Black Report 3 Trends In Inequality Of Health

1. Previous chapters have set out the problems of defining and measuring inequalities of health and have described contemporary inequalities. In this chapter we will attempt to draw conclusions about trends, especially during the last 30 years. For want of other data of a comparable kind covering a span of decades rather than a few years, mortality data will be given most attention. After a brief overview of general changes we will discuss in turn the data for men of economically active ages, women of economically active ages, children and elderly people.

DECLINE IN DEATH RATES

2. For about 100 years mortality rates for both sexes, taking one decade with the next, have declined. Figure 3.1 shows the trends, after discounting for changes that have taken place in the age-structure of the population. It will be seen that the rates for males have remained markedly higher than for females, and in recent decades the difference has become relatively greater. This is shown more clearly in Figure 3.2. Even since 1946 the excess of the male over the female rate has increased at all ages – and especially between the ages of 10 and 30. Although attempts have been made to explain the difference between the sexes comparatively little systematic work exists. Research has often concentrated on explaining the health experience of a single sex. We consider that this is unsatisfactory whenever it would be relevant to study the experience of both sexes. We also consider that as a consequence some social inequalities may have been overlooked or minimised. Even when comparisons have been undertaken they have sometimes been made crudely, perhaps because of the shortage of well-grounded studies. For example, an MRC report took the view that mortality from bronchitis among men in the coal industry owed little to direct occupational effects, and more to general socio-economic or environmental factors, because a high correlation between the bronchitis SMRs for men and those for their wives had been observed (Medical Research Council, 1966, pp 101-102). But others pointed out that the exposure of some working wives to comparable occupational hazards as well as the exposure of families to some of the dust, and the effects of dust, picked up by the men, seemed likely to have been underestimated. (McLaughlin, 1966, and Higgins, 1959; as quoted by OPCS, 1978, P-32).

Figure 3.1 Death Rates (standardised to 1901 population): England and Wales

Source: McKeown, T. The Role of Medicine, London, Nuffield Provincial Hospitals Trust 1976 p.30

Figure 3.2

Historical change in the risk of death for males and females throughout the lifetime

3.3. The aetiology of the mortality and morbidity differences between men and women remains to be disentangled. Women suffer uniquely from some diseases: there have been major changes in the last 50 years in their experiences of childbearing and of paid employment, and there have been changes too in the pattern of personal, married and family life outside employment. Such factors must be assumed to play some part. Certainly the amount of waking time spent outside paid employment has increased, and more families spend parts of the year on holiday far away from their local environment. It would be wrong, for example, to assume too readily that all wives share the same living conditions or even standards as their husbands. Some men have the advantage, for example, not only of a preferential diet at home but subsidised meals at work. Where both husband and wife are in paid employment the meals they get in the day, as well as working conditions and the nature of the work, may be radically different. These general reservations have to be borne in mind in examining certain types of mortality date – as for example correlations between SMRs for occupational orders for married men and their wives; and between cause-specific SMRs for married men and those for married women grouped according to their husbands occupation order, which are included among the indices of an association between occupation and mortality listed by the OPCS (OPCS, op cit., 1978, pp.68-70).

Men of Economically active ages

3.4. Table 3.1 summarises statistical data produced in the Registrar General’s Decennial Supplements about the trends in inequality in male mortality, as measured by the relative changes in the mortality of different “social” or more strictly “occupational” classes. The table shows the mortality of different classes relative to the national rate.

3.5. The unadjusted data cannot be taken at face value and require careful discussion. After a long period of decline in inequality the statistics appear to suggest that a sharp change took place in the 1950s which persisted into the 1970s. Between 1949-53 and 1959-63 inequality between occupational classes in mortality experience appeared to have widened. Indeed in the supplement published in 1971 the Registrar General stated, “the social class gradient increases with successive censuses so that in 1959-63 the Standardised Mortality Ratio for social class I was only about half that of social class V”. (Registrar General’s Decennial Supplement, England and Wales, 1961, 1971, p.22) However, changes were introduced in 1960 in the classification of occupations and these account for most of the change in the relative mortality experience of classes I and V indicated by the unadjusted data in Table 3.1. In 1961 approximately 26 per cent of occupations were allocated to a class different from that to which they would have been allocated on the 1950 classification (Ibid, p.19).

3.6 Adjusted data for each class for 1959-63 were not given in the Supplement published in 1971. But estimates could be based upon information given in the text. Thus, mortality rates per 100,000 for all men and for men in class V in 1959-63 were given (Table D6) on the basis of both the 1950 and 1960 occupational classifications. These were given only for men in particular age-groups, but weighted totals for men of all ages could be calculated. One estimate was of 128 (substituting for the unadjusted figure of 143) for class V, the comparable figure for 1949-53 being 118 (Townsend, 1974). Although the Registrar General felt overall that it was “impossible to disentangle real differential changes in mortality in this context from apparent differences due to changes in classification” (Decennial Supplement for 1961, P.22), he went on to draw 2 rather important conclusions.

3.7 First, by looking at certain closed professional groups (for example, doctors, lawyers, teachers and clergy) it could be seen that “not all the improvement in social classes I and II is due to differences in classification”, (Ibid, p.22 and p.27). Second, ‘the most disturbing feature of the present results when compared with earlier analyses is the apparent deterioration in social class V …. Even when the rates are adjusted to the 1950 classification it is clear that class V men fared worse than average”. (Ibid, p.29).

