Marshall S W (Injury Prevention Research Unit, Department of Preventive and
Social Medicine, Otago Medical School,. PO Box 913, Dunedin, New Zealand), ...
Vol. 22, No. 2 Printed in Great Britain
International Journal of Epidemiology © International Epidemtologlca) Association 1993
Social Class Differences in Mortality from Diseases Amenable to Medical Intervention in New Zealand STEPHEN W MARSHALL.* ICHIRO KAWACHI," NEIL PEARCE* AND BARRY BORMAN 1 Marshall S W (Injury Prevention Research Unit, Department of Preventive and Social Medicine, Otago Medical School, PO Box 913, Dunedin, New Zealand), Kawachi I, Pearce N and Borman B. Social class differences in mortality from diseases amenable to medical intervention in New Zealand. International Journal of Epidemiology 1993; 22: 255-261. Social class differences in mortality from causes of death amenable to medical intervention were examined. All deaths in New Zealand males aged 15-64 years during the periods 1975-1977 and 1985-1987 were identified. Strong social class gradients in mortality from causes of death amenable to medical intervention were observed during both periods. Furthermore, social class inequalities were more pronounced for amenable causes of mortality than for non-amenable causes. However, a marked decline in the age-standardized mortality rate from amenable causes was observed, with the rate falling by 30% over the 10-year study period. This decline was twice as large as the drop in the non-amenable mortality rate. Despite the fall in the death rate from amenable causes, social class inequalities in mortality persisted among New Zealand men, with the lowest socioeconomic group experiencing a death rate from amenable causes of mortality that was 3.5 times higher than men in the highest socioeconomic group.
The concept of amenable or 'avoidable' mortality originated from the work of Rutstein et at.1-2 who proposed a list of 'sentinel health events' to be used as indicators of the quality of medical care. The list was based on diseases for which mortality was judged to be largely avoidable given appropriate medical intervention. This idea was subsequently taken up in studies of mortality variations over time and across geographical areas, at first by Charlton et a/.3"5 and later by an increasing number of other investigators.6"10 Time-trend studies in different countries have in general shown that mortality from amenable causes has declined faster over the past decadesythan most other causes of death." In the time-trend study by Charlton and Velez,4 declines in mortality rates from amenable causes were shown in all six countries studied (England and Wales, Sweden, Italy, France, US and Japan). This consistency in mortality trends between countries which otherwise differed in their social, environmental, genetic and diagnostic factors, appeared to suggest that improvements in health care
were a factor in the observed declines. Rapid declines in mortality from amenable causes were also observed in Finland7 and the Netherlands.9 In New Zealand, overall age-standardized mortality rates (AMR) among working males declined between 1975-1977 and 1985-1987.12 As in several other countries, a strong social class gradient was found in both periods, with males in the lower socioeconomic groups experiencing the higher rates of mortality. One of the hypotheses put forward to explain the observed social class gradient in male mortality is unequal access to health care.12"15 This implies the presence of a strong social class gradient for diseases where the mortality rate can be expected to be amenable to the provision or non-provision of the appropriate medical services. Therefore, using the list of amenable causes of death developed by Charlton et al.* we have examined the time-trends and social class patterns of mortality in New Zealand males over the 10-year period 1975-1977 to 1985-1987.
• Injury Prevention Research Unit, Department of Preventive and Social Medicine, Otago Medical School, PO Box 913, Dunedin, New Zealand. •• Department of Community Health and * Department of Medicine, Wellington School of Medicine, Wellington, New Zealand. t Health Statistical Services, Department of Health, Wellington, New Zealand.
METHODS The present study used the list of amenable causes of mortality developed by Charlton et al.*-6 (Table 1). Although all these conditions were used in the original studies by Rutstein et al.'-2 (with the exception of chronic rheumatic heart disease), the list is not a com255
Downloaded from https://academic.oup.com/ije/article-abstract/22/2/255/752473 by guest on 26 November 2017
256
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
TABLE 1 Selection of amenable causes of death (after Chartton)
Cause of death
1CD code (8th Revision)
ICD code (9th Revision)
Hypertensive disease Tuberculosis Asthma Chronic rheumatic heart disease Appendicitis Acute respiratory disease Bacterial infections
400-404 010-019 493 393-398 540-543 460-466, 470-474 004, 034, 320, 381-383, 390-392, 680-686, 710, 720 201 550-553 574-575 280-281 480-486,490
401-405 010-018, 137 493 393-398 540-543 460-466,487 004, 034, 320, 381-383, 390-392, 680-686, 711,730 201 550-553 574-575 280-281 480-486,490
Hodgkin's disease Abdominal hernias Acute and chronic cholecystitis Deficiency anaemias Pneumonia and bronchitis
plete selection. Examples of conditions mentioned by Rutstein et at. but not selected by Charlton et al?A include: epilepsy, peptic ulcer, and cancers of the larynx, colon and rectum, thyroid, neuroblastoma, and malignant melanoma. Other studