Black Report 2 The Evidence of Inequalities in Health


2.1 Inequalities in human health take a number of distinctive forms in Britain today. In this report, for the reasons set out in Chapter 1, most attention is given to differences in health as measured over the years between the social (or more strictly occupational) classes. These differences are highlighted in Table 2.1 by comparing rates of mortality among men and women in each of the Registrar General’s 5 classes. Taking the 2 extremes as a point of comparison it can be seen that for both men and women the risk of death before retirement is two-and-a-half times as great in class V (unskilled manua1 workers and their wives), as it is in class I (professional men and their wives). If attention is confined to age-standardised deaths rather than all deaths of those aged 15-64 then the ratio for class V males becomes a little under twice (1.8) that of class I (OPCS) 1978, P37)

2.2 This great gap in the life chances of men and women at the 2 polar ends of the occupational spectrum is, however, not the only source of health inequality for, as Table 2.1 also indicates, the risk of death for men in each social class is almost twice that of their wives.

Table 2.1: Death rates by sex and social (occupational) class (15-64 years)

Rates per 1000 population England and Wales 1971

Social (occupational) class Males (all) Females (married, by husbands occupation) Ratio M/F
I (Professional) 3.98 2.15 1.85
II (Intermediate) 5.54 2.85 1.94
IIIn (Skilled non-manual) 5.80 2.76 1.96
IIIm (Skilled manual 6.08 3.41 1.78
IV (Partly Skilled) 7.96 4.27 1.87
V (Unskilled) 9.88 5.31 1.86
Ratio V/I 2.5 2.5

Source: Occupational Mortality 1970-72. (Microfiches and 1978, P-37)

Note: The decennial supplement of Occupational Mortality for 1970-2 provides data on the class of married and widowed women classified by (1) their present or former husband’s occupation, and (2) their own occupation where this is applicable. The difference between these 2 measures is only significant in the case of women in class I (ie professional workers). When classified by their own occupation such women have somewhat lower rates of mortality from most causes than those of women allocated to class I by their husband’s occupation (see pp 153-156). In this table women with husbands have been classified by their husband’s occupation, women of other marital statuses are attributed to their own occupational class.

Sex Differences in Health

2.3 One of the most distinctive features of human health in the advanced societies is the gap in life expectancy between men and women. This phenomenon carries important implications for all spheres of social policy but especially health, since old age is a time when demand for health care is at its greatest and the dominant pattern of premature male mortality has added the exacerbating problem of isolation to the situation of elderly women who frequently survive their partners by many years. The imbalance in the ratio of males to females in old age is the cumulative product of health inequalities between the sexes during the whole lifetime. These inequalities are found in every occupational class (Table 2.1), demonstrating that gender and class exert highly significant and different influences on the quality and duration of life in modern society.

Regional Differences in Health

2.4 Rates of age-specific mortality vary considerably between the regions which make up the United Kingdom. Using mortality as an indicator of health the healthiest part of Britain appears to be the southern belt (below a line drawn across the country from the Wash to the Bristol Channel). This part of the country has not always exhibited the low rates of mortality that are found there today. In the middle of the nineteenth century, the South East of England recorded comparatively high rates of death, while other regions like Wales and the far North had a rather healthier profile. The fluctuation in the distribution of mortality over the years suggests that social (including industrial and occupational) as much as “natural” factors must be at work in creating the pattern of regional health inequalities. Table 2.2 depicts regional variation in mortality standardised for age and for both age and occupational class. Once again it is clear that these variables exert, at least statistically, an independent influence on human health.

Table 2.2: Regional Variations in Mortality

SMR: standardised for
Standard Region Age Age and class
Northern, Yorkshire & Humberside 113 113
North West 106 105
East Midlands 116 116
West Midlands 96 94
East Anglia 105 104
South East 90 90
South West 93 93
Wales I 114 117
Wales II 110 113
England and Wales 100 100

Source: occupational Mortality 1970-72, P-180

Race, Ethnicity and Health

2.5 One of the most important dimensions of inequality in contemporary Britain is race. Immigrants to this country from the so-called new Commonwealth, whose ethnic identity is clearly visible in the colour of their skin are known to experience greater difficulty in finding work and adequate housing (Smith, 1976). Given, for example, these social and economic disabilities it is to be expected that they might also record rather high than average rates of mortality and morbidity.

2.6 This hypothesis is difficult to test from official statistics since “race” has rarely been assessed in official censuses and surveys. Moreover, it far from clear what indicator should be utilized in any such assessment (eg skin-colour, place of birth, nationality): that most significant may indeed depend upon the precise issue of interest.

2.7 The pattern of social and economic disadvantage experienced by black Britons is connected with occupational class and is reflected in the working of the labour market. But other factors may also be important and amongst adult males at least, the variables of occupational class and race do not compound one another in a linear fashion when place of birth is used as a means of measuring race. As Table 2-3 indicates, the age standardised mortality ratios of immigrant males compare favourably with their British born equivalents in occupational classes IV and V, but less so higher up the scale in classes I and II. The interpretation of these ratios is made difficult at the higher end of the occupational scale because they are based on small numbers.

TABLE 2.3 Mortality by Country of Birth and Occupational Class (SMR) (Males 15-64)

Country of Birth I II IIIN IIIM IV V All
India and Pakistan 122 127 114 105 93 73 98
West Indies 267 163 135 87 71 75 84
Europe (including UK & Eire) 121 109 98 83 81 82 89
UK & Eire (including England and Wales) 118 112 111 118 115 110 114
England and Wales 97 99 99 99 99 100 100
All birth places 100 100 100 100 100 100 100

Source: Occupational Mortality, 1970-72, pp 186-187

In the poorer occupational classes, where the SMR is based on larger numbers of deaths, men born in India, Pakistan or the West Indies seem to live longer than their British born counterparts. It should be remembered, however, that the percentage of workers in class V among the British born is less than 7 while the equivalent percentage of those born in, for example, India and Pakistan is 16. In addition, of course, the average British born male classified as an unskilled manual worker is likely to be older than his foreign born counterpart and, is more likely to have acquired, this low occupational status after a process of downward social mobility associated with failing health.

