Social Policy Research 126 - August 1997
Changing mortality ratios in local areas of Britain 1950s-1990s

Britain became a signatory to the World Health Organization's Targets for Health in 1985. The primary target commitment was to reduce inequalities in health, measured by mortality differences, by 25% by the year 2000, including reducing geographical variations. The most detailed study to date of mortality trends by area in Britain, by Daniel Dorling at Bristol University, shows that inequalities in health have been rising since 1985, making it extremely unlikely that this international commitment will be met. The research found that:

• Although absolute mortality rates for all groups in society have fallen steadily since the 1950s, the gap between people living in different areas has widened, particularly since the 1980s.
• People under 65 living in the worst tenth of areas were almost twice as likely to die in 1990-92 than people living in the tenth of areas with the lowest mortality rates. This is the highest inequality ever recorded and compares to a difference of one-and-a-half times as likely 20 years ago.
• Those living in the worst tenth of areas in terms of mortality in the early 1990s are in those areas where indicators of poverty are also the greatest.
• More people in Britain live in areas of high relative mortality now than have at any time since the early 1950s - 1 in 12 now live in areas where more than 15% of deaths are above expected levels, after standardising for age differences.
• Men aged between 15 and 44 are the group who have experienced the least improvement in their life chances in recent years, but there is wide area variation to these changes over time.
• Nationally there were 77,000 excess deaths in the 1990-92 period, representing about 4% of all mortality.

Variation between areas

Where a child is born in Britain today is more important than ever in determining that child's life chances, particularly his or her chances of survival. Such a finding is the opposite to common preconceptions of relatively even life chances across the country. This inequality has arisen partly because absolute mortality rates have fallen steadily over the years, but mainly because those improvements have not been distributed evenly amongst the population as a whole. For instance, an infant girl in Leeds is now more than twice as likely to die in the first year of life as is an infant girl growing up in a town in Dorset. Proportionately, eight times as many boys aged between 1 and 4 years died in Manchester as compared to rural Gloucestershire between 1990 and 1992. There has been a 50% increase in the absolute death rate of boys aged between 5 and 14 in Salford since 1981, similarly with adult male mortality in Hammersmith. These are not the most extreme discrepancies but they are places in which the long-term trend is worsening. We are becoming less equal in death.

Using comparable 1951 local authority boundaries, the three areas with the highest mortality rates in the 1990s (Oldham, Salford and Greenock) had mortality ratios only a fifth higher than the national average in the early 1950s. Their rates are now rising towards being a third higher than the national rate. Almost a thousand deaths a year would be avoided, were the mortality rates not excessive in just these three places. 'Excess deaths' are calculated as deaths above the average rate, not deaths above the lowest rate, which would have produced far higher numbers of 'avoidable' mortalities and a much more dramatic summary. Nationally there were 77,000 excess deaths in the 1990-1992 period, representing about 4% of all mortality. Because mortality statistics require census statistics, this is the latest date for which reliable local estimates can be made for small areas.

Increasing health inequalities

Nationally, mortality rates are higher in the north and in urban areas. In recent years, a person living in Glasgow was 66% more likely to die in any given year than someone living in the districts of rural Dorset and 31% more likely than a resident of Bristol. At the end of the 1960s the excess chance of dying in Glasgow, relative to these two places, was much lower, at 42% and 21% respectively. Crude death rates have fallen for most people in places like Glasgow, but not as quickly as they have in cities further south or in rural areas (where they were also lower to begin with). There has been a divergence in mortality along geographical as well as social lines.

To produce a fair measure of geographical inequality in mortality all parts of Britain need to be included. This is done by dividing the population into ten groups of equal size (deciles), the worst group living in the places with the worst mortality records for people under the age of 65 and the best group living in the places that contain the tenth of the population with the best mortality records at any one time. Historical records do not provide enough detail to look at variations in mortality over the age of 65 (although this is of great contemporary interest).

Table 1 shows the age and sex standardised mortality ratio of the under-65 population of Britain living in each decile group of areas in 1950-53 and 1990-92. At the start of the 1950s people in the worst decile areas were 31% more likely to die than average. By the early 1990s that differential had grown to 42%, the largest divergence ever recorded, resulting from steady growth in the level of geographical inequality in mortality since 1981 (when rates of inequality were marginally better than in the early 1960s). At the other extreme, people living in the areas with least deaths were 18% less likely to die than average in 1950-53. By 1990-92 this differential had grown to 24% less likelihood of death in that period. Where you live is more important in assessing your chances of dying in the early 1990s that it has been at any time since World War II. The table also shows that those living in the worst health decile areas in the early 1990s are in those areas where indicators of poverty are the greatest.

Mortality by age and sex

The rise in geographical variations in mortality is different for different groups of the population subdivided by age and sex. There is little geographical variation in the level of mortality amongst very elderly people for the simple reason that mortality is so common amongst this group. The study considered each sex and six age groups of the population separately. The main findings are as follows.

