Black Report 4 The Evidence for Inequality in Health Service Availability and Use

4.1 In this Chapter we review available evidence on class inequalities in health service use, both absolutely and in relation to need.

4.2 It is of course true that in seeking to explain class inequalities it health, which is our principal task, many factors other than health service usage prove relevant. As Martini, Allan, Davison and Backett (1977), amongst others, have shown differences in health outcome measures between populations may be far more a function of “variations in the socio-demographic circumstances of the population than [to] the amount and type of medical care provided and/or available”. Nevertheless, any inequality in the availability and use of health services in relation to need is in itself socially unjust and requires alleviation. This remains true whatever the proportional contribution which the health service makes to health, though its priority in social policy may well properly depend upon that proportionality.

4.3 Moreover since, as Cartwright and O’Brien (1976) point out, “One of the fundamental principles of the National Health Service was ‘to divorce the care of health from questions of personal means or other factors irrelevant to it’ (HMSO 1944)”, the extent to which this object has been achieved has been a matter of considerable interest. Thus, in 1968 Titmuss argued, on the basis of then available evidence, that inequality in receipt of care remained, that: “higher income groups know how to make better use of the Service; they tend to receive more specialist attention; occupy more of the beds in better equipped and staffed hospitals …….” (Titmuss, 1968). In 1969 Rein (participating in the American debate about finance of health care) argued, on the basis of different evidence, that the British Health Service is in fact equitable in the treatment provided (Rein, 1969). But his assumption was that need for health care was uniform between classes, and he did not relate utilisation to need. A number of subsequent studies many of which we refer to below, have sought to cast further light on the extent of social equality/inequality in availability and use of health services.

4.4 For our present purpose it is useful to look separately at GP consultation rates, hospital care, preventive services and services for the disabled and infirm. This is partly because such a distinction reflects the availability of information and the foci of research and also because (as we shall show) while some uncertainty remains as to the existence of inequalities in the first 2 cases there are no grounds for doubt in the ease of preventive services in particular.


4.5 The evidence linking class with rates of consultation with General Practitioners does not permit of simple interpretation. GHS data are available here, and are the best source of information. However as Table 4.1 shows, the trend, while clear enough for the sample of males and females in general, does not apply to males aged 0-4 or females aged 0-14 and is uneven for some other age groups. The GHS results have fluctuated from year to year and the consulting rate has not generally shown as marked a class gradient as have the measures of morbidity discussed in earlier chapters (see also Tables 2.11-2.14 for aggregated 1974-76 data). The average number of consultations has shown still less of a class gradient.

Table 4.1 Doctor Consultations: persons consulting a GP (NHS) in a 2 week reference period, by sex, age and socio-economic group rates per 1000

GB 1977

Socio-economic group Total 0-4 5-14 15-44 45-64 65+ Total 0-4 5-14 15-44 45-64 65+
Professional 69 [16] 74* 51 40* [6] 120 [11] 111 134 107 [3]
Employees/managers 102 250 107 69 93 132 116 166 79 129 109 127
Intermediate 104 206 107 69 93 132 116 166 79 129 109 127
Skilled manual 103 164 91 90 110 123 134 190 101 145 125 137
Semi-skilled manual 112 154 81 79 159 134 143 161 98 144 148 155
Unskilled manual 138 [6] 124 92 172 170 157 [8] 36* 216 160 161
All person in each sex/age group 104 181 91 79 115 137 133 173 92 138 128 150

Source: GHS 1977

i. bracketed figures are number of observations only where base figures less than 100

ii. *indicates less than 10 observations.

4.6 Picken and Ireland (1969) found that there was no significant relationship between class and level of consultation for adults in a study of Scottish general practice. However, “children from upper social classes and from smaller families tend to consult relatively more often than other children”. This may be taken as implying a greater inequality in use of services by children. It is however, also consistent with the view that among adults inequalities in need are greater, since here these outweigh the factors producing relative under-utilisation by adults on behalf of their children.

4.7 It has been argued that the proper basis for comparing rates of consultation is not one of simple population but of the need for care. Estimates of class trends on this basis have been made using GHS data. Brotherston, for example, divided number of GP consultations by number of restricted activity days (each in a 2 week reference period) – as given by the GHS – for each occupational group. His “use/need ratios” clearly declined in going from SEG 1 to SEG 6 (Brotherston 1976). Table 4.2 shows use-need ration calculated on the basis of the GHS for 1974-76 (without the indexing convention used by Brotherston) which similarly show an overall pattern of decline from SEG I to SEG 6.


GB 1974-1976

SEG Males Females
1 0.23 0.23
2 0.21 0.24
3 0.20 0.22
4 0.18 0.22
5 0.20 0.20
6 0.17 0.19
All 0.19 0.22

4.8 Forster (1976) used aggregated 1971 and 1972 GHS data and also found statistically significant trends in consultation rate/morbidity where the ‘morbidity’ measure was rate of chronic sickness or rate of sickness absence from work or school (but a non-significant trend in the case of acute sickness, ie restricted activity). He showed that in proportion to reported sickness and sickness absence from work, the semi-skilled and unskilled in fact made less use of GP services than other groups did. In a later review (1979) Forster found significant correlations between being unskilled and unemployed, on the one hand, and mortality and acute and chronic sickness on the other hand, but no significant correlation between environmental factors and health needs. Mainly negative but non-significant correlations were found between the need indicators and GP availability and between the proportion of unskilled persons and GP availability.