2.9 This rather favourable comparison between immigrant and British born males may also reflect the underlying tendency for migrants to select themselves on the grounds of health and fitness. Men and women prepared to cross oceans and continents in order to seek new occupational opportunities or a new way of life do not represent a random cross section of humanity. A better comparison for exploring health inequality would ideally involve second or third generation immigrants, but these are the very groups that are difficult to trace for statistical purposes. What little evidence that has been accumulated however does suggest that the children of immigrants do suffer from certain specific health disabilities related to cultural factors such as diet or to their lack of natural immunity to certain infectious diseases (Thomas: 1968; Oppe: Gans: 1966). Studies based on small samples of immigrant children have pointed to the possibility of higher than average morbidity associated with material deprivation but the evidence is scarce and somewhat inconclusive and needs to be augmented by further research (Hood, et al, 1970).

Social Class and Health

2.10 Social class is a further concept by means of which inequalities in industrial society may be examined. It reflects income, property, occupation and education, and much else. The data presented in the remainder of this chapter employ occupation as a means of approximating social class and for this reason, as Chapter One indicated, the variable will often be referred to (more accurately) as occupational class.

2.11 The measurement of “health” in official statistics and government publications is usually achieved by utilising data on rates of mortality and morbidity. In fact these measures are both indicators of ill-health but, since indicators of health in its “positive” sense are fewer, less available, and relevant principally to children (eg height, weight) they provide the most feasible and readily accessible substitute.

2.12 One alternative means of depicting the level of human health in a population is to take measures of bodily growth, development and decay. In Britain this kind of data is not routinely collected except in the first week of life. Some data on developmental processes in childhood have been accumulated as part of ad hoc and longitudinal surveys and this does indicate the existence of class differentials in height and weight and in patterns of dental health (Miller et al 1974; Todd, J E 1975; Goldstein, 1971). Table 2.4 gives a breakdown of class differentials in birth weight.

TABLE 2.4 Occupational Class and Birth Weight

Birth weight class (mother’s husband) Fatherless
% who were (I II) III (IV and V)
Less than 2500 grams 4.5 5.6 8.2 9.5
More than 3000 grams 81.0 76.3. 72.7 66.7

Source: Chamberlain (1975)


2.13 The persistence of class differentials in both mortality and morbidity is evidenced in regular reports provided by the OPCS. Since 1970 the investigation of occupational differences in self-reported sickness and medical consultation has been carried out as part of the General Household Survey (GHS). The GHS is carried out annually by the Office of Population Censuses and Surveys. Based upon a sample of 15,000 households in Great Britain, the survey provides data on a range of topics including health, education, employment, housing and migration. Since its inception In 1970, it has been possible to estimate the occurrence of morbidity as reported by sampled men and women but although this information has been updated every year from 1970 to 1977, detailed analyses of trends on an annual basis aredifficult because of the restricted size of sample and there is no opportunity to examine a longer time series. In 1977 and 1978 the health questions were however redesigned.

2.14 Although it has been well known for over a hundred years that occupation and social class are implicated in the aetiology of many diseases there are still no regular and reliable official statistics of provision of (let alone need for) medical treatment by the social or occupational class of the patient. The National Morbidity Survey carried out in 1955-56 collected data on a voluntary basis from general practitioners scattered all over Britain who provided some data on the occupational class characteristics of adult males and their children – but not women. The Hospital In-patient Enquiry in England and Wales does not make use of occupational categories in the publication of data because the quality of the information is felt to be too poor. See Chapter 4.

2.15 The empirical evidence of class differences in morbidity is therefore partial. We know from the GHS that men, women and children from lower class households generally report higher rates of chronic and acute sickness and we also know, by their own account, that they tend, especially during adult working life, to consult more frequently with their general practitioners. These data are based on self recollections and as such do not provide a wholly satisfactory standard measure of morbidity.

2.16 The analysis of differences in rates of mortality has a considerably longer history. Since the mid-nineteenth century decennial reports of occupational mortality have been provided by the Registrar General. In the century or more that has elapsed since this practice began the measurement of occupations has been complicated by the evolution in the occupational structure of the older industrial economies like Britain. Other changes in measurement conventions during the twentieth century, when added to the fact that many established occupations have disappeared while new ones have grown in significance, makes it difficult to analyse trends in mortality by occupational class over the whole time period for which data are available. Class differences in the risk of premature death have survived into the late twentieth century despite the dramatic decline in deaths from infectious disease. Class and occupational differences in rates of age-specific mortality offer the best alternative means of exploring the health gap in present day Britain. Every death in Britain is a registered and certificated event in which both the cause and the occupation of the deceased or his or her next of kin are recorded. By treating the actual incidence of death among members of the Registrar General’s ‘social’ classes as a numerator and by taking the denominator from the Census it is possible to derive an estimate of class differentials in mortality.

2.17 An analysis of the relationship between occupational class and mortality is carried out every decade and presented in the form of a decennial supplement. The data analysed in these successive supplements present certain problems of inference. The basic source of these problems is the method of estimating mortality in the different occupational classes. Because it is based on 2 different types of evidence and, because the measurement of occupations in one of these, ie the Census, is thought to be more accurate than in the other, there is some scope for error. In the future, these difficulties will partly be overcome by use of the longitudinal survey which the OPCS started in 1970, and which consists of a linkage of one per cent of individual census returns to the system of death registration, (Fox and Goldblatt). Death certificates are used only to indicate the event of death and not as a source of substantive data in their own right. All of the relevant statistical information is thus drawn from the census. Some of the early results of this survey are considered below.

2.18 In all previous decennial supplements occupational class differences in rates of age specific mortality for most major causes of death show gradients of varying slope with class I located at the most advantaged extreme and class at the most disadvantaged.

2.19 Although the risk of early death is an obvious indicator of ill-health its precise relationship to sickness, disease or morbidity is not straightforward. In present day Britain accidents are the major cause of death during childhood and early adult life and there is no necessary relationship between an accident and an individual’s state of health at least in these age ranges. In the first year of life and during old age the rate of mortality is a rather more efficient measure of morbidity. These age-specific differences in the relevance of mortality to the measurement of ill-health should be borne in mind in the interpretation of the evidence presented below.

2.20 Having noted some of the major problems of obtaining and interpreting the statistics of mortality and morbidity it is now necessary to examine them in greater detail to see what they reveal about class based patterns of advantage and deprivation in health in Britain today.