Table 3.1 Mortality of men by occupational class (1931-1971)

(standardised mortality ratios)

	Men aged 15-64
Occupational Class	1930-2	1949-53	1959 -63 unadjusted	1959 -63 adjusted	1970-2 unadjusted	1970-2 unadjusted
I Professional	90	86	76	75	77	75
II Managerial	94	92	81		81
III Skilled manual and non manual	97	101	100		104
IV Partly skilled	102	104	103		114
V Unskilled	111	118	143	127	137	121

Notes: a. Corrected figures as published in Registrar General’s Decennial Supplement, England and Wales, . 1961: Occupational Mortality Tables, London, HMSO, 1971, p.22

b. Occupations in 1959-63 and 1970-72 have been reclassified according to the 1950 classification.

3.8. The next decennial supplement, covering occupational mortality during 1970-72, shows little or no change in the mortality “advantage” of classes I and II, but though there was an improvement in the mortality of social class V, relative to other classes, this improvement fell short of restoring the position the class had reached in 1949-53, (OPCS, occupational Mortality, Decennial Supplement, 1970-72, England and Wales, 1978 p.174). And, compared with 1959-63, the mortality of class IV relative to other classes had deteriorated. The report on the years 1970-72 went on to point cut that the age-standardised death rate per 100,000 living at ages 15-64 (using all men in 1970-72 as the standard and after adjusting to the 1950 classification) had declined between the early 1960a and the early 1970s for class V men from 134 to 123, and therefore, according to this criterion, their health had improved “historically”. But this attempt to distinguish changes relative to a historical benchmark from those relative to class structure also demands qualification. The age-standardised death rate for class IV actually increased betveen the early 1960s and the early 1970s and the rate for class III declined only marginally. Table 3.2 gives the figures for the 3 periods for each class.

TABLE 3.2

Recent trends in death rates by occupational class

men aged 15-64 (England and Wales)

	age-standardised death rate per 100,000 living at ages 15-64
Occupational Class	1951	1961	1971
I Professional	103	82	79
II Managerial	108	87	83
III Skilled manual and non manual	116	106	103
IV Partly skilled	119	108	113
V Unskilled	137	134	123

Note: Adjustments have been made by the OPCS to improve comparability between censuses.

Source: OPCS, Occupational Mortality, Decennial Supplement. 1970-72 England and Wales, London, HMSO, 1978, p.174 (supplemented by the OPCS).

3.9 The trend is clearly uneven, in terms of both historical decline in the absolute rates for each class and relativity between classes. Table 3-3 illustrates both these features for different age-groups. First, mortality rates for younger men declined during the whole period of more than 2 decades but the decline was arrested or even reversed for class III and classes IV and V combined for the 3 ten-year age-groups over 35. In the case of men aged 45-64 mortality rates in 1970-72 were either the same as or worse tbom those in 1949-53. During the 1960s a deterioration in the rates for men aged 35-54 in classes III, IV and V (and little or no improvement for older men in these classes) took place.

3.10 Second for each 10 year age group the mortality rates of men in classes III, IV, and V worsened during these 2 decades relative to men in classes I and II.

Table 3.3

Mortality rates per 100,000 and as percentage of rates for occupational classes I and II (1951-71, England and Wales, men and married women)

Occupational class	Age	Men rates per 100,000			Married Women rates per 100,000
		1949-53	1959-63	1970-2	1949-53	1959-63	1970-2
I and II	25-34	124	81	72	85	51	42
III		148	100	90	114	64	51
IV and V		180	143	141	141	77	68
I and II	35-44	226	175	169	170	123	118
III		276	234	256	201	160	154
IV and V		331	300	305	226	186	193
I and II	45-54	712	544	554	427	323	337
III		812	708	733	480	402	431
IV and V		895	842	894	513	455	510
I and II	55-64	2097	1804	1710	1098	818	837
III		2396	2218	2213	1202	1001	1059
IV and V		2339	2433	2409	1226	1129	1131
		as per cent of I and II
I and II	25-34	100	100	100	100	100	100
III		119	123	125	134	125	121
IV and V		145	177	196	166	151	162
I and II	35-44	100	100	100	100	100	100
III		122	134	151	118	130	131
IV and V		146	171	180	133	151	164
I and II	45-54	100	100	100	100	100	100
III		114	130	132	112	124	128
IV and V		126	155	161	120	141	151
I and II	55-64	100	100	100	100	100	100
III		114	123	129	109	122	127
IV and V		112	135	141	112	138	135

Source: OPCS

Table 3.4 Changes in death rates by occupation and age (occupied and retired men)

Occupational Class	25-34			35-44			45-54			55-64
	1951	1961	1971	1951	1961	1971	1951	1961	1971	1951	1961	1971
I	162	82	67	230	166	166	756	535	506	2347	1699	1676
II	114	81	74	225	177	170	704	545	565	2050	1820	1717
III	148	100	90	276	234	226	812	708	733	2396	2218	2213
IV	156	119	118	290	251	270	779	734	826	2103	2202	2301
V	214	202	199	386	436	391	1027	1119	1059	2567	2912	2635
All occupied and retired men	153	108	98	280	237	230	816	704	728	2312	2174	2145
I clergymen	123	42	57	223	109	178	654	409	573	2007	1437	2545
dentists	95	74	64	155	133	159	824	460	583	2320	1393	1608
physicians	140	103	98	230	202	235	736	653	555	2119	1929	1579
accountants	120	87	77	180	186	172	644	512	629	1903	1715	1947
II Teachers	82	72	65	181	134	131	509	412	445	1667	1299	1290
Artists	153	77	55	278	203	238	773	704	795	2180	2009	1953
Journalists, authors, etc.	147	126	102	252	226	237	944	561	644	2139	1658	1875
Innkeepers	162	165	109	422	398	322	1288	1010	1062	3395	3199	2842
III Commercial Travellers	92	87	82	229	187	183	687	592	678	2058	1904	1816
Police	67	61	68	225	152	152	773	610	516	3496	4654	6144
Bricklayers	136	77	79	254	203	205	713	684	705	2363	2104	2377
Boot and Shoe repairers	199	136	127	314	334	356	973	834	952	2559	2734	3374
IV Paintsprayers	142	93	118	324	283	282	791	780	805	2440	2332	2222
Postmen	110	70	52	217	174	142	758	517	563	2255	2166	1762
Telephone operators	208	253	170	308	340	488	1167	770	1040	2646	2914	2452
Fishermen	109	119	85	332	327	329	1055	1063	1444	2958	3184	3423
V Railway porters	155	125	98	335	339	191	905	964	854	2404	2365	2323
Office cleaners	72	157	88	383	285	246	840	794	759	1936	1940	1687
Dock labourers	178	110	151	328	265	243	994	947	904	2739	3053	3065