4.9 A weakness in this method of relating use to need derives from the fact that many of those with restricted activity may not visit a GP, whereas others may visit a GP for reasons other than restricted activity. In other words, comparison of these rates may not be purely indicative of differences in receipt of care when sick.

4.10 A different analysis of GHS data focusing upon those individuals reporting restricted activity or long-standing illness who went to a GP specifically For that illness has been attempted. For this purpose data for 1973-1975 and the first quarter of 1976 were aggregated, as were SEGs (1 and 2) and (4 and 5) (as numbers were small). It then appeared that, for both males and females, the percentages of those reporting restricted activity during a 2 week reference period who also visited a GP because of that restricted activity during the same period rose with declining SEG (see figs 4.1 and 4.2). However only about 13 per cent of those who reported a long-standing illness visited a GP during the 2 week reference period because of that illness. The small numbers do not permit any definite conclusion as to the trend by socioeconomic group. Two other qualifications are necessary. First, more of the semi-skilled and unskilled are likely to have both acute and chronic sickness, and restricted activity may thus be under-estimated among these groups. Second, people in the manual groups are likely to under-report their sickness, though less likely to under-report visits to the GP, thus elevating the apparent ratio.

Fig 4.1

Males seeing a GP for their restricted activity
Females seeing a GP for their restricted activity

4.11 Of course, comparison of rates of consultation alone is not a wholly adequate -conceptualization of inequality in care. Both Cartwright and O’Brien (1976) and Buchan and Richardson (1973) have studied GP consultations in depth. Both investigations showed that middle class patients tended to have longer consultations than did working class ones. More problems were discussed at consultations with middle class patients than with working class ones. Cartwright and O’Brien also found that middle class patients were, in a sense, able to make better use of the consultation time, as measured by the number of items of information communicated and the number of questions asked. Moreover even though working class patients tended to have been with the same practice for longer, the doctors seemed to have more knowledge of the personal and domestic circumstances of their middle class patients. In an earlier study Cartwright had found that middle class patients were more likely to be visited by their GP when in hospital than were working class patients (Cartwright, 1964). For cultural reasons then, and also (as discussed below) because there is a tendency for the ‘better’ doctors to work in middle class areas, the suggestion is that middle class patients receive a better service when they do present themselves than do working class patients.

4.12 Although the data are limited and further analyses remain to be carried out to bring out the full meaning of the 1971-1976 GHS data (like those of Forster and Brotherston), the Survey remains the best national source of information about the relationship between GP consultations and need for health care. As we have said we have strong reservations about the restricted nature of the indicators and about fluctuations especially for particular socioeconomic groups from year to year (see Appendix 5). Nevertheless the evidence can be taken to suggest that the level of consultation among partly skilled and unskilled manual workers does not match their need for health care (at least as implied by the GHS health indicators for all socioeconomic groups). Table 4.3 illustrates trends from year to year in the excess of ill health and GP consultatives of semiskilled and unskilled over professional groups.

4.13 The Table compares the rates of self-reported sickness with the corresponding rates of persons consulting and numbers of consultations for each of the years which data were collected. (The basic data are given in Appendix 5). It illustrates the yearly fluctuations – in part due to sampling – occurring in the findings especially in the relationship of sub-groups.

Table 4.3 Morbidity and GP attendance indicators for social-economic groups: Semi-skilled and unskilled as per cent professional

1971 1972 1973 1974 1975 1976
Long standing illness
Semi-skilled 158 163 157 160 157
unskilled 196 213 218 197 196
Limiting long standing illness
Semi-skilled 272 203 233 179 174 222
unskilled 371 290 333 274 244 292
Acute sickness (restricted activity)
Semi-skilled 126 133 110 134 102 80
unskilled 155 181 129 150 85 102
Acute sickness (restricted activity)
Semi-skilled 169 168 169 130 122
unskilled 268 206 215 121 196
GP consultations: persons consulting (rate per 1000)
Semi-skilled 122 132 121 146 122 91
unskilled 128 157 145 141 111 111
GP consultations: (rate per 1000)
Semi-skilled 133 132 125 146 129 91
unskilled 143 175 164 147 121 125
Long standing illness
Semi-skilled 274 214 182 197 176
unskilled 320 275 204 253 246
Limiting long standing illness
Semi-skilled 245 303 257 213 259 248
unskilled 298 355 332 246 346 348
Acute sickness (restricted activity)
Semi-skilled 105 128 115 115 134 95
unskilled 107 141 113 122 128 94
Acute sickness (restricted activity)
Semi-skilled 205 137 117 136 127
unskilled 238 148 122 129 133
GP consultations: persons consulting (rate per 1000)
Semi-skilled 113 147 110 118 116
unskilled 112 141 123 111 115
GP consultations: (rate per 1000)
Semi-skilled 108 150 110 123 114
unskilled 117 150 120 107 102

4.14 According to the available indicators, with one partial exception, (The relative consultation rates of semi-skilled and unskilled groups exceed the relative rates of restricted activity, though in most of those instances they no longer do so when days of restricted activity are taken) the observed socioeconomic or class differences in ill-health (long standing illness or disability, limiting long standing illness or disability, persons experiencing restricted activity in a 2 week reference period and number of days of restricted activity because of illness or injury – to which might be added absence from work for reasons of sickness though these were reported for only 3 years and so are not included) are larger than the corresponding group or class differences in GP attendance (rates for persons consulting and total number of consultations). If we were to combine different measures of chronic and acute ill health and compare “need” with “care received” (on the basis of some principle of weighting) then this on would become still stronger.