2.21 Contemporary trends in occupational mortality have been extensively reviewed in the most recent decennial supplement of occupational mortality (OPCS 1978). A summary of some of the most relevant findings will be presented here.

Figure 2.1

Mortality by Occupational Class and Age

Source: Occupational Mortality 1970-72 HMSO 1978 p196

Note: Relative mortality (%) – ratio of rates for the social class to the rate of all males (females)

2.22 Class differences in mortality are a constant feature of the entire human lifetime. They are found at birth, during the first year of life, in childhood, adolescence and in adult life. In general they are more marked the start of life and in early adulthood. Average life expectancy provides a useful summary of the cumulative impact of these advantages and disadvantages throughout life. A child born to professional parents, if he or she is not socially mobile, can expect to spend over 5 years more as a living person than a child born to an unskilled manual household. Figure 2.1 illustrates the consistency of class gradients in mortality through the lifetime.

Table 2.5: Stillbirths and infant Deaths by Sex, Age and Occupational Class

Occupational Class
I II IIIN IIIM IV V All infants Ratio V:I
Stillbirths per 1000 live and still births
Males 8.63 10.16 11.44 12.26 12.73 17.16 12.36 2.0
Females 8.92 10.01 11.54 12.81 13.41 17.67 12.67 2.0
Perinatal deaths (stillbirths and less than 1 week) per 1000 live and still births
Males 17.44 19.79 22.02 23.16 25.27 33.93 23.80 1.9
Females 15.17 17.36 19.05 20.98 22.42 30.24 21.14 2.0
Early neo-natal deaths (less than 1 week) per 1000 live births
Males 8.89 9.73 10.70 11.04 12.70 17.06 11.59 1.9
|Females 6.31 7.43 7.60 8.27 9.14 12.64 8.58 2.0
Late neo-natal deaths (1-3 weeks) per 1000 live births
Males 1.23 1.39 1.64 1.81 2.02 3.06 1.84 2.5
Females 0.99 1.29 1.27 1.53 1.84 2.41 1.57 2.4
Post neonatal deaths (1-11 months) per 1000 live births
Males 3.47 4.09 4.57 6.20 7.31 14.61 6.48 4.2
Females 2.32 3.22 3.11 4.99 5.97 11.62 5.11 5.0
Total infant mortality (0-11 months)
Males 13.60 15.21 16.91 19.06 22.03 34.73 19.91 2.5
Females 9.61 11.94 11.99 14.79 16.95 26.67 15.27 2.1

Source: Occupational mortality 1970-72 p157 HMSO 1978.

2.23 At birth and during the first month of life the risk of death in class V (unskilled manual workers) is double the risk in class I (professional workers). When the fortunes of babies born to skilled manual fathers (class IIIM) are compared with those who enter the world as the offspring of professional workers (class I) the risk of mortality is one and half times as great. From the end of the first month to the end of the first year, class differentials in infant mortality reach a peak of disadvantage. For the death of every one male infant in class I, we can expect almost 2 deaths in class IIIM and 4 deaths in class V. Among female infants these ratios are even more disadvantageous to the offspring of manual workers. The latest rates of infant mortality (1975-77) suggest that the position of classes IV and V may be improving.


Infant Mortality by Sex, Occupational Class and Cause of Death

2.24 What causes these differentials in life chances among Britain’s youngest citizens? Figure 2.2 compares class gradients for different causes of death, demonstrating that the steepest curves are found for accidents and respiratory disease, causes of death which are associated with the socio-economic environment. (see chapter 6). Other causes associated with birth itself and with congenital disabilities have less steep class gradients.

Table 2.6: Childhood Mortality

Occupational Class
Sex and age group I II IIIN IIIM IV V All children Ratio V:I
Males aged 1-4 Deaths 165 468 338 1407 660 384 3778
Deaths per 100,000 children pa 60.58 62.02 74.53 75.53 92.89 128.72 78.60 2.12
Males aged 5-9 Deaths 96 337 216 998 412 259 2579
Deaths per 100,000 children pa 27.51 31.41 38.93 42.40 44.06 69.10 41.57 2.5
Males aged 10-14 deaths 78 313 172 703 355 199 2032
Deaths per 100,000 children pa 28.32 31.07 34.81 34.64 40.06 56.57 36.57 1.99
Males aged 1-14 339 1118 726 3108 1427 842 8389
SMR 74 79 95 98 112 162 100 2.2
Females aged 1-4 Deaths 146 377 268 1087 570 309 3060
Deaths per 100,000 children pa 57.48 54.02 62.16 61.85 84.21 109.09 67.36 1.90
Females aged 5-9 Deaths 97 245 147 604 294 157 1685
Deaths per 100,000 children pa 26.62 23.81 27.46 27.30 32.53 43.43 28.58 1.63
Females aged 10-14 deaths 55 204 95 408 222 108 1187
Deaths per 100,000 children pa 20.83 21.33 19.89 21.29 26.44 32.76 22.57 1.57
Females aged 1-14 288 826 510 2099 1086 574 5932
SMR 89 84 93 93 120 156 100 1.75

Source: Occupational Mortality 1970-72 Table 7.8 p159

2.25 Between the ages of one and fourteen, the risk of mortality continues to be closely correlated with class. Among boys the ratio of mortalty in V as ompared to I is of the order of 2: 1, among the girls it varies between 1.5:-1 to 1.9:1. Once again the cause of this difference can be traced to environmental origins. The most steep gradients in childhood are found for accidents (33 per cent of total causes). For deaths caused by fire, falls and drowning the risk for boys in class V is 10 times the risk for their peers in class I. The corresponding ratio for deaths caused to youthful pedestrians by motor vehicles is more than 7:1. The other major causes of death showing steep class gradients in childhood are infective and parasitic diseases (5 per cent of total) and pneumonia (8 per cent of total). For most other causes, there is less clear evidence of class disadvantage.

FIGURE 2.3: Class and Mortality in Childhood (Males and Females 0-14)

Class and Mortality in Childhood (Males and Females 0-14)

2.26 Class differences in mortality for all adults aged 15-64 are somewhat less marked than in childhood, but this conceals a large difference for those in their 20s and 30s, and a smaller difference for those approaching pension age, ie class disadvantage becomes less extreme as men and women grow older and the frequency of death increases.