Source : OPCS

3.11 Classification of mortality by occupation serves only as an indicator of possible causes of inequalities. A large amount of work has concentrated on excess of particular causes of death within occupations or occupational orders, for example, textile workers and diseases of the blood, and miners and circulatory and respiratory diseases (the latest review is to be found in OPCS, 1978, op cit, chapter 5).

3.12 In 1959-63 more class V men died at every age than in 1949-53, from cancer of the lung, vascular lesions of the central nervous system, arteriosclerotic and degenerative heart disease, motor vehicle accidents, and other accidents. Some diseases, like lung cancer and duodenal ulcers which showed no trend with social class, or, like coronary disease, an inverse trend 40 to 50 years ago, were by the 1960s producing higher mortality among social classes IV and V than I and II. In the report for 1959-63, there were 49 out of 85 separate causes of death applying to men (and 54 out of 87 applying to women) in which SMRs for classes IV and V were higher than for I and II. For only 4 causes of death among men (and 4 among married women) was the class gradient reversed. (See Table II in Registrar General’s Decennial Supplement (1959-63), op cit.).

3.13 Some comparison can be made with data for 1970-72. For 92 causes of death which were picked out for men aged 15-64 in the latest OPCS report the mortality ratios for both classes IV and V were higher than for I and II in as many as 68 – which represents a proportionate increase compared with 10 years earlier. For only 4 causes were mortality ratios for I and II higher than for IV and V:- accidents to motor vehicle drivers, malignant neoplasm of the skin, malignant neoplasm of the brain and polyarteritis nodosa and allied conditions (OPCS, 1978, op cit, Table 4A).

3.14 But it remains difficult to explain excess mortality in terms of occupation. This is not only because factors other than the effects of occupation contribute to premature illness and death. It is because both the lifetime and total effects of occupation have not been measured or are difficult to measure. There is mobility between occupations during life, which makes difficult proper evaluation of the specific effects upon health of particular occupations. Strictly, therefore a better measure is required of length of exposure to the effects of an occupation. Secondly, the definition of an occupation and of its conditions, involves a variety of factors, each of which are likely to be related to health. They include working indoors or outdoors , the proportion of time standing or walking about; the number of hours of work; working early or late hours-of the day, or varying times of work from week to week; degree of mental and physical exertion; dexterity or agility involved; degree of warmth, light, quiet, isolation, vibration and humidity; availability of different facilities (Toilet, first aid, telephone, cloakroom or locker for outdoor clothing; coffee and tea; meals); job security; earnings and fringe benefits. We consider that delineation and measurement of such factors will help to explain differences in the health experience of people in different occupations. And changes in the nature of work itself and in the distribution of different types of work working conditions, amenities, remuneration and fringe benefits, and not only changes in the degree of protection offered against specific risks of occupations, will explain trends in health experience. We recommend elsewhere that steps be taken in research and administrative statistics to improve our knowledge of both matters (Chapter 7).

3.15. This brief review of trends in mortality for men of economically active age shows:

i. there was greater inequality of mortality between occupational classes I and V both in 1970-72 and 1959-63 than in 1949-53;

ii. between 1959-63 and 1970-72 the mortality rates of different age-groups over 35 in occupational class III and classes IV and V combined, either deteriorated or showed little or no improvement; and relative to the mortality rates of occupational classes I and II they worsened.

Women of Economically Active Age

3.16. With reservations about occupational class I (numerically a very small category – less than 1 per cent of married women) the data set out in Table 3.5 show the same “spread” of mortality for married and single women as for men. For both married and single women in class IV, and for single women in class V mortality increased relative to women generally during the 1960s.

Table 3.5: Mortality of women by occupational class (1961-1971)(England and Wales)

	Women aged 15-64
	Married		Single
	1959-63	1970-2	1959-63	1970-2
I	77	82	83	110
II	83	87	88	79
III non-manual	103	92	90	92
III manual		115		108
IV	105	119	108	114
V	141	135	121	138

Source: Registrar General’s Decemial Supplement: 1961 p91, 503. OPCS, Decennial Supplement,, 1970-72, p.211.

3.17. Table 3.6 gives more detailed information for different age-groups. Except among the youngest age-groups the “spread” of inequality among married and single women in narrower than among men. But among some age-groups the inequalities between those in classes I and II and those in classes IV have grown. Between 1959-63 and 1970-72 it can be seen that SMRs for class IV deteriorated at all ages for men, married women and single women For class V the experience in the period is mixed, with a tendency, at least for men, for SMRs to increase at earlier ages and steadily decrease at older ages. An increase in SMRs of single women at some ages is noteworthy, but the small numbers of deaths involved has to be borne in mind, (eg only 15 at ages 15-24, and only 83 and 175 at ages 45-54, and 55-64).