4.15 There is no regular source of class-related information on use of hospital services comparable with that obtained on GP consultation rates from the GHS, even though questions on hospital attendance are asked in the Survey. In the case of hospital inpatients, the percentage concerned (-2%) is too small for any class breakdown of GHS data to be statistically meaningful. Attendance at outpatients are higher and although since 1972 these rates have not been published on an SEG basis, information is available, as shown in Table 4.4. This suggests that there are no systematic class gradients in outpatient attendance for either males or females.


Great Britain. 1974-1977 data combined

Socio-economic group
0-14 15-44 45-64 65+ 0-14 15-44 45-64 65+
96.2 79.0 87.9 132.2 76.9 103.3 101.6 104.2
96.3 94.4 98.4 112.0 67.4 96.9 110.3 97.4
113.0 100.7 122.3 142.0 77.9 107.3 123.8 134.5
89.5 112.9 117.0 123.5 75.1 96.4 112.6 122.7
82.5 102.8 132.0 122.5 68.6 86.6 114.1 121.9
108.2 133.5 113.7 104.7 58.9 90.9 107.6 130.9
All groups
94.2 105.4 115.8 123.2 72.6 98.0 114.2 123.7

(Note that regular attendances for antenatal care are excluded. Note also that the low response rate (GHS rates are only about 2/3 those reported in Hospital Activity Analysis) may affect the age and SEG distributions).

Source: GHS (unpublished data)

4.16 Data given in the 1972 GHS Report seemed to suggest that men in social class V aged 15-64 had particularly high rates of attendance. After retirement this was no longer so. Referring to this “relatively much greater use of ‘out-patients’ by unskilled males than by males of other groups” “the Report indicated that attendance for consultant out-patient, casualty, and ancillary services could not be distinguished. However the decline in rate at retirement at least suggests that “the higher rates amongst unskilled males of working age than amongst males of other groups may reflect a rather particular use of out-patient facilities related to their greater risk of exposure to accident or injury compared with other groups” (GHS 1972 p.214). Though later data, presented in Table 4.4, do not show this discontinuity at retirement, the possibility of this special use of hospital out-patients is a matter of interest.

4.17 A study of a random sample of patients attending the accident and emergency departments of hospitals in the Newcastle-upon-Tyne area is relevant (Morgan et al 1974; Holohan et al 1975). The area was served by 3 hospitals, so far as accident and emergency services were concerned, and a random sample of 1% of those seen in a period of 3 months (except those subsequently admitted to hospital, or who were victims of road accidents), was interviewed. In fact, the overall class distribution of the sample (135 males and 97 females) was the same as the class distribution in the general population of the Tyneside conurbation – though this has a lower proportion of classes I and II than has the national population. 78.4% of the patients in the survey were self-referred, and there was no class gradient in extent of self-referral. Only 10% of patient had attended for reasons other than accidental injury. This study suggests that people self-refer themselves to A and E departments principally because they are seen as more “available” than the GP: they consider the organization of primary care (surgery hours, appointments system) makes it unlikely they will be seen promptly.

4.18 Such little evidence as is available on hospital in-patient care describes not the proportions of various social groups spending time in hospital but the social composition of the group admitted. For England and Wales, admission rates on an occupational basis were available from HIPE only prior to 1963, when data ceased to be centrally collected on that basis (mainly due to doubts as to the quality of the information). The older data which are available do, however, suggest that the rate of usage of hospital beds rises with declining class. In Scotland, Carstairs and Patterson (1966) analysed hospital admission rates and duration of stay data and found clear upwards trends in both with declining class. Moreover, SHIPS (Scottish Hospitals In-patient Survey) does continue to code social class, and more recent figures confirm this class gradient both in admission rates (given by SDR in Table 4.5) and length of stay (given by SBDR).


Class Males Females Males Females
I 79.5 95.9 63.7 92.5
II 80.9 98 73.3 93.6
III 94 90.4 93.9 91
IV 115.1 107.4 116.4 106.7
V 141.4 161.1 151.7 153.9

Source: SHIPS

4.19 A rough attempt at relating the composition of the hospital population to that of the population at large has been made in the Welsh Office using Welsh HA and census data. Though this was done only for a number of occupational categories, it suggests that extractive occupations (eg coal mining) and metal working occupations are greatly over-represented in the hospital population, compared to ‘other industrial occupations’ and (still more) compared to ‘clerical and sales’ occupations.