TABLE 2-7: SMRs (16-64) by Sex and Class

Men 77 81 99 106 114 137 1.8
Married women 82 87 92 115 119 135 1.6
Single women 110* 79 92 108 114 138 1.25

Source: Occupational mortality, 1970-72, HMSO, p.211.

*based on a small number of deaths.

2.27 The risk of death in class V is between one and half to two times the risk in class I for adult males and females. Within the range 15-64 years, class differences in mortality are greatest for the earlier ages as Table 2.8 indicates. Among males of less than 44 years the ratio of deaths in V as compared to I is 2.45, after 45 years it falls to 1.7

Table 2.8: Class and the Relative Age Gradient for Males

Class (a) SMR 15-44 (b) SMR 45-64 Relative age gradient (b)-(a)
I 71 78 +7
II 75 82 +7
IIIN 91 100 +9
IIIM 94 108 +14
IV 121 113 -8
V 174 132 -42
Ratio V/I 2.45 1.7

Source: Occupational Mortality 1970-72, HMSO 1978, P.14.

Note: Relative age gradient. Difference in the SMR between the age 15-44 and 45-64. A positive gradient indicates a higher SMR in the older age group.

2.28 In adult life, class differences in mortality are found for many different causes. As in childhood the rate of accidental death and infectious disease forms a steep gradient especially among men; moreover an extraordinary variety of causes of deaths such as cancer, heart and respiratory disease also differentiate between the classes.

Figure 2.4: Occupational class and mortality in adult life (Men and married Women) [by husband’s occupation] (15-64)

Occupational class and mortality in adult life (Men and married Women) [by husband’s occupation
2.30 The steep class gradient found in deaths attributable to accidents is only in part a reflection of the greater risks of manual work. Accidents at work account for only 20 per cent of the total of deaths recorded as accidents and as would be expected, most of these lost lives were male manual workers. The majority of accidental deaths were attributable to motor vehicle traffic accidents (54 per cent) which were not work related. For these deaths there was a steady class gradient. (see Occupational Mortality, PP-57-59).

2.31 The preceding analysis has relied on grouped occupations as an indicator of class. This form of indicator, which is the one conventionally adopted by the Registrar General, is of course no more than an artefact. There are several other ways in which social class might be represented for purposes of statistical presentation. Income, property, education and housing tenure are among the possible alternative variables which might be used to stratify the population. We have relied on occupation in this survey of evidence because this is the form in which OPCS provides a detailed analysis of mortality. Occupation is of course directly correlated with variables like income, education, and property, all of which can also be expected to differentiate between social groups within the population in terms of life expectancy. But of course in a review such as this we are constrained by the availability of data.

2.32 In future years, the OPCS will be able to produce more detailed analyses of the socioeconomic variables associated with premature mortality on the basis of the new longitudinal survey. This survey has already produced some preliminary results which suggest some of the limitations of occupation as a uni-dimensional measure of social class. When the population is divided into 3 tenure groups: owner-occupiers, private tenants and local authority tenants, it has been found that class gradients, while maintaining their distinctive slope, vary considerably.

TABLE 2:9 Mortality by Tenure and Class (SMR: Males 15-64 years)

Class Owner occupied Privately rented Local authority tenancy
I 79 93 99
II 74 104 99
IIIN 79 112 121
IIIM 83 99 104
IV 83 100 106
V 98 126 123

Source: Unpublished data, Medical Statistics Division, OPCS, preliminary results of the LS 1970-75.

2.33 People who live in homes which they own themselves have lower rates of mortality than those who rent their homes from private landlords who in turn have lower rates than those who are the tenants of local authorities. Housing tenure is one possible measure of the accumulation by individuals or households of fixed property or assets. Here it can be seen that this variable shows a very close relationship with the risk of premature death.

2-34 The influence of social class on the variations in mortality after retirement has traditionally been difficult to estimate. This is because occupation, which is mainly used by the Registrar General to indicate class is less indicative of the lives of retired men and their wives and is often unspecific, or open to selection from more than one occupation in life, at death registration. In this surey of evidence, we are again able to draw on the preliminary results of the longitudinal survey to throw some light on the distribution of mortality among retired men in different classes.

Table 2.10 Occupational Classes and Mortality amongst Retired Men aged 65-79, 1970-75

Class Observed Expected SMR
I 101 110.5 91
II 629 729.4 86
IIIN 406 499 93
IIIM 1377 1329.5 104
IV 732 769.1 95
V 503 422 119
Armed forces 9 17.8 51
Inadequately described 762 701.6 109
All retired men 4597 4578.9 100

2.35 Table 2.10 demonstrates the influence of last recorded occupation on the mortality rate of retired men. Men who followed non-manual professional or managerial occupations during their working lives tend to have below average death rates, while those who followed manual occupations tend to have higher death rates. The differences between the Registrar General’s 6 occupational classes do not form a linear class gradient, in these results, class IV manual (ie semi-skilled manual occupations) deviates from the trend of other manual workers. It remains to be seen whether this reflects a real divergence or whether it is due either to the vicissitudes of sampling or some problem arising from taking the “last recorded” as distinct from, say, “the longest occupation” in life.

2.36 The relationship between occupational class and the risk of death in old age present fewer problems of interpretation. During working life most active men and their wives and families live above a minimum or subsistence level of income, at least for most of the time. Retirement and old age, however, are characterised by widespread social and material deprivation, especially among manua1 working class households, who are less well insured against the event of poverty at a time of life when the human body is in its declining phase and when the risk of death is proportionately so much greater; but also among some non-manual households, where an occupational pension may be a very small addition to the state’s pension, or depreciate rapidly in value.

2-37 The duration of the human lifetime is one of the best means of approximating the lifelong pattern of health of individuals and whole populations. As we have seen, the risk of premature death in Britain today is systematically related to socioeconomic variables. This association is not new or unusual. Death rates have always been relatively high among the underprivileged and materially deprived sections of communities. Why this should continue to be so in an era characterised by new patterns of disease, increased purchasing power, and state provision of free medical care is more perplexing. In infancy and childhood where the class gradients are steep, the major causes of death are in many ways directly linked to poverty and to environmental risk. In adulthood the relationship between health and class becomes more complex and in old age social and economic deprivation becomes a common experience. The problem of explanation will be discussed more fully in Chapter 6.

