SMR of this group has increased by 7.4 percentage points.

Housing wealth and community health
The social polarisation which has occurred in Britain over the past two decades - the well-documented increasing gap between the rich and the poor - can only partially account for this spatial polarisation. The 'Housing wealth and community health' project is investigating, on various geographical levels, the role played by migration and other factors in producing this polarisation.

One of the aims of the project is to pay particular attention to the possible role of changes in the housing market in the polarisation of health. Changes in access to owner-occupied and local authority accommodation may have contributed to the spatial polarisation of mortality in the 1980s. This was a decade in which the supply of council housing declined sharply, with the result that medical factors may have become a more important criterion for the allocation of council housing. Poor health may also have acted as a barrier to owner-occupation. As housing choices become more dependent on both health and wealth, it is possible that poorer people in worse health are more likely than before to move to ( or not leave) already high-mortality areas and regions, while richer people in better health move to lower-mortality areas and regions. Such a process would contribute to the geographical polarisation of health. We are currently collecting evidence to support this hypothesis, by using a variety of different sources of secondary data (ONS mortality data, Census data, Building Society wealth data, the British Household Panel Survey and the Longitudinal Study) as well as local case studies.

High mortality rates among the homeless and vulnerably housed
Our local case studies focus on Brighton and Oxford. These studies have suggested that pockets of high mortality in areas which are relatively affluent and which have close to the national rate of mortality might be largely explained by one of the outcomes of the changes to the housing market in the 1980s and 1990s: increasing homelessness. In Brighton, for instance, the electoral wards with the highest mortality are those with relatively large numbers of people sleeping rough and with many people who are vulnerably housed, a group that includes those living in temporary bed and breakfast accommodation, in poor quality (but expensive) bedsits or in hostels for the homeless. In Oxford, if the deaths of those men whose residence at time of death was recorded as a hostel for the homeless are taken into account, then the distribution of elevated ward mortality is dramatically reduced (see Figure 1).

Figure 1: SMRs for men by ward with and without hostel deaths




Analysis of data collected by the charity Crisis on deaths of homeless men in London which were recorded as 'No Fixed Abode' also suggest that rough sleepers have significantly elevated mortality rates compared to the general population. Table 1 compares the death rates for these two groups. The standardised mortality ratio for male rough sleepers for the age group 16-64 is 2587.

The death rate for male rough sleepers aged 16 to 64 is 26 times higher than the death rate for the general population.

Table 1: SMRs for male rough sleepers in London

Moreover, in both Brighton and Oxford, there is evidence that the homeless and vulnerably housed are more likely than other people to have migrated from other areas. This suggests that the differential migration of those who are most socially excluded, in combination with various locally-specific cultural, social and economic factors, produces a spatial polarisation of mortality. Similarly, our initial analysis of the British Household Panel Survey (BHPS) has found that the lifetime selective migration of the affluent also tends to move them towards areas that subsequently experience much lower rates of mortality. Our current research is considering the extent to which the geographical concentration of the vulnerably housed in particular wards across Britain explains part of the polarisation on a national scale. We are also examining the extent to which lifetime migration in general produces the geography of mortality seen across Britain today.