Report of Darwin Lecture

The 1991 Darwin Lecture was delivered in the London School of Economics on 6 November by Dr John Fox, Chief Medical Statistician at the Office of Population, Censuses and Surveys. His subject was "Social Factors and Health".

Introduction

Dr Fox stressed the progress which has been made in the last 15 years to increase the variety of survey data available for analysis. Earlier work focused primarily on the relationship between occupation and mortality among men of working age, virtually ignoring women, children and the elderly.

Risk Factors and Indicators

The range of social factors analysed for health differences in recent years includes the following:

• "Social class" - indicated by occupation
• Unemployment

• Tenure and car ownership
• Housing conditions and education
• Ethnic origin
• Migration, income and wealth
• Location - "deprived" areas
• Family roles
• Marital status (divorce, bereavement)
• Health behaviour

The range of health indicators has also widened to include:

• mortality (by cause of death)
• hospitalisation (by diagnosis)
• consultation of general practitioner (by diagnosis)

The richness of survey data now available is partly due to the variety of study designs:

• Routine (at national level)
• unlinked
• linked (eg infant mortality related back to birth records for social class, birth weight and marital status of the mother)
• cross sectional (eg the general household survey)
• longitudinal (information on some samples over different periods of time are brought together)
• Ad hoc
• cross-sectional
• follow-back
• prospective
• longitudinal (eg Whitehall study of Civil Servants)

Findings

Among the findings which have emerged from these surveys are the following:

• Differences do exist and are widespread
• The differences are not consistent for different diseases and subgroups
• Car access and tenure are important indicators: these indicators (eg local authority versus owner occupiers) are not subject to the same problems of interpretation as occupation and social status and they are not so changeable
• Relative differences show signs of widening. This seems to be particularly true over the period since the last war for men aged 30-64. A widening has not been observed for women over the last 20 years or for men in other age groups.
• More causes of death portray differences than in the early part of the century. For example, in 1920 less than half the causes of death showed class differences but it is now the exception for a cause to show the opposite trend; in particular, ischaemic heart disease and lung cancer showed no correlation in the l920s but do now.
• It can be hard to say whether observed patterns are the result of differential incidence or subsequent events such as use of medical services; for most cancers the former appears to be the case but for some cancers, particularly those with long survival, the differences do sometimes relate to subsequent events.

Explanations

Dr Fox considered possible explanations for these observations. He noted that the differences are not consistent, showing variable gradients although usually in the same direction. Clearly some factors influence many conditions. The Black report suggested several possible influences:

• Artefact. This is not really relevant with today’s databases and wide range of different measures - eg cars and tenure.
• Selection. The suggestion here is that the constitutionally weak go down the social ladder. However, longitudinal studies looking at people holding constant status show exactly the same differences, as do other studies using constant measures such as educational status.
• Material Circumstances.
• Behaviour.

Dr Fox said that over recent years he had been increasingly struck by the arguments over the relative roles of individual and society. For example, deaths due to lung cancer are very closely related to smoking. Therefore differences between social groups come down to different smoking habits. But why are there these differences and have they always existed? In fact they didn’t in the 1950s when the health effects were unknown. The current differences are therefore a reflection of how the know ledge was used when it became available – rather than, for example, putting additional tax on cigarettes to discourage poorer people we relied on education which is absorbed more quickly by higher classes.

Conclusion

Although he had concentrated on the UK where until recently the quality of data had been ahead of other countries, Dr Fox stressed that the issues he had raised have become the subject of international debate. There are now eight or nine countries with comparable data who are tackling social factors and health. In the UK the recent "Health of the Nation" green paper has been criticised for failing to include inequalities in health among its targets, but this will be implicit in tackling, say, lung cancer and heart disease.

Concluding, Dr Fox hoped that he had convinced his audience that the UK database is much richer than it was ten years ago and remains the richest in the world.

Robert Peel