2-38 Mortality, it must be remembered, is an indirect measure or indicator of health and one which reflects many other aspects of material welfare, such as the risk of accidents, which are not normally associated with ill-health as such. Even so, the evidence surveyed in preceding pages suggests that occupational class is closely related to the likelihood of premature death.

Morbidity data

2.39 Morbidity data provide a second way of looking for, or at, class inequalities in health. Moreover there is a sense in which the extent of ill-health in a social group is a better indicator of its health vis-a-vis another social group than is relative mortality rate. Morbidity data are available from a variety of cohort studies and ad hoc surveys.

2.40 Such morbidity data are of 2 kinds, though both are scant at the national level. The first is based upon examination of, or symptom identification in, the ‘social group as a whole or in a properly selected sample. An approach of this kind has sometimes been used in the attempt to assess the prevalence of specific diseases within research studies. Social or occupational class is sometimes noted.

2.41 The second kind of data derives from analysis of medical consultation rates, hospital admission rates, etc. Not only do we have few data of this kind by occupational class, but there is the disadvantage that rates reflect not only the incidence of disease but also the processes by which an individual defines him(her) self as ill, seeks medical attention, and has his(her) definition confirmed/legitimated by medical authority. Since we know that there are class-related differences in the propensity of an individual with a given set of symptoms to refer himself for treatment or attention, as well as in the subsequent medical response, we recognise that data of this kind cannot be interpreted clearly.

2.42 Nevertheless illustration by reference to the sort of data available from these 2 kinds of source will confirm, broadly speaking, the picture which mortality data indicate. We shall not deal here with studies specifically of childhood morbidity (of which there are a number) since we shall be drawing on these elsewhere in our report.

2.43 An example of the first sort of morbidity data is provided by a survey of the prevalence of chronic bronchitis in Great Britain carried out by members of the (Royal) College of General Practitioners around 1960 (College of General Practitioners, 1961). Ninety-two GPs, distributed throughout the country, were asked to select similarly sized age/sex stratified random samples from their practice lists. All were to be aged between 40 and 64. Since bronchitis is diagnosed principally on the basis of symptoms, a questionnaire covering many possible symptoms was administered by each doctor, who was then asked to note whether he thought the patient suffered from chronic bronchitis or any other chest disease. The project’s steering group subsequently identified those questions which appeared to discriminate most effectively between ‘bronchitis’ and ‘non-bronchitis’. The total sample examined (787 men and 782 women) corresponded on social class distribution and marital status with expectation from the 1951 census. The wives of 442 men selected were also interviewed.

2.44 Overall, the GP’s themselves diagnosed 17% of male patients and 8% of female patients as suffering from chronic bronchitis. In terms of the 3 criteria identified by the steering group (‘standard diagnosis’), these figures fell to 8% and 3%. We consider here only the relationship between bronchitis prevalence (defined in each of these 2 ways) and occupational class of male patients (the 785 who could be classified) and 442 wives. In terms of GPs diagnosis the percentage suffering from chronic bronchitis rose with descending class from 6% in class I (0% of wives) to 26% in class V (15%wives). The picture, broadly paralleling that obtained from mortality rates, was maintained if the more rigorous ‘standard diagnosis’ was used: 3% in classes I and II (4% of wives), 10% in class III (3% of wives) and 10% in classes IV and V (9% of wives). “In other words”, the Report adds, “here is evidence that the social gradient in reported mortality from chronic bronchitis is unlikely to be due to differences in the diagnostic skills and habits of the doctors certifying death”.

2.45 A different means of acquiring morbidity data from general practice has been through GPs recording details of consultations. In the analysis of consultations with 76 practices (120 practitioners) between May 1955 and April 1956 carried out under the auspices of the College of General Practitioners and the (then) General Register Officer (now OPCS), occupation was recorded.

2.46 Results from the study (Logan and Cushion 1960) showing consultation rates for each of a wide range (or group) of conditions for males, females, children by social class, are of considerable interest though not easy to interpret. The authors summarise their conclusions relating to working males as follows (p.21);

“The picture that emerges from this analysis of morbidity amongst working males is by no means a clear and obvious one. Nevertheless, some fairly definite correlations can be recognised. With a certain amount of simplification and ignoring numerous exceptions the position can be summarised approximately in the following scheme, where +indicates morbidity above, and – below average.”

agricultural occupations non-manual occupations manual occupations
psychoneurotic disorders
cardiovascular disorders
respiratory disorders
gastric disorders

2-47 Another way of looking at the results suggested is by comparison of condition specific SMRs with Standardised Patient Consultation Ratios (defined as actual number of men aged 15-64 consulting at least once with a given diagnosis per cent of the number ‘expected’ to have consulted on the basis of patient consultation rates at corresponding ages of men in all occupations). The results of this comparison, for selected diagnoses, is given in Table 2.11.

2.48 If we then compare the class gradients for SPCRs and SMRs respectively we find that in those cases where there is a clear association of consultation or sickness and mortality and working class status (notably pneumonia and bronchitis), then the gradient is steeper for SMR. This suggests more severe sickness or less adequate treatment with declining class.

TABLE 2.11


SPCR Class SMR Class
respiratory tuberculosis 102 85 105 102 91 58 63 102 95 143
malignant neoplasms 75 111 94 91 111 94 85 104 95 113
diabetes mellitus 89 123 100 108 74 134 100 99 85 105
coronary disease/angina 89 108 102 89 93 147 110 105 79 89
hypertension 120 127 99 70 89 123 106 103 83 101
influenza 83 82 103 113 107 58 70 97 102 139
pneumonia 70 87 90 121 132 53 64 92 105 150
bronchitis 49 70 99 118 146 34 53 98 101 171
gastric and duodenal ulcer 48 78 99 88 116 68 76 101 99 134

Source: Logan and Cushion, 1960 p. 16

2.49 Very approximately, however and bearing in mind the complexities and ambiguities of direct estimate of disease prevalence or incidence in the community, it seems reasonable to conclude that the picture presented by mortality data is sustained.