4.20 Against this, of course, has to be set the possibility that systematic differences exist in the nature, or quality, of care received within the hospital. While there is little evidence for this, a recent study of childbirth by Cartwright (1977) should be noted. This found that women from class V not only experienced the lowest degree of intensive care their pregnancies, but were very much less likely to have their babies induced in hospital (though the reasons for this are unclear as, indeed are its implications).


4.21 Although neither administrative returns nor the GHS provide information on utilization of community health and preventive services by socia1 or occupational class, there is here a substantial body of research upon which to draw. As we shall illustrate from this research, it is well established that those in the manual classes make considerably less use of these services than do those in the non-manual classes. Moreover, the ambiguity which surrounded the relation of utilization to ‘need’ in the case of GP consultations is not encountered here. Assessment of morbidity, or need for care, is not at issue in comparing rates of attendance for ante-natal care (though this may be more essential for those rendered at risk by social factors), cervical screening, radiography, or immunization of children. Even those who have been led to doubt the importance of medical care for improving health (relative to the importance of socioeconomic and demographic factors), such as Martini et al, have largely excepted these preventive services from their general conclusion.

4.22 In the case of family planning and maternity services substantial evidence shows that those social groups in greatest need make least use of services and (in the case of antenatal care) are least likely to come early to the notice of the service. Cartwright (1970) found clear class gradients in the proportion of mothers having an antenatal examination, attending a family planning clinic, and discussing birth control with their GP. Unintended pregnancies were more common among working class women. Bone (1973) also found that women from the non-manua1 classes make more use of family planning services than those from the manual classes. This was true both for married and for unmarried women. Brotherston has presented Scottish data showing that late antenatal booking is more common in poorer social groups, although the situation seems to be improving in all classes,(and has suggested that late presentation for antenatal care is an effective predictor of subsequent infant morbidity and mortality within families) (Brotherston 1976).

Table 4.6 Late Antenatal booking

% married women in each occupational class making an antenatal booking after more than 20 weeks of gestation

Scotland 1971-3

Occupational Class 1971 1972 1973
I 28.4 27.2 27.0
II 35.3 32.3 29.8
III 36.3 33.4 30.6
IV 39.3 37.8 35.3
V 47.1 44.2 40.5

Source Brotherston (1976)

4.23 Similar differences have been found in presentation for post-natal examination (Douglas and Rowntree, 1949) and (by Gordon, 1951) immunisation, ante-natal and post natal supervision and uptake of viatmin foods. Tha National Child Development Study (1958 birth cohort) found substantial differences in immunisation rates in children aged 7, as well as in attendance at the dentist.

Table 4.7 Use of health services by children under 7 by occupational class of father

GB 1965

% who had never visited a dentist 16 20 19 24 27 31
% not immunised against
smallpox 6 14 16 25 29 33
polio 1 3 3 4 6 10
diptheria 1 3 3 6 8 11

4.24 Among women, it has been found that those in classes IV and V are much less likely to be screened for cervical cancers even though mortality from this condition is much higher in these classes than in the non-manual classes. Table 4.8 taken from Sansom, Wakefield and Yule (1972), shows that while women from classes IV and V accounted for over one-third of all women living in the study area (Greater Manchester), they made up only about one-sixth of women who had a smear test done.

TABLE 4.8 Cervical Screening of women (all ages) by Occupational Class

% women in each class in 1966 compared with two populations of screened women: Manchester area

Class 1966 Census (NW Region less Merseyside 1965 profile of screened women (n=5000) 1968 profile of screened women (n=34,851)
1 2.6 6.5 7.7
2 15.5 19.3 20.1
3 48.2 57.3 57.0
4 25.4 10.6 11.1
5 8.3 6.3 4.1

Source: Sansom, Wakefield and Yule (1972)

4.25 Further studies show that working class people make less use of dental services (Gray et all 1970; Bulman et all 1968) and of chiropody. (Clarke, 1969) and receive inferior dental care (Sheiham and Hobdell, 1969). Many of these studies are admittedly old, and their findings cannot necessarily be accepted as still valid. Nevertheless, taken together, and in the absence of conclusions to the contrary, a clear relationship between social class and use of preventive services seems to have been demonstrated.

4.26 All these services have in common that the individual has to exercise his own discretion and initiative in obtaining the service. The class gradients reported may thus often compound differences in the personal choices or decisions made, in information possessed, and in the availability of services (as reflected in proximity access and quality). That is, to some extent at least the problem may inhere in geographical differences in provision between areas of largely middle and largely working class residence. We take up below this interaction between class and geographic disparities. What of those (admittedly few) services which have an ‘outreach’ capacity’? Do these show some tendency to compensate, or positively discriminate in favour of those in particular need? An important example is health visiting. Both Jeffreys (1965) and Cartwright (1970) found that health visiting did not seem to favour one class or another. Cartwright’s survey of mothers showed “no variation between mothers in the different classes in the number of times they had been visited by health visitors”, suggesting that there is no compensatory positive discrimination. However the picture may not be quite so bleak.

4.27 In a partial follow-up of 852 children from the 1970 birth cohort study who at age 3 ½ were resident in the SW of England and Glamorgan, Butler and his co-workers covered similar ground. First, like Jeffreys and Cartwright, they found no class differences in the percentages of children who had received a home visit from a health visitor in the first year or second and third year of life. Nor was there a difference in the percentages who had received 5 or more visits. However, when the data were reanalysed by their ‘social index’ (which takes account not only of fathers’ occupation, but also of amenities and crowding in the home, parental education, etc) pronounced differences appeared. Thus, in the first year, whereas 46.9% of those in the ‘advantaged’ or ‘neutral’ groups on this index received 5+ home visits, 63.8% of those ‘disadvantaged’ did so. The same difference persisted in the second and third years. It may be, therefore, that health visitors are responding to a more subtle notion of advantage/disadvantage than is captured by ‘occupation’ alone.

4.28 A further indication that ‘outreach’ can be effective in reaching those in need has been described by Epsom (1978) in an account of the use of a mobile health clinic in Southwark.


4.29 There is little information on class inequalities in the care received by the infirm and disabled, though we now enter that awkward and neglected area where health care shades into the variety of other forms of social service provision. That is, to make comparisons of the care, or services received by those who are disabled or infirm (including the aged infirm and the long-term chronic sick) would necessarily to be consider not only health care in the strict sense, but social work support, meals on wheels, home help, sheltered housing, mobility aids, sheltered work and rehabilitation etc etc: all or many of which may be crucial to the well-being of an infirm or disabled person. It is of obvious importance where the attempt is to be made to take relevant services to those in need without awaiting demand, to be aware of the extent of need and of use. The surveys carried out as a result of the Chronically Sick and Disabled Persons Act of 1970, though varying somewhat from one local authority social service department to another, give some indication of the characteristics of this substantial group of people and of their needs – met and unmet.

4.30 Social or more strictly occupational class does not figure prominently as a classificatory variable in these surveys. In one after another attention is drawn to the economic hardship under which the majority of disabled people live and their dependency on State benefits: eg in the Newcastle study (1972), well over 90% were “heavily dependent” on these benefits; only about 10% were receiving any income from employment, and that generally very little. This is despite the fact that among those who did, or had, worked the range in occupations was broad: III NM (eg clerical), IIIM (Skilled trades), V (labourers) were all substantially represented. Not only does social class, as defined by the Registrar General have little value as an indicator of the resources or way of life, of this group, but other classificatory variables (notably degree of handicap, or whether or not living alone) are seen as of greater importance in improving coverage. Blaxter, in her sophisticated study of a sample of people disabled in adult life (Blaxter, 1976), shows how the resultant problems (financial and other) as well as the ability to find solutions to these problems was a function of class. Yet even here, when receipt of services (eg home nursing, home help) is considered, class does not figure as a dimension of analysis.

4.31 The fact is that, in contrast f or example to current priorities in improving coverage of antenatal care, social class is not seen as a major dimension of unmet need. Without denying that extent of handicap is likely quite properly to remain paramount, it may well be that it is the narrowness of the current operational concept of class which is partly at fault here. Few would deny that class (as reflected not only in accumulated financial resources and, to some extent, in the availability of a supportive social network, but also in the narrow sense of occupation) mediates the effects of a given impairment.

4.32 The slight, and unfortunately somewhat old, empirical evidence available does indicate that class inequalities are indeed to be found here. Townsend, for example, in a study of the elderly in residential homes, geriatric and psychiatric hospitals carried out some years ago, found not only that the manual and non-manual elderly were likely to be in different kinds of institution (the latter less commonly in hospital), but of a gap in the standard of living and care available in institutions catering principally for one or the other group (Townsend, 1962). Townsend and Wedderburn, in a national study of people aged 65 and over, found disparities in receipt of a number of services. Tables 4.9 and 4.10 show how publicly provided chiropody services, and domestic help, fail fully to compensate for unequal ability to purchase such services (Townsend and Wedderburn, 1965).


Source of Chiropody treatment Occupational Class (%)
I II III non-manual III manual IV V
Public or voluntary service 2 6 6 9 8 8
Privately paid 20 18 14 8 9 8
Non-professional or none, need felt 6 8 13 11 13 10
Non-professional or none, no need felt 72 68 67 72 71 74
Total 100 100 100 100 100 100
Number 82 557 396 1193 1040 457

Source: Townsend and Wedderburn (1965)

Note a further 42 persons were classed in armed services occupations, 10 had no occupation and 290 were unclassifiable.

Table 4.10 Percentage of old persons of different occupational class who were receiving public or private domestic help, or said help was needed: Britain 1962

Source of domestic help Occupational Class (%)
I II III non-manual III manual IV V
Local authority 1 2 4 6 4 4
Privately paid 42 27 12 5 3 2
Other (eg family) or none, need felt 10 7 6 6 6 4
Other (eg family) or none, no need felt 47 64 78 83 87 90
Total 100 100 100 100 100 100
Number 81 555 396 1188 1033 447

Source: Townsend and Wedderburn (1965)

4.33 Equally a matter of concern is the difference in attitude between old people of different social classes which this work demonstrates. Table 4.10 shows that 90% of those in class V neither receive domestic help (from outside their family) or feel the need of it. Among the elderly who are incapacitated “nearly half those in Social Classes I and II …. already had privately paid or local authority domestic help, and nearly half the others said they needed help. But only a sixth of those in Social Class V who were severely incapacitated had such help already and only a fifth of the remainder felt the need for it” (Townsend and Wedderburn, 1965; 46). Whilst we cannot tell if these differences still exist, the study shows all too clearly the way in which norms and values associated with class may influence subjective perceptions of need.