Table 3.6 Trends in Standardised Mortality Ratios according to occupational class and age

	Occupational Class	15-19		20-24		25-34		35-44		45-54		55-64
		1959-1963	1970-1972	1959-1963	1970-1972 a	1959-1963	1970-1972	1959-1963	1970-1972	1959-1963	1970-1972	1959-1963	1970-1972
Men	I	72		59	74	73	67	69	72	76	70	78	82
	II	106		85	85	72	74	73	74	77	78	84	84
	III nm/m	97		90	78/90	89	90/90	97	99/97	100	106/101	102	98/111
	IV	118		100	137	107	118	104	117	104	115	101	112
	V	142		149	164	181	199	181	169	158	147	134	128
Married Women	I	(38)		(79)	76	83	79	75	82	78	83	76	83
	II	(41)		(64)	82	76	81	79	80	82	83	85	91
	III nm/m	97		97	85/97	99	92/100	102	93/108	102	91/111	102	92/120
	IV	(88)		92	115	103	119	106	121	104	120	106	118
	V	(159)		159	182	163	163	153	161	144	143	136	128
Single women	I	97		79	132	(67)	96	82	76	86	115	83	117
	II	103		70	105	56	63	65	69	82	69	99	83
	III nm/m	78		72	91/80	74	72/83	73	76/81	86	86/92	104	102/126
	IV	95		98	107	93	96	97	103	104	107	116	121
	V	197		213	232	145	180	132	139	105	137	119	125

Source: Registrar General’s decennial Supplement (1959-63)… op cit, Tables 3A(i), 3B(i), and 3C(i)

a= 15-24

3.18. When causes of death are divided into 13 broad groups for women aged 15-64 there is markedly higher mortality among the partly skilled and unskilled classes (whether defined by their own or a husband’s occupation) in the case of:

i. infective and parasitic diseases;
ii. circulatory disease;
iii. respiratory disease;
iv. diseases of the genito-urinary system, and though less markedly;
v. congenital anomalies;
vi. diseases of the blood;
vii. endocrine and nutritional diseases, and
viii. diseases of the digestive system.

3.19.In the case of (ix) benign neoplasms there is no trend by class but in (x) mental disorders, (xi) diseases of the nervous system, (xii) malignant neoplasms and (xiii) accidents poisoning and violence, there was higher mortality in 1970-72 among classes I and II.

Infant Mortality

3.20 Inequality in mortality among infants reflects that among adults, for both England and Wales and Scotland. Table 3.7 shows that although deaths per 1,000 live births in England and Wales have diminished among all classes the relative excess in combined classes IV and V over I and II increased between 1959-63 and 1970-72. Inequality remained marked in 1975 (Morris, 1979, p.87). As the Court Committee commented, between 1950 and 1973 the perinatal mortality rate declined by 45% for those of professional and 49% for those of managerial class but by only 34% for those of unskilled manual class (Court Report, P7l).

3.21 Scottish trends are similar. During the 1960’s infant mortality rates of each occupational class continued to decline but the class gradient remained broadly the same. The same had been true of earlier decades ( Morris and Heady 1955)

Table 3.7 Trends in infant mortality by occupational class (England and Wales)

	Ratios of actual to expected deaths of infants
	1930-32	1949-53	1959-63	1970-72
I	53	63	73	66
II	73	73		77
III	94	97	98	94
IV	108	114	119	111
V	125	138		175
	Infant deaths per 100 legitimate live births
I	32	19	–	12
II	46	22	–	14
III	59	28	–	16
IV	63	35	–	20
V	80	42	–	31

Source 1959-63 calculated by Julian Tudor Hart, Lancet 22/1/72 p192. 1970-72 OPCS

3.22 Neonatal and post-natal mortality rates for Scotland are shown in table 3.8. It can be seen that the neonatal rates for class V remained about twice as high, and the post-natal rates 6 times as high for class V as for class I in 1975, compared with 1946. In the period 1946-1960 there was some narrowing of the gap between I and V but a reversal of this trend for 1960-1975.

Table 3.8 Neonatal and Post natal Mortality rates (per 1000 live births) by occupational class (Scotland)

	Mortality per 1000 live births			% decrease
	1946	1960	1975	1946-60	1960-75
Neonatal mortality
I	16.7	13	7.6	22	41
II	25	17.2	8.7	31	49
III	29.3	17.1	11.2	42	34
IV	31.1	20.7	10.8	33	38
V	36.9	21	14.6	43	30
Post-natal mortality
I	5.5	2.7	1.8	51	33
II	12.8	4.3	3.8	66	12
III	22	7.2	4.7	67	35
IV	29.3	10.2	5.1	65	50
V	36.1	12.8	10.8	64	16

3.33. Rates of infant mortality have maintained a steady pattern of decline in the post war era. This pattern of decline has been recorded for all occupational classes. Table 3.9 rises the trends for the different components of infant mortality for England. As elsewhere rates listed in Table 3-9 are not strictly comparable because the conventions of classification have changed. These changes mainly affect the figures for 1950 and 1964 but by grouping classes IV and V together the problems of changes of classification can be minimised and comparison enhanced.