2.50 Given the significance of class as a variable in the analysis of the length of the life-span in contemporary Britain, it is disappointing that there are no official statistics of the rate or outcome of medical therapy by the occupation or social class of the patient. Infective and parasitic disease with a steep class gradient amongst children accounts for 5 per cent of all deaths in the 1-4 age group. Between 15-64, such disease accounts for less than 1 per cent of all deaths. Since the advent of chemotherapy, the effectiveness of medical intervention in diseases caused through infection by microparasites has been exceptional and it seems likely that many of these childhood deaths ought, in theory, to be prevented. Why they are not, in the absence of reliable official statistics, remains a matter for conjecture. The same problem obstructs any attempt to examine the effectiveness of the National Health Service as an agent of equality in contemporary Britain. It is impossible, in fact, even to begin to evaluate what impact this great innovation in social policy has had on the relative life chances of men, women and children in the manual working class. If rates of mortality are used to evaluate its achievement, then recent experience would not appear to be particularly favourable.

2.51 For the analysis of social class and occupational differences in morbidity, the only regular source of information provided by central government is to be found in the General Household Survey (GHS). This annual sample survey has included substantial coverage of health topics and the report includes an analysis of socioeconomic group of personal reports of sickness and medical consultation. An account of the use of the GHS for the analysis of inequalities in health is provided in Appendix 1.

2.52 There are problems in interpreting this type of morbidity data. The subjective experience of ill-health is framed by customary expectations and by the degree of inconvenience and cost attached to occupancy of the sick role. Even pain which we might assume has a certain objective reality independent of culture, is influenced by the perceptions, the consciousness and the bodily reaction of the sufferer. Indeed some would argue that a medicalised civilisation like our own has lowered the threshold of pain by heightening the fear of disease, by increasing the need for pain avoidance while at the same time providing chemical potions which anaesthetise bodily sensation. For these cultural reasons, and others, which must include individual physiology or psychology and the structure of social organisation there is a variability in the pattern of individual human response to sickness which complicates any attempt to examine the relationship between subjective and objective experiences of ill-health and understand the incidence of self-reported sickness by class or any socially defined status group.

2.53 The GHS data indicate that the personal experience of sickness tends to be greater amongst the socially and materially deprived especially during middle and old age. How are we to interpret the excess of self-reported sickness in the households of partly skilled and unskilled workers? Do the observed class differences reflect group comparisons of the same phenomenon? Are professional people more likely or less likely to respond to the symptoms of illness as a deviation from normal experience? Are they more likely to adopt the sick role and restrict normal social and economic activity? The answers to these questions are unclear, not least because there is no universally agreed definition of sickness outside of medically defined disease which allows controlled comparison to be made in the context of everyday life.

2.54 From 1971 to 1976 information on the distribution of 2 basic forms of ill- health was collected. These 2 were ‘long-standing illness’ and ‘sickness causing a restriction of activity during a two-week reference period’. In everyday language these 2 forms might be referred to as chronic and acute sickness respectively. In addition there were data on the self-reported rate of consultation with general practitioners. The data show that rates of sickness and the use of medical services varies in relation to a number of variables including sex, age, SEGs and region. In what follows the GHS data for 1974, 1975 and 1976 are aggregated so as to raise the sample size and improve the significance of the results.

Childhood Morbidity

2.55 In childhood the rates of reported morbidity do not follow the linear relationship with occupational class or, strictly, the closely associated socioeconomic grouping which is characteristic of the mortality rate. In general the incidence of both long-standing sickness and restricted activity is greater among boys than girls and this is likewise reflected in a greater tendency to consult general practitioners. In Table 2.12 comparison of socioeconomic group 6 with the average for all socioeconomic groups tends to show higher rates for the former for girls but only in the case of long,-standing illness for boys. At least in 1974-76 there were relatively low rates of restricted activity and consultation among boys in unskilled manual households.

TABLE 2:12 Sickness and Medical Consultation in Childhood (0-14 years) (rates per 1,000 population)

Restricted activity Consultation
Socio-economic group Boys Girls Boys Girls Boys Girls
1. Professional 91.5 74.7 105.4 73.2 104.0 83.8
2. Managerial 105.8 72.9 95.3 71.7 94.3 89.1
3. Intermediate 94.9 77.9 98.1 88.8 93.8 106.1
4. Skilled manual 101.4 81.4 88.2 77.5 86.6 87.2
5. Semi-skilled manual 108.0 70.6 87.0 75.0 98.5 76.9
6. Unskilled manual 123.3 105.8 58.3 83.3 76.2 92.3
All 102.5 78.1 90.5 77.8 91.7 88.5
Ratio 6/All 1.20 1.35 0.64 1.07 0.83 1.04

Source: GHS, 1974-76.

2.56 The results are clearly uneven, and the familiar gradients depicting correlations between health and class cannot be distinguished. Indeed reported sickness is in the case of boys more marked and in the case of girls nearly as marked in the homes of professional and non-manual workers than in the homes of unskilled workers, Part of the problem of interpreting rates of ‘restricted activity’ depends upon the interpretation of rates of ‘long standing illness’. Among the working class more of those suffering from the latter probably also suffer bouts of restricted activity which may or may not be reported. Conclusions have to be carefully qualified.

2.57 The lack of any clear class relationship in the distribution of self-assessed sickness in childhood and with it the somewhat lower rates of consultation with doctors is somewhat surprising, in view of the pattern of inequality found in mortality rates. This is especially so in the case of acute sickness such as infective or parasitic disorders which one might reasonably expect to be more prevalent in the more crowded environment of the typical working class home. (Douglas and Bloomfield, 1958). Chronic disorders do appear to be reported more frequently among the offspring of manual workers but even here there are exceptions and the higher incidence does not produce a concomitant rate of medical consultation. The problem of the 2 measures of morbidity is that they are expressed as rates per 1,000 population and not also in terms of severity or intensity, and similarly the crude consultation rate is not broken down in relation to the 2 measures of morbidity and further expressed in frequency.

Morbidity in Early Adult Life

2.58 Early adulthood is demarcated here as the phase of life between 15 and 44 years. For women this stage in the lifetime is one which brings them into close and frequent contact with the medical profession because of their biological role in human reproduction. This fact is reflected in the higher rate of consultation among women in all socioeconomic groups. During this time of life females also exhibit higher rates of restricted activity. This pattern of female dependence on health care, accompanied as it is by the higher rate of acute sickness, does not seem to be related to class. If anything, the wives of professionals and of other categories of non-manual worker experience a greater restriction of normal activity through sickness as well as being more prone to consult with general practitioners. The pattern is not repeated among men although linear class gradients are not clearly in evidence for either acute sickness or medical consultations. Occupational class is nevertheless an important source of differentiation. Male manual workers are more prolific users of the general practitioner service and they tend to report a greater than average incidence of restricted activity on account of acute sickness. The difference in rates of consultation is no doubt partly a reflection of the degree of flexibility which workers in different classes have in making arrangements to withdraw from work. Manual work is more likely to require that absence on account of sickness be sanctioned by a medical practitioner and this expediency, if nothing else, must help to swel1 the ranks of male patients in surgery waiting rooms during the working week.