4.34 In considering the needs of the disabled, aged infirm, and chronic sick, not only is it difficult clearly to distinguish needs for strictly medical services from needs for other supportive services, but it is similarly difficult, to distinguish needs related to the condition ‘itself’ (ie medically defined) from those relating to its social and economic consequences. In this context an inquiry into the circumstances of the long-term sick (ie receiving sickness, invalidity, and industrial injury benefits for periods of between 1 month and 1 year) is of relevance (Martin and Morgan, 1975). This shows not only that the sample of benefit-recipients as a whole contained a much higher proportion of semi-and unskilled manual workers (which was to have been expected), but that the longer the spell of sickness, the higher the proportion of unskilled workers. Moreover, and relevant to this section of our report, the longer the period of incapacity the less likely the sick person is to be able to return to the same type of work with the same employer as before. Skilled manual workers are more likely to have their jobs kept open for them than are semi- or unskilled workers. Unsurprisingly also, receipt of sick pay from the employer is also related to duration of invalidity and to level of job (SEG). Clearly, in considering the needs of the disabled, long-term sick, and aged infirm, financial problems and (in some cases) problems of subsequent re-employment are both pertinent and class-related. The implication of this for policy (which we take up in Chapter 9) is that the equalization of provision of health and social services is not justifiably separable from the equalization of the social and economic consequences of long-term invalidity or sickness.


4.35 It has long been known that differences are to be found between the health services available in well-to-do areas and in poor areas. Thus, in 1957 Martin, in a study of social aspects of prescribing, found that the average total cost of drugs per prescription was higher in wealthier areas. Noyce, Snaith and Trickey (1974) analyzed health expenditure at the regional level, and found a positive correlation between the percentage of the population in professional and managerial socio-economic groups and both community health expenditure and hospital revenue expenditure, and a negative correlation between expenditure and proportion of population in unskilled and semi-skilled occupations. They concluded “There are no regions of above-average spending which are not also high socioeconomic status regions. Indeed, if one knew no other facts it would be possible to explain two thirds of the variation in community health expenditure by a knowledge of what proportion of the population in each region were managers, employers, or professional workers”.

4.36 In 1971 Tudor Hart contrasted the availability of medical care in poor industrial areas of high need and affluent salubrious areas of lower need in memorable terms (Tudor Hart, 1971). He wrote:

“In areas with most sickness and death, general practitioners have more work, larger lists, less hospital support and inherit more clinically ineffective traditions of consultation than in the healthiest areas; and hospital doctors shoulder heavier case-loads with less staff and equipment, more obsolete buildings and suffer recurrent crises in the availability of beds and replacement staff. These trends can be summed up as the inverse care law: that the availability of good medical care tends to vary inversely with the need of the population served.”

4.37 West and Lowe (1976) analyzed data on need for and provision of child health services for each of the 15 pre 1974 hospital board regions of England and Wales, collected for the Court Committee. Some of their findings are given in Table 4.11 which shows how, in particular, regional provision of GPs and health visitors is negatively correlated with a number of indicators of need (including stillbirth rate, level of infant mortality and birthrate to teenage mothers). There are very few positive correlations in the Table: midwives alone seemed to be relatively well distributed throughout the regions. The authors go on to suggest: When data become available for area health authorities even greater differences between need and provision will probably be uncovered between areas than between regions (CF Morris 1975 pp 53, 77). In the meantime it has at least been established that variations in expenditure (Richard.1976) and in the provision of services (Buxton and Klein 1975; Jones and Masterman 1976) are greater at the sub-regional than at the regional level.

4.38 Differences in the availability of services to the various social groups are likely to be more adequately expressed in terms of variation in provision between relatively small geographical areas. As Jones & Masterman point out “The knowledge that he or she lives in a relatively well provided region is of little consolation to the patient unable to enjoy satisfactory facilities in his or her locality”. The most satisfactory level of disaggregation may well depend upon the particular service under consideration, being in some way a function of the catchment area of the service or facility. Thus, insofar as our principal concern here is with primary and community care, the crucial question becomes that of whether small typically working class areas are less well provided with health services than are small areas of typically middle class residence.