3.34. The greatest improvements have been recorded in the rate of post-neo-natal mortality (death from the fifth week to the end of the first year of life) where in classes III to V as well as in the ‘illegitimate’ category rates have fallen by more than 60 per cent during the last quarter of a century. This decline represents a narrowing of the class differential between I plus II and the rest even though the rate for IV plus V and for illegitimate births in 1975-6 was still higher than the rate of I plus II in 1950. It should be noted that the data summarised in the table are collapsed into only 3 categories, compared with 6 in the decennial supplement for 1970-72.

3.35. For neo-natal mortality (death during the first month of life) the degree of improvement has been rather less. Occupational classes IV plus V have made the slowest progressand their failure to maintain parity with the result is particularly marked aver the last decade.

3.36. The same conclusion emerges from the trends for stillbirths. The greatest progress over the 25 year period was made by class I plus II and the least progress for IV plus V. These trends, along with those for neo-natal deaths represent a gradual widening of the gap between the 2 classes at the top and the bottom of the scale. The most recent annual data given below in Table 3.10 do not follow this trend and suggest a catching up process on the part of classes IV and V.

3.37 Over the 25 year period the percentage of illegitimate births has almost doubled. This statistical trend reflects to some extent, real changes in the social meaning of illegitimacy with something of a lessening of the stigma attached to being born outside of wedlock. A growing (but still tiny) minority of women today actually choose to remain unmarried and yet have children and such women are often highly educated and employed in secure and well-paid occupations. For these women, and their children, illegitimacy carries few of the sanctions and hardships which are traditionally associated with it and this pattern of social and cultural change may well have contributed to the fall in the high rate of mortality associated with illegitimate birth.

Table 3.9 Trends in infant mortality 1950-76 (rates per 1000 live births) England and Wales

	rates per 1000 live births			% improvement
Still Births	1950	1964	1975/6	1950/64	1964/76	overall 1950/76
I & II	18.9	11.8	7.8	38	34	59
III	21.5	15.6	9.8	27	37	53
IV & V	24.6	17.2	12	30	30	51
Illegitimate	29.3	21.3	12.7	27	40	57
Neonatal deaths (under 4 weeks)
I & II	13.7	9.2	7.9	33	14	42
III	15.9	11.8	9.3	26	21	42
IV & V	18.4	13.2	11.7	32	22	47
Illegitimate	28.4	19.3	15	32	22	47
Postneonatal deaths (1-11 months)
I & II	5.5	3.5	3	36	14	45
III	10.3	5.4	4	48	26	61
IV & V	15.1	7.6	6.1	50	20	60
Illegitimate	19.9	9.2	7.4	54	20	63
Illegitimate births as % of all live births	5.34	7.6	9.2	+42	+21	42

Recent changes in Infant Mortality: 1975 and 1976

3.38. The most recent data published by OPCS on infant mortality is for the years 1975-76. (Occasional paper No 12, OPCS, 1978). These data are the first set to be published in a continuous series derived from a new linkage of birth and death registration. Table 3-10 presents the data for 1975 and for 1976. Occupational classes I and II and IV and V have been aggregated for purposes of comparison with the earlier ad hoc studies carried out during the fifties and sixties.

Table 3.10 Infant mortality by occupational class 1975-1976

	rates per 1000 live legitimate births					rate per 1000 illegitimate births
	occupational class				Other
Stillbirths	I &II	III	IV & V	I-V
1975	8	10.1	12.6	10.1	10.2	12.9
1976	7.7	9.6	11.3	9.4	8.3	12.6
% improvement	3.7%	4.9%	10.3%	6.9%	18.6%	2.3%
Perinatal (stillbirths and under 1 week)
1975	15	18.3	22.8	18.4	21.2	26.4
1976	13.9	16.8	20.5	16.8	22.12	24.3
% improvement	7.3%	8.2%	10.1%	8.7%	4.8%	8%
Neo-natal (under 4 weeks)
1975	8.4	9.8	12.3	9.9	12.9	16.1
1976	7.5	8.7	11.1	8.9	16.2	13.9
% improvement	10.7%	11.2%	9.8%	10.1%	25.6%	13.7%
Post neonatal (1-11 months)
1975	3.2	4.2	6.5	4.4	7.5	7.6
1976	2.7	3.8	5.8	4	8.9	7.1
% improvement	15.7%	9.5%	10.8%	9.1%	-18.6%	6.6%
% fall in no of births 1975-6	-0.63%	-4.09%	-1.5%	-2.59%	-18.9%	-2.05%

Source: Social and Biological Factors in Infant mortality 1975-76. OPCS

“Other” includes the armed forces, inadequately described occupations, persons who were unoccupied and occupations not stated. Note also between 1975 and 1976 the rate of unemployment in the UK (excluding school leavers) climbed from 3.9 to 5.4.

3.39. During 1975-76 rates of infant mortality continued to decline in all occupational classes. Up to the end of the first week of life the percentage improvement was somewhat higher in classes IV and V than other classes. For neo-natal mortality (death during the first month of life), there was less variation between the 5 classes and for post neo-natal mortality combined classes I and II showed the most improvement. It is always hazardous to draw inferences on the basis of results for only 2 years however and despite the decline in stillbirths among classes IV and V the trends in general in infant mortality do not yet suggest much change in the pattern of relative inequality of the last 2 decades.

