Income and Health 2 Inequalities in health

Part of the way in which income distribution raises overall standards of health is, as we have seen, by improving the health of the poor and narrowing the health gap.  Later, we shall see that this is by no means the whole story: the health benefits of income redistribution are too large to be accounted for in terms of health gains among the poor alone. Although the primary objective should probably be the maximisation of overall standards of health, equity in health is an important and desirable goal in itself.  Indeed, the reduction of health inequalities is the first of the World Health Organization’s European Targets.

Most of the data on socio-economic differences in health come from the Registrar General’s decennial figures showing differences in death rates between people classified by occupational class. Those figures show that the reduction in life expectancy at birth for people in unskilled occupations amounts to some seven or eight years – a loss of about ten percent of life.  At most ages death rates are about twice as high among unskil1ed manual workers and their families as they are among professional classes. This means that twice as many of a persons circle of friends and contemporaries would die prematurely each year in one class as in the other.  When it comes to illness, the differences are of a similar size: twice as many lower as upper class people say they suffer from chronic illnesses that limit their activity.

How many excess deaths result from the higher death rates in lower socio-economic groups? We can calculate the proportion of deaths among men and women of working age which would be prevented if the death rates of all manual social classes were reduced to the level of all non-manual social classes. This produces a figure of some 22,000 excess deaths a year which, because they consist of deaths occurring between the ages of 16 and normal retirement, are all premature deaths. Unfortunately problems of classifying a large proportion of the total population including the old, the young and women who are not employed, means that there is no reliable way of estimating excess deaths at other ages to produce a total associated with socio-economic disadvantage. But if we assume that socio-economic disadvantage has the same effect on the death rates of all people at all ages as it does to economically active people of working age, we would expect a total of between 100,000 and 200,000 excess deaths annually. Most of these additional deaths would of course be in later life.

Over the last generation or so these health differences have shown no sign of diminishing.  Indeed, the tendency for the health gap to widen even during decades when most people – ignorant of the trends in relative poverty – assumed that Britain was becoming a more egalitarian society, led to doubts about whether the figures could be taken at face value.

For example, it was suggested that people in lower classes had worse health, not because of the circumstances in which they lived, but because people with worse health are more likely to suffer downward social mobility. Instead of being less healthy because they lived in less good conditions, it was suggested that people ended up in less good conditions because they were unhealthy.

Data from a number of large cohort studies (which collect information from a sample of the population over a number of years) have now been used to measure the relationship between health and social mobility. They found that although illness does influence social mobility, this relationship only accounts for a small part of the overall class differences in health. Much the larger part of the health disadvantage of lower classes has to be seen as a product of their socio-economic disadvantage.

A number of other more technical problems of measurement and comparison were also raised but, on investigation, were found not to make much difference to the picture we have.

Reducing health inequalities

It is almost certainly unrealistic to believe that it is possible to develop policies which would reduce health inequalities without tackling the underlying socio-economic inequalities.  First, the excess deaths are not the result of just one or two preventable diseases.  Instead, the vast majority of diseases are more common lower down the social scale. Nor can the imbalance be attributed to the known behavioural influences on health.  For example, differences in death rates among diseases unrelated to smoking are just as great as for diseases which are related to smoking. In fact, modern knowledge of the causes of disease is still quite unable to account for most of the class differences. Very little is known about how to prevent many of the important causes of death. Even with a disease such as heart disease, which has been much more thoroughly researched than most, the known risk factors explain a good deal less than half the class differences in death rates. But even where we do know the risk factors they are often only partially controllable. Sometimes this is because health advice has only a very minor influence on behaviour, and sometimes it is because even if everyone followed all the do’s and don’ts of healthy living to the letter, it would only lead to a minor reduction in the underlying physiological risk. For instance, the amount of fat in the diet is only one of a number of determinants of blood cholesterol levels, and they in turn, are only one of several determinants of heart disease. Even if adults changed their diets enough to lower their blood cholesterol levels by ten percent for forty years, only one in fifty men and one in four hundred women would have prevented a heart attack before their 55th birthday.

Most of the published evaluations of projects which have attempted to influence health related behaviour suggest that health advice has, with a few exceptions, only a minor influence on most people’s behaviour. In part this probably reflects people’s understandable scepticism about how much difference behaviour change will make to their health. But the frequent and well documented failure of so many health education projects suggests that it is unwise to approach behaviour change as if people’s drinking, smoking, exercise and dietary patterns were not bound up with every other aspect of their lives.  It is precisely because behaviour is related to personal circumstances that the behavioural risk factors so far identified are systematically related to social class.

Finally, it would be wrong to think that health differences result from differences in medical care.  First, although differences in medical care exist, they are relatively small in comparison to the differences in health.  But even if they were not, medicine is largely ineffective in relation to some of the most important modern causes of death. The fact that the first symptom of heart disease and stroke is often sudden death clearly limits the scope for medical intervention. Among some of the most common cancers the lack of any long-term improvement in case-fatality rates suggests that medical treatment has rather little impact on death rates. Although there are class differences in case-fatality rates for cancers and heart disease, they do not seem to be related to differences in medical care and are anyway overshadowed by the scale of class differences in the incidence of these diseases. The controversy which exists about the influence of medical care on death rates is less about how large it is as about whether it is measurable at all.

It has to be accepted that health inequalities are, first and foremost, a testament to the continued strength of socio-economic influences on health. Health inequalities could only be reduced while leaving the underlying socio-economic inequalities intact if known risk factors accounted for more than a small fraction of the overall burden of ill-health and if they could easily be prized from their social context. However, whether or not the aim is considered desirable, the fact that this is not the situation leaves us little choice. Even tackling known risk factors in such a way as to reduce, rather than widen, health inequalities is beset with difficulties. Take smoking for example.  It is often said that health education campaigns widen class differentials in smoking and that it would be better to increase the tobacco tax. But it has been calculated that smokers among the poorest twenty percent of households already spend nine percent of their disposable incomes just on tobacco tax. Paying additional tax may well add to the health problems of those who are not deterred from smoking. A realistic policy for reducing inequalities in health must then address itself to the underlying socio-economic inequalities.

Income almost certainly provides the best means of tackling health inequalities. Income differences come close to the core of the underlying socio-economic inequalities; they exert a powerful influence on health. Income distribution and levels of relative poverty are highly sensitive to policy on taxes and benefits.  Data from the Health and Lifestyles Survey suggests that the association between social class and health is primarily an association between income and health. Although there must be other features of socio-economic status which affect health, few are as accessible to policy as income.