Infant mortality (under 1)

Nationally, infant mortality has seen the most dramatic falls since the start of the period under study here, but regionally divides are growing between different parts of the country. Although in most parts of Britain rates are still falling steadily, the exceptions warrant further investigation (see map 1 and map 2). For instance, male infant mortality rates have risen steadily in the northern boroughs of Blackburn, Halifax and Preston, where the death rates of boys aged under one year are now roughly twice the national average. In 1981 these areas had rates which were only just above the national average. It should be noted, however, that in all these places infant mortality rates are still at half the levels they stood at in the 1950s, despite having risen so quickly during the 1980s. There is less variation amongst infant girls' mortality, where rates only rose steadily recently in Paisley borough.

Early childhood mortality (aged 1 to 4)

Mortality is very rare in this age range, but where geographical variations do occur they can be dramatic and certainly warrant further study. Rates of male early childhood mortality in the old boroughs of Dewsbury, Bethnal Green, Manchester and St. Helens, have as much as doubled since 1981, so as to be now approaching the rate reported in these areas in the 1950s. Absolute, but smaller, rises were also recorded for young girls, again in Manchester, but also in Birkenhead, rural Isle of Wight and Newport.

Adolescent mortality (aged 5 to 14)

Amongst this age group, mortality rates are also very low, although similar places appear when those areas leading the trend towards geographical divergence in mortality rates are listed. An adolescent child's chances of dying rose during the 1980s in absolute terms in Salford, rural Camarthenshire, St. Helens, Nottingham and Bethnal Green, but in all cases rates were still lower than recorded in the 1950s. For adolescent girls, absolute mortality rates rose in Barrow-in-Furness in the most recent period.

Younger adult mortality (aged 15 to 44)

Mortality rates are much more important amongst adults in terms of the actual numbers of people who die each year. Those areas with the highest and lowest number of 'excess deaths' are shown in Tables 2 to 5. Amongst younger adult men, mortality rates fell sharply in Britain from 1950 to 1981, but have remained practically static since then, with 0.11% of this population dying per year. However, this proportion is very, and increasingly, unevenly distributed across the country. In Hammersmith, Port-Glasgow and Southwark a man this age is more than twice as likely to die than average and in too many areas to list here, adult male mortality rates are high and rising. For women, variations are also high, with absolute mortality rates rising in several parts of the country, including in many Scottish counties.

Older adult mortality (aged 45 to 64)

The great majority of premature deaths in Britain occur to people in this age range and here there has been a continuation of the improvement in rates which began in earnest in the 1970s. Despite this, geographical variations have increased again amongst this group. For men and women in the early 1960s, 6.9% and 5.4% of deaths in this age group would have been avoided if no area had over-average mortality. By the early 1990s these proportions had risen to 9.6% and 8.9% respectively. More deaths could be attributed to geographical variations in mortality rates. Where people live has become more reliable as a predictor of whether they are likely to die before the age of retirement. However, only Shoreditch records a steady absolute increase in mortality, where the mortality rate for men has risen since 1981 and is now almost twice the national average.

Conclusions

By using the same age and sex groups and same geographical areas as were used in 1951, this analysis shows that a large degree of inequality persists.

This study did not set out to speculate on the causes of the maintenance and selective widening of these divisions in society, save to say that many of the trends which have been shown here have occurred too quickly, or involve too many deaths, to be ascribed simply to changing socio-economic structures, or changing causes of death, or simply the reflection of past health variations. Explaining the patterns of life chances will be far more difficult than describing them.

About the study

The improvements in mortality rates for different age and sex groups are compared across a constant set of areas. The areas used are those for which standardised mortality ratios were published in the Registrar General's Decennial Review of 1951. Calculations are also made on the basis of contemporary administrative geography and for more recent dates (up to 1995), although records with detailed geographical location were only available up to 1992. Whereas figures on changes in mortality by social group are available for most of this century using a consistent set of definitions, data on mortality using a consistent set of areas have had to be constructed separately especially for this study.

Data from the Censuses of Population are Crown Copyright and are reproduced by permission of the Controller of Her Majesty's Stationery Office and accessed through the ESRC/JISC purchase. Data on mortality were kindly provided by the Office for National Statistics. The research on which these findings are based was also funded by the Economic and Social Research Council Health Variations Programme.

Further information

A short report from this project, Death in Britain: How local mortality rates have changed: 1950s to 1990s, by Daniel Dorling, is published by the Joseph Rowntree Foundation. (Price £11.95 plus £1.50 p&p, ISBN 1 85935 031 3).

Click on the 'order report' icon in the left margin to order online.

The technical report, Changing life chances in Britain, 1950s to 1990s, is available from the Department of Geography, University of Bristol, University Road, Bristol BS8 1SS (Price £11.95).