3.40. Attention needs to be called to illegitimate births and “other” legitimate births in Table 3-10. Neonatal and post-neonatal mortality rates for these 2 categories are high. The latter category includes the armed forces, the unemployed and others who could not be assigned to an occupational class. Mortality rates during the period from the end of the first week to the end of the first year of life have currently increased by over a fifth. However, in 1975-76 an improvement occurred in the rate of stillbirth. How can such variations be explained? The “other” category accounted for 727 deaths in 1975 and 644 in 1976. The variations in the rates may be an artefact of the measurement process induced by the problematic nature of classification. On the other hand the category includes some seriously deprived families. Between 1975 and 19769 unemployment in Britain jumped to a level unknown previously in the period following the second World War, The numbers unemployed have remained substantially in excess of one million throughout the late 1970s and the number unemployed for 6 months or more has steadily increased. Perhaps the increase in infant mortality recorded here only among the category which includes the unemployed is a reflection of the way in which the economic “health” of the nation imposes upon the physical welfare of the new born in the manner suggested by Brenner. Brenner found that infant mortality rates in the United States were related to economic recessions, with a lag of from one to two years of the peak average mortality behind the peak of unemployment (Brenner 1973, p 155) The hypothesis specified that as a result of maternal deprivation or lack of medical care, in addition to psychological stress, economic decline would be associated with elevated infant mortality rates. More recent work is believed by Brenner to have confirmed “that undesirable changes such as unemployment and income loss are substantially more generative of pathology” (Brenner 1979, p22).

Maternal Mortality

3.41 The trend by occupational class of maternal mortality are shown in table 3.11. During a period of less than a decade mortality fell by more than a third. Although that of class I fell less sharply than other classes inequality between the more numerous class II and classes Iv and V remained about the same. The table shows that mortality among women in class V was nearly double that in classes I and II.

Table 3.11 Maternal mortality by occupational class: married women 15 and over (England Wales)

	Rate per 100,000 births
Occupational class	1962-65	1970-72	% decline
I	16	13	-19
II	22	11	-50
III non-manual	23	13	-39
III manual		15
IV	32	19	-41
V	44	23	-48
All married women	26	16	-38

Mortality of Children

3.42 The ratio of class V to class I deaths is higher in the first to the twelfth months of life than in later years of childhood. The OPCS report for 1970-72 shows than in infancy (after the first month) the ratio was 4.2 for males and 5.0 for females (OPCS, 1978, op. cit., Table 7.7). As table 3.12 shows this ratio is lower at older ages. The trends by class among children of different age have varied. Between one and four years of age there has been a small reduction in the class differential (especially for girls), little or no change between the ages of 5 and 9 and an increase in the differential between the ages of 10 and 14. Excepting stillbirths, fewer females than males in 1970-72 died in childhood in every age-group and class.

Table 3.12 Mortality of children 1-14 by occupational class (England)

	Age
	1-4			5-9			10-14
Occupational Class	1959-63	1970-2		1959-63	1970-2		1959-63	1970-2
		males	females		males	females		males	females
I	69	61	57	33	28	27	30	28	21
II	73	62	54	35	31	24	29	31	21
III non- manual	89	75	62	41	39	27	31	35	21
III		76	62		42	27		35	21
IV	93	93	84	41	44	33	30	40	26
V	154	129	109	67	69	43	41	56	33
V as % of I	223	211	191	203	246	159	137	200	157

3.43 For boys aged 1-14, mortality ratioes for classes IV and V in 1970-72 were both nigher than for classes I and II for 23 of 38 causes of death, compared with only one cause (asthma) where the ratios were lower. For girls the corresponding figures were 22 and 0 respectively.

The elderly

3.44 The occupational class differential in mortality diminishes in the late 30s and 40s and further diminishes as the pensionable ages are approaches. But classification by occupation becomes less meaningful for the elderly. Information about occupation and cause of death recorded on the death certificates of persons of 75 years and over is sometimes imprecise or inaccurate. In the case of widows, especially if dying in the late seventies or afterwards, they may be classified according to the last occupation of husbands dying many years earlier. This may be a weak indicator of life chances and lifestyles over lengthy periods. Again there is evidence that men who had worked for some years before retirement in unskilled occupations tend nonetheless to be listed at death as having worked in skilled occupations if in fact that had been the case previously in their lives (OPCS, 1978, op cit, P.7).

3.45. It is hoped that more reliable data for the elderly will emerge from the longditudinal survey being carried out by the OPCS. Some of the first results have been given above in Chapter 2 (Table 2.10). For men the class gradient corresponds quite closely with that based on less reliable information for 1959-63 and published in a previous Decennial Supplement (Registrar General’s Decennial Supplement, 1959-63 op cit, Table 3A(i)). At that stage the data for 1949-53 and 1959-63, even taking account of changes of classification, indicated relative deterioration in the rates for class V men aged 65-74 (as for younger age groups). (Ibid, p.24).

3.46. Data about the mortality of men aged 65-74 in individual occupations in 1970-72 shows there were very large differences between some groups of manual workers and some groups of non manual workers. For example, the mortality ratio for former miners and quarrymen was 149 gas, coke and chemicals makers 150, and furnaces, forge, foundry and rolling mill workers 162, compared with administrators and managers with a ratio of 88 and professional, technical workers and artists with a ratio of 89. (OPCS 1978, op. cit, p.167). Compared with 1959-63 the mortality ratios of several manual occupations (including the 3 listed above) deteriorated, relative to the ratio for all men. (Registrar General’s Decennial Supplement, 1971, op cit, Table 3A(i)).