2.59 For both sexes, long-standing sickness varies systematically by class. In each case, the incidence in socioeconomic group 6 is considerably higher than in group 1 and the occupational categories in between these extremes are also hierarchically arranged in a manner which accords with the expected gradient of disadvantage. This form of chronic sickness occurs with about the same frequency for both sexes in each group.

TABLE 2:13 Sickness and Medical Consultation in Early Adulthood (rates per 1,000 population)

Longstanding illness Restricted Activity Consultation
Socioeconomic group Males Females Males Females Males Females
1. Professional 145.4 138.2 84.0 106.4 75.5 140.4
2. Managerial 149.7 141.9 63.1 93.3 61.3 133.9
3. Intermediate 164.0 145.4 85.1 105.5 72.3 130.6
4. Skilled manual 161.9 167.2 89.7 95.2 85.1 142.5
5. Semi-skilled manual 173.8 170.3 81.5 99.3 80.5 146.0
6. Unskilled manual 197.4 202.3 110.4 95.3 93.5 145.9
All 163.2 157.8 84.5 99.0 78.7 138.6
Ratio 6/All 1.21 1.30 1.31 0.96 1.19 1.04

Source: GHS, 1974-76

Morbidity in middle age, 45-64

2.60 In the 2 decades before retirement occupational class exerts an important influence on all rates of morbidity and consultation among men and to a lesser extent among women also. On each of the 3 health indicators depicted in Table 2:14t class gradients are to be found. In general, rates of longstanding sickness are higher amongst men excepting for those in professional or managerial occupations who report lower rates than their wives (who nevertheless exhibit considerably lower rates than other women in the population). At this phase of the lifetime, chronic sickness is quite prevalent, affecting more than 3 out Of 10 men and women on average and in unskilled manual households the percentage increases to 50 for males and to 40 for females. Chronic disorder of the kind which resists treatment and cure but instead persists as a routine discomfort to the individual must be one of the most unfavourable dimensions of middle age in the working class.

2.61 Restricted activity on account of sickness is less prevalent. It affects about 1 in 10 on average, a rate which compares quite favourably with early adulthood. Sex differences are small and the differentiating influence of occupational class is more evident among males. Much the same pattern is found for rates of consultation. Professional and managerial males have the lowest rates of all and the highest is found amongst the unskilled manual category of men. For all 3 health indicators in Table 2.14, SEGs 1 and 2 tend to exhibit a markedly lower incidence than their counterparts lower down the socioeconomic scale.

TABLE 2:14 Sickness and Medical Consultation in Middle Age (45-64) (rates per 1,000 population)

Longstanding illness Restricted Activity Consultation
Socioeconomic group Males Females Males Females Males Females
1. Professional 228.9 291.3 71.1 92.2 75.6 94.7
2. Managerial 257.0 265.7 75.4 77.0 74.8 99.8
3. Intermediate 368.0 329.7 98.4 94.6 122.1 122.4
4. Skilled manual 357.7 315.1 102.6 102.7 112.4 109.2
5. Semi-skilled manual 387.6 380.8 101.0 114.9 124.9 121.5
6. Unskilled manual 485.5 401.6 120.0 111.9 145.5 122.6
All 348.6 329.4 96.5 99.8 110.0 113.6
Ratio 6/All 1.39 1.22 1.24 1.12 1.32 1.08

Source: GHS, 1974-76

Morbidity in Old Age

2.62 As one might predict, rates of chronic sickness are extremely high for both sexes after 65, the customary age of male retirement. At least half the population in this age group report some form of long-standing sickness, although among retired professional workers and their wives, the percentage is lower. For the rest of the population, males and females alike, the effects of ageing, which seem to have become quite common among socioeconomic groups 5 and 6 during middle age, become a routine feature of personal experience. At this stage in the lifetime females report the higher incidence of long-standing sickness and, not only is there a much lower rate in group 1, but there appears to be some class gradient. Much the same might be said of class differences amongst males, and males who have retired from manual occupations show a somewhat higher propensity to long-standing sickness than do their white-collar counterparts.

2.63 The remaining 2 health indicators taken from the GHS for 1974-76 which are presented in Table 2:15 show little, if any, evidence of distinctive class divisions. Restriction of activity amongst the aged population is more frequently reported by the higher classes, perhaps, in part, a reflection of their lower rates of chronic sickness and their higher expectations. Amongst the over-65s women report higher rates of restricted activity than men in every class. These sex and class differences are reflected in rates of consultation. Older women outnumber older men at the doctor’s surgery and the most infrequent attendees are those with the highest age – specific risk of mortality, men retired from unskilled manual work.

TABLE 2:15 Self-reported Sickness and Medical Consultation in Old-Age (65+) (rates per 1,000 population)

Longstanding illness Restricted Activity Consultation
Socioeconomic group Males Females Males Females Males Females
1. Professional 400.0 376.2 90.9 148.5 133.3 158.4
2. Managerial 476.6 525.6 93.6 109.3 157.3 164.6
3. Intermediate 503.2 553.4 94.2 121.5 121.2 142.7
4. Skilled manual 541.9 556.4 105.7 120.0 133.2 145.5
5. Semi-skilled manual 549.8 592.4 99.6 142.0 140.0 173.7
6. Unskilled manual 542.5 586.2 83.0 124.0 109.3 154.8
All 521.0 564.6 98.0 126.8 134.0 156.0
Ratio 6/All 1.04 1.04 0.85 0.98 0.81 0.99

Source: GHS, 1974-76

2.64 Up-to-date evidence relating to class differences in hospital treatment unfortunately cannot be added (see Chapter 4). Hospitals absorb much of the resources of the National Health Service in terms of both finance and skilled manpower. The occupational or other socioeconomic characteristics of patients receiving treatment are not analysed in the Hospital In-Patient Enquiry. Past research, however, has suggested that there are substantial class differences in the benefit which accrues to individuals on account of hospital episodes. Ferguson and Macphail (1954) in a survey of 700 male patients completed during the first decade of the NHS found differences in the numbers of patients of different occupational status whose health condition improved following hospital treatment and discharge. Whilst extrapolation over a period of 30 years is of uncertain validity here, their study remains worth quoting.