Table 4.11 Correlation Co-efficients between indicators of need and provision: Child Health – related services in 15 pre-1974 RHBs (Statistically significant coefficients only)

No. paediatric medical staff No. paediatric beds No. obst/gyn medical staff No. special care baby units No. midwives No. GPs No. HVs No. LA med. staff No. school nurses
Live birth rate -0.44 +0.47 -0.62 -0.49 -0.54
birth rates, mothers 15-19 years +0.75 -0.73 -0.79
birth rate, mothers 35-44 years +0.56
low birth weight rate +0.45
% population <15 years +0.62 -0.63 -0.60
stillbirth rate +0.48 -0.52 -0.74
infant mortality -0.65 -0.50
mortality:1-4 years
mortality:5-14 years +0.50

>0.44=P<0.05, >0.59=P<0.01

Source West and Lowe (1976)

4.39 Whilst it is not possible to give a clear answer to this question, an exploratory study by Skrimshire (1978) is usefully indicative. This involved interviews with samples drawn from 3 areas of council housing :in Newham (East London) and from 2 homogeneous areas in the Midlands: One solidly middle class, the other a council estate with a history of social problems. Unlike Newham which has long suffered from severe deprivation of the environment both Midlands areas were situated in “a Socially mixed county with a teaching hospital, and an attractive environment: a highly desirable county in which to work, with no reputation for recruitment problems in general practice” (unlike Newham). Questions covered morbidity (GHS questions were used) and experiences in seeking help from the GP. Although much was found to be common to the 2 working class areas, differentiating their inhabitants from the middle class residents of the third area, this was not the whole picture. Indeed, the principal conclusion of the study is as follows:

“The provision of health care and the subjective experience of seeking that care are all partly determined by the socio-economic structure of society on an area basis so that a working class person is at a greater disadvantage if he lives in a predominantly working class area than if he lives in a socially mixed area. The data … are consistent with a theory of structural determination of need and demand for health care from an area, operating both through environmental and social conditions on the level of health, and through the social processes and life experiences that further affect demand, particularly in the case of childhood illness. The level and quality of available medical manpower, relative to need and demand, is likely also to be strongly affected by the environment and social class composition of an area through the operation of the market for recruitment”.

4.40 It is likely that similar conclusions would follow from a consideration of race or ethnicity. However, information on use of services by ethnic groups is sparse. Coombe has referred to hesitation in seeking ante-natal care among immigrants, and their difficulties in securing adequate dietary information (Coombe, 1976). There is evidence of some lack of appreciation among health services staff of the special needs of some immigrant groups, as well as a clear lack of adequate facilities in some of the areas in which they have been obliged to concentrate.


4.41 Generalization about inequality of utilization is made difficult partly because of sampling errors in the case of national surveys and of partial information in the case of local studies, and because of the (as yet unresolved) problem of relating utilization to need.

4.42 Inequalities appear to be greatest (and most worrying) in the case of the preventive services. Severe underutilization by the working classes is a complex resultant of under-provision, of costs (financial, psychological) of attendance, and perhaps of a life-style which profoundly inhibits any attempt at rational action in the interests of future well-being. Such factors are not, in this case, outweighed by the costs of present disruption of normal social functioning. We have also seen, however, how services provided on an ‘outreach’ basis can serve to reduce at least some of the costs of attendance, with beneficial results.

4.43 The situation is not clear cut in the case of GP attendance, partly because attendance rates cannot be compared with any precise measure of need. Excepting, for some years, children, more of those in lower than in higher SEGs consult a doctor, and their total consultations are, relatively greater. But on most of the health indicators their need for care is greater still. It is hard not to conclude that poorer groups make relatively low use of GP services, irrespective of the separate question of the adequacy of the services to which they typically have access.

4.44 Middle class parents are, however, more likely than working class parents to seek medical attention for their children. Since the (direct) costs of attendance may be presumed similar, this may imply that working class adults are likely to be typically more sick than are middle class ones before help is sought. Moreover, we have seen also that middle class patients typically receive better care from their GP – a consequence, once more, of both interpersonal and ecological factors.

4.45 Hospital outpatient departments are used more by the working class than by the middle class. In the case of Accident and Emergency departments there is evidence of some use in place of the GP, access (as to the GP) being on the basis of self- (or “lay”) referral. It has been suggested that this preferred use of outpatient in treatment of “traumatic’ conditions (suffered at work, or in the home) is principally a result of their greater availability: they are open 24 hours a day, no appointment is needed, and availability of diagnostic aids is certain. Could it be that here the working class patient feels better able to count on good, and equal, treatment?

4.46 In the case of inpatient departments too evidence suggests greater use by the working class. It may be noted that, since admission is generally on the basis of GP referral, a higher proportion of working class patients than of middle class patients consulting a GP must be subsequently admitted to hospital. This in turn must imply that more working class patients have illnesses requiring hospital admission, or that the working class patients seeing his GP is typically sicker and/or that he or she is seen as less likely to receive adequate care at home. And indeed, evidence from a survey of the elderly suggests that this is so: public provision of domiciliary services seems not fully to compensate for differential ability to purchase such services.

4.47 It is hard to resist the conclusion that this pattern of unequal use is explicable not in terms of non-rational response to sickness by working class people, but of a rational weighting of the perceived costs and benefits to them of attendance and compliance with the prescribed regime. These costs and benefits differ between the social classes both on account of differences in way of life, constraints, and resources, and of the fact that costs to the working class are actually increased by the lower levels and perhaps poorer quality of provision to which many have access.