Trends in Mortality by Age

47, The trends during the life-span have been shown for both sexes. Inequalities are largest during the first year of life especially after the end of the first month. After some diminution in the differential between the classes during childhood it widens again and reaches a second smaller, peak in early adulthood. The differential narrows in late middle age for women and not only men (Table 3.4) and appears to remain small among the elderly. These trends with age apply whether we take the difference between classes V and I or IV and I (Table 3.13). However we must emphasise that although the relative difference between classes V and I and IV and I diminishes in middle and old age the absolute difference in numbers of deaths increases. For example, at 60-64 that additional numbers of deaths of men per million population in class V than in class I was 10,622, compared with 868 per million at 20-24 (OPCS, 1978, op.cit., p37)

Table 3.13 Male mortality rates at different ages in occupational class IV and V as a percentage of those in class I

	Class IV as per cent class I				Class V as per cent class I
	1959-63	1970-72			1959-63	1970-72
Stillbirths		148				199
Perinatal (still births, and less than 1 week)		145				195
Early neonatal (less than 1 week)		143				192
Late neonatal (1-3 weeks)		164				249
Post neonatal (1-11 months)		211				421
Total infant mortality (under 1 year)		162				255
1-4 years		153				212
5-9 years		160				246
10-14 years		141				200
Total 1-14		151				219
15-19		148			197	161
20-24		184			254	236
25-34	145	176			246	297
35-44	151	163			263	235
45-54	137	163			209	209
55-64	130	137			171	157

3.48. The structure of inequality in mortality rates during the life-span does not appear to have changed much in recent decades. Table 3.13 shows that although the ratio of class V deaths to class I deaths diminished for some age-groups during the 1960s it increased for others, and the ratio of class IV deaths to class I deaths increased in all age-groups. If we combine classes IV and V then between 1959-63 and 1970-72 their mortality worsened, relative to classes I and II, for each ten-year age-group between 25 and 65. It should also be noted that class III also slipped back (Table 3-4). These trends for adults, as those quoted earlier for infants, are very serious, and need to be carefully analysed and explained.

Morbidity

3.49 As pointed out earlier it is difficult to trace morbidity data by class for any span of years. The General Household Survey has now been running since 1971 but it is still too early to distinguish reliable trends in health from that source. Two examples are given below. Table 3.14 shows that absence from work because of sickness or injury is sharply related to class but that the precise rates are liable to fluctuation from year to year. The average number of days lost through illness or accident among unskilled manual men was 41/2 times that among professional men in 1971 and 1972 (the data are not given for 1977).

Table 3.14 Working males absent from work due to illness or injury (England and Wales 1971)

Socio-economic group	Absent from work due to illness or injury in a two week reference period – rate per 100			Average number of work days lost per person per year
	1971	1972	1977 (rates given only to nearest 10)	1971	1972
Professional	37	21	20	3.9	3.1
Employers and managers	37	39	20	7.2	6.2
Intermediate and junior non-manual	44	48	50	7.6	6.0
Skilled manual	57	56	60	9.3	9.4
Semi-skilled manual	56	68	70	11.5	10.5
Unskilled manual	88	99	60	18.4	17.6
All groups	52	54	40	9.1	8.4

3.50 Table 3.15 also shows a class gradient during the 1970s for restricted activity (acute sickness), long-standing (chronic) illness and GP consultations but the rates are even more uneven fromyear to year and in some years, for some age-groups, there is no perceptible gradient. The figures illustrate the problems of drawing conclusions about trends in self-reported sickness, for some major sex/age-groups if not for the population as a whole, during a short span of years.

Table 3.15 rates of long standing and acute illness and consultations per 1000 of occupational classes IV and V, as a per cent class I (1971-76) Britain

Sex/class/health indicator	1971	1972	1973	1974	1975	1976
Males Class IV
long-standing illness		158	163	157	160	157
acute sickness	126	133	110	134	102	80
(percent of class II)	120	120	125	119	124	99
GP consultations	133	132	125	146	129	91
Males class V
long-standing illness		196	213	218	197	196
acute sickness	155	181	129	150	85	102
(percent of class II)	148	163	146	134	103	127
GP consultations	143	175	164	147	121	125
Females Class IV
long-standing illness		274	214	182	197	176
acute sickness	105	128	115	115	134	95
(percent of class II)	111	134	110	123	128	130
GP consultations		180	150	110	123	114
Females class V
long-standing illness		320	276	204	253	246
acute sickness	107	141	113	122	128	94
(percent of class II)	114	146	109	131	122	129
GP consultations		117	150	120	107	102

INEQUALITIES AND DISTRIBUTION

3.51. When examining indicators of health for different occupational or socioeconomic classes for a span of years any changes that may be taking place in the relative size of particular classes may be as important as any changes in the inequality between classes in assessing trends in the overall health of the population. Some commentators have pointed out that while inequalities in health between the unskilled manual class and other classes may not have diminished, or may even have increased that class has become smaller and therefore there has still been an “improvement” in the distribution of health. This change has been regarded as compensation for the lack of any closing of the gap between classes.

3.52. Two comments should be made. The first is that changes in occupational classification have caused commentators to believe that the reduction of class V since 1931 has been greater than it has. This is shown by the adjusted and unadjusted figures in Table 3.16. Since 1961, for example, the fall in proportion of men in class V has been small and in absolute numbers has not fallen at all.

3.53. The second is that relatively poor health experience applies to other manual classes and especially class IV and that though this class too has fallen in proportion to population it continues to make up, together with class V, more than a quarter of the economically active male population, Mortality indicators for class IV, relative to other classes, have shown some deterioration between the early 1960s and early 1970s and, as discussed above, it would be wrong to confine discussion of health inequalities to class V.