2.65 It is not surprising that men doing different kinds of work and receiving varying levels of economic reward for their labour receive different levels of benefit from medical therapy. While it is possible for the health care professionals to provide equality of treatment in hospital, what they cannot do is equalise the domestic and occupational circumstances of the patients they discharge. Non-manual work is, on average, less physically demanding, more secure and better rewarded than manual employment and therefore more compatible with the event of sickness and short or even long term physical impairment. Ferguson and McPhail observed clear differences in the numbers of men in different occupational groups who were able, after discharge, to return to their old jobs or to find alternative “suitable” work. (see table 2:15) Manual workers, especially the semi-skilled or unskilled, were much more likely to find themselves made redundant by the event of sickness, and the consequent loss of income and self-esteem can have only added extra burdens to the problems of recovery. As the authors of this study perceptively concluded,

“The transition from the sheltered atmosphere of the modern hospital ward to the icy chill of the workaday world is indeed a testing time and it is not surprising that many soon break down. The ex-patients who showed the heaviest mortality at early ages, the strongest tendency to relapse and the poorest record in point of early return to work were the group of unskilled labourers and it is significant that – apart from those suffering from such conditions such as advanced malignant disease – the proportion of men back in employment after leaving hospital was even more closely related to the nature of employment and home conditions than to the estimate made by the medical staff at the time of the men’s discharge from hospital. In many cases early recurrence of breakdown came of bad social and environmental conditions rather than any inevitability on medical grounds.”

TABLE 2:15 Occupation and failure to maintain improvement

Occupation group % failing to maintain improvement
Manual skilled
Manual semi-skilled
Manual unskilled

Source: Ferguson and McPhail (1954: pp. 137-8)


2.66 There are marked inequalities in health between the social classes in Britain. In this chapter mortality rates are taken as the best available indicator of the health of different social, or more strictly occupational classes and socioeconomic groups. Mortality tends to rise inversely with falling occupational rank or status, for both sexes and at all ages. At birth and in the first month of life twice as many babies of unskilled manual parents as of professional parents die, and in the next 11 months of life 4 times as many girls and 5 times as many boys, respectively, die. In later years of childhood the ratio of deaths in the poorest class falls to between 1 1/2 and 2 times that of the wealthiest class, but increases again in early adulthood before falling again in middle and old age.

2.67 A class “gradient” can be observed for most causes of death and is particularly steep for both sexes in the case of diseases of the respiratory system and infective and parasitic diseases.

2.68 Other aspects of class than merely occupational category have an impact on health although few data relating mortality to education, income etc are available. This is however illustrated by evidence that in all classes owner occupiers have lighter mortality than those paying rent.

2.69 Available data on (self reported) morbidity tend to reflect those on mortality. Rates of “long standing illness” (as defined in the GHS) rise with falling socioeconomic status and tend to be twice as high among unskilled manual males and about 2 1/2 times as high among unskilled manual females as males and females respectively in the professional classes. Inequalities are smaller in childhood and early adulthood and larger in middle age. If severe or “limiting” long-standing illness is isolated from long-standing illness then the poorer groups are found to be at a still greater disadvantage. Rates of sickness absence from work are also widely unequal.

2.70 On the other hand measures of “restricted activity” (regarded as a rough index of acute or short-term ill-health) are less unequal (or less unequally reported) between classes. For most years of the 1970s for which data were collected there was a class gradient but it was either less steep or uneven, and for children (particularly boys) there tended either to be no gradient or (in some years) an inverse gradient. Part of the problem of measuring ‘restricted activity’ is of course whether some, or a large number, of those already categorized as having ‘limiting long standing illness’ and not saying that their ‘normal’ activities were further restricted in a preceding period of 14 days, should be included. There remain of course problems about interpreting self-reported sickness and especially in judging whether the same conditions are as likely to be reported by some occupational groups as others.


BUTLER, N. R. and Bonham, D.G.  Perinatal Mortality, E & S Livingstone Ltd., Edinburgh, 1963

CHAMBERLAIN, R.  et al British Births   1975

DOUGLAS, J.W.B., and Bloomfield, J.M. , Children Under Five, London, Allen & Unwin, 1958.

DOUGLAS, J.W.B., The Home and The School, London, MacGibbon & Kee, 1964.

FERGUSON, T., MacPhail, A.N., Hospital and Community, London, Oxford University Press, 1954-

FOX, J.  “Household mortality from the OPCS Longitudinal Study” Population Trends. Winter 1978

GANS, B., “Health Problems and the Immigrant Child”, in CIBA Foundation, Immigration; Medical and Social Aspects, 1966.

GOLDSTEIN, H., “Factors Influencing the Height of Seven-Year Old Children – Results from the National Child Development Study”, Human Biology, Vol. .13, 1971-

HOOD, C., Oppe, T.E,., Pless, I.B. and Apte, E., West Indian Immigrants : A Study of One-year olds in Paddington, Institute of Race Relations,

KHOSLA, T., and Lowe, C.R., “Height and Weight of British Men”, Lancet, Vol. 1, 1960.

MILLER, F.J.W., et al, The School Years in Newcastle upon Tyne, 1952-1962, oup, 1974.

OPCS, Occupational Mortality 1970-72, Decennial Supplement, London, HMSO , 1978.

OPCS, The General Household Survey, 1976, London, HMSO, 1978.

OPPE, T.E., “The Health of Vest Indian Children”, Proc. Roy. Soc. Med. ,  57 1967, pp. 321-323.

RUTTER, M., Tizard, J and Whitmore, K. , Education, Health and Behaviour, London, Longmans, 1970.

SMITH, D. , The Facts of Racial Disadvantage  ; A National Survey, London, PEP, 1976.

TANNER, J.M. (ed) Human Growth , London, Pergamon, 1960.

Thomas, H. E., “Tuberculosis in Immigrants”, Proc.Roy.Soc. Med.,61

Todd, J.E., “Children’s Dental Health in England and Wales, 1973”, London, HMSO,   1975