4.48 Class differentials in use of the various services which we have considered derive from the interaction of social and ecological factors. Differences in sheer availability and, at least to some extent, in the quality of care available in different localities provide one channel by which social inequality permeates the NHS. Reduced provision implied greater journeys, longer waiting lists, longer waiting times, difficulties in obtaining an appointment, shortage of space, and so on. A second channel is provided by the structuring of health care institutions in accordance with the values, assumptions and preferences of the sophisticated middle class ‘consumer’. Inadequate attention may be paid to the different problems and needs of those who are less able to express themselves in acceptable terms and who suffer from lack of command over resources both of time and money. In all cases, for an individual to seek medical care, his (or her) perveptionof his (or her) need for care will have to outweigh the perceived costs (financial and other) both of seeking care and of the regime which may be prescribed. These costs are class related.

4.49 It is the interaction of these two sets of factors which produces the inequalities documented in this chapter.


Blaxter, M. The Meaning of Disability (Heinemann, 1976)

Bone, M. Family Planning Services in England and Wales (HMSO 1973

Brotherston, J.     “Inequality:  is it inevitable?” in Carter and Peel (eds Equalities and Inequalities in Health (Academic Press 1976)

Buchan, I. C. & Richardson, I. M,   Time Study of Consultations in General Practice (Scottish Home and Health Department, 1973

Bulman, J. S., Richards, N. D., Slack, G. L., Willcocks, A J – Demand and Need for Dental Care (OUP, 1968)

Butler, H et al Community and family influences on 0-5s: Utilization pre-school day care and preventive health care (unpublished, 1977)

Buxton,  M J & R E Klein      “Distribution of hospital provision:  policy themes and resource variations” BMJ1(1976)  299

 Cartwright, A Human Relations and Hospital Care (Routledge,   1964)

Cartwright, A Parents and Family Planning Services (Routledge, 1970)

Cartwright, A The Dignity of Labour (Tavistock 1979)

Cartwright, A & M O’Brien  “Social Class variations in health care” in M Stacey (ed) The Sociology of the NHS (Soc Rev Mon 22 1976)

Carstairs, V & P E Patterson  “Distribution of hospital patients by social class” Health Bull 24 (1966)

Coombe, V. “Health and social services and minority ethnic groups” J Royal Soc   Health 96 (1976)

Davie, R, Butler, N & H Goldstein    From Birth to Seven (Longman 1972)

Douglas, J W B & G Rowntree  “Supplementary maternal and child health services” Pop Studies 2 (1949)

Epsom, J D ‘The Mobile health clinic: a report on the first year’s work” in D Tuckett & J M Kaufert Basic Readings in Medical Sociology Tavistock 1978)

 Forster, D P “Social class differences in sickness and general practitioner consultations” Health Trends 8 (1976) 29

Forster, D P “The relationship between health needs, socio-environnn indices, general practitioner resources, and utilization” J Chronic Dis (1979) 32 4 333

General Household Survey Reports 1971-1977

Gordon, I           “Social status and active prevention of disease” Mon Bull Min Health 10 (1951)

Gray, P G et al       Adult Dental Health in England and Wales in 1968 (HMSO 1970)

Jefferys, M An Anatomy of Social Welfare Services (Joseph 1965)

Jones, D R & S Masterman    “NHS Resources:    scales of variation” Brit J Prev Soc Med 30 (1976) 244

Martin, J & M Morgan    Prolonged Sickness and the Return to Work(HMSO,   1975;

Martini, C J M et al   Health indexes sensitive to medical care variation Int J Health Services 7  (1977) 293

Morgan, W., Walker, J. H., Holohan, A M and I T Russell      “Casual attenders” Hospital and Health Services Review 70 (1974) 189

Morris, J N  The Uses of Epidemiology (London, Longman 1975)

Newcastle, City and County of Report on the Survey of Chronically Sick and Disabled Persons Resident in Hewcastle-upon-Tyne (1972)

Noyce, J., Snaith, A. H. & A J Trickey  “Regional variations in the allocation of financial resources to the community health services” Lancet (1972)

Rein, M “Social class and the utilization of medical care services” J American Hospital Assn 43 (1969)

 Picken, B & G Ireland “Family patterns of medical care utilization. Possible influences of family size, role and social class on illness behaviour” J Chronic Dis 22 (1969)

Rickard, J H “Per capita expenditure of the English area health authorities BMJ J (1976) 299

Sansom, C D, Wake field, J & E Yule  “Cervical cytology in the Manchester area: changing patterns of response” in J Wakefield (ed) Seek Wisely to Prevent (HMSO. 1972)

Sheiham, A & M H Hobdell “Decayed, missing and filled teeth in British adult populations” British Dental Journal 126 (1969) 401

Shrimshire, A Area Disadvantage, Social Class and the Health Service Department of Social and Admin Studies, University of Oxford, (1978)

 Titmuss, RM, Commitment to Welfare (London: Allen and Unwin, 1968)

Townsend, P The Last Refuge (Routledge, 1962)

Townsend, P and D Wedderburn  The Aged in the Welfare State (Bell Occ Papers on Social Administration, 1965)

J Tudor Hart “The inverse care lawLancet 1 1971

West R R and Lowe C R., “Regional variations in need for and provision and use of child health services in England and Wales” BMJ (Oct 1976) 843