Table 3.16 Percentage of economically active men in different occupational social classes 1931, 1951, 1961, 1966, 1971 (England and Wales)

Occupational class	1931		1951		1961	1966	1971
I	1.8	2.2	2.7	3.2	4.0	4.5	5.0
II	12.0	12.8	12.8	14.3	14.9	15.7	18.2
III	47.8	48.9	51.5	53.4	51.6	50.3	50.5
IV	25.5	18.2	23.3	16.2	20.5	20.6	18.0
V	12.9	17.8	9.7	12.9	8.9	8.8	8.4
Total	100	100	100	100	100	100	100
Number (Thousands)	13,247		14,067		14,649	15,686	15,668

Percentage have been weighted to allow for changes in classification between 1931 and 1961 censuses. The second column for 1931 and 1951 are based on the classification at those times.

CONCLUSION

3.54. Our review of trends in inequalities of health has produced some disturbing conclusions.

3.55. As explained earlier in the chapter trends are not easy to trace, either because of inconsistencies in the categorization of data or changes in occupational classification. Our conclusions make allowances for these. problems. We have also had the opportunity of comparing trends in infant mortality with trends in mortality of people at later ages. Analyses in the literature have tended to concentrate attention either on infant mortality or mortality of males of economically active age rather than on the population of both sexes of different age.

3.56. Perhaps the most important general finding in the chapter is the lack of improvement, and indeed in some respects deterioration, of the health experience not merely of occupational class V but also class IV in health relative to occupational class I, as judged by mortality indicators during the 1960s and early 1970s. The more specific conclusions underlying this finding, are as follows. (These conclusions apply to England and Wales. Scottish experience has been rather similar, though certain differences are noted in the text).

i. Mortality rates of males are higher at every age than of females and in recent decades the difference between the sexes has become relatively greater.

ii. For men of economically active age there was greater inequality of mortality between occupational classes I and V both in 1970-72 and 1959-63 than in 1949-53.

iii. For economically active men the mortality rates of occupational class III and combined classes IV and V for age-groups aver 35 either deteriorated or showed little or no improvement between 1959-63 and 1970-72. Relative to the mortality rates of occupational classes I and II they worsened.

iv. For women aged 15-64 the standardised mortality ratios of combined classes IV and V deteriorated. For married and single women in class IV (the most numerous class) they deteriorated at all ages.

v. Although deaths per thousand live births in England and Wales have diminished among all classes the relative excess in combined classes IV and V over I and II increased between 1959-63 and 1970-72.

vi. During period of less than a decade maternal mortality fell by more than a third. Although that of class I fell less sharply than other classes inequality between the more numerous class II and classes IV and V remained about the same.

vii. Among children between 1 and 4 years of age, there has been a small reduction in the class differential (especially for girls), for children aged 5 to 9 little or no change, but for children aged 10 to 14 an increase in the differential. For boys aged 1-14, mortality ratios for classes IV and V in 1970-72 were both higher than for classes I and II for 23 of 38 causes of death compared with only one cause (asthma) where the ratios were lower. For girls the corresponding figures were 22 and 0 respectively. There is evidence that as rates of child death from a specific condition decline to very low levels class gradients do disappear. The gradual elimination of death from rheumatic heart disease over the post war period provides evidence of this (Morris, 1959).

CHAPTER 3 – References

ADELSTEIN, A.M., and WHITE, G.C., “Causes of Children’s Deaths Analysed by Social Class (1959-63)”, in Child Health: A Collection of Studies. London, HMSO, 1976.

BRENNER, M.H., “Foetal, Infant and Maternal Mortality daring Periods of Economic Instability”, International Journal of Health Services, Vol 3, 1973.

BRENNER, M.H., “Influence of the Social Environment and Psycho pathology: The Historical Perspective”, in Barrett, J.E., et al (eda), Stress and Mental Disorders t New York, Raven Press, 1979.

Fit for the Future The Report Of the Committee on Child Health Services ( Court Report), Cmnd, 6684 London, HMSO, 1977

FOX, A.J., and ADELSTEIN, A.M., “Occupational Mortality: Work or Way of Life?” Journal of Epidemiology and Community Health, vol 32,No 2, June, 1978.

HIGGINS, I.T.T., et al. “Population Studies of Chronic Respiratory Disease; A Comparison of Miners, Foundry Workers and others in Staveley, Derbyshire”, British Journal of Industrial Medicine. Vol 16, 1959.

McKEOWN, T., The Role of Medicine. London, The Nuffield Provincial Hospitals Trust, 1976.

McLAUGHLIN, A.I.G., “Chronic Bronchitis and Occupation”, British Medical Journal. Vol 1, 1966.

Medical Research Council, “Chronic Bronchitis and Occupation”, British Medical Journal, Vol 1, 1966

MORRIS, J.N., “Health and Social Class” The Lancet February 7th, 1959.

MORRIS, J.N., “Social Inequalities Undiminished”, The Lancet, January 13th» 1979.

MORRIS, J.N., Uses of Epidemiology. Churchill Livingstone, Edinburgh, 1975 (3rd ed).,

MORRIS, J.N., and HEADY, J.A. “Social and biological factors in infant mort V mortality in relation to the father’s occupation 1911-1950 The Lancet 1 (1955) 554.

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

Registrar General’s Decennial Supplement. England and Wales. 1961: Occupational Mortality Tables. London, HMSO, 1971•

Scottish Home and Health Department, Towards an Integrated Child Health Service Edinburgh, HMSO, 1973.

Social and Biological Factors in Infant Mortality. 1975-76. Occasional Paper No 12, OPCS, 1978.

TOWNSEND, P. “Inequality and the Health Services¹¹, The Lancet. 15 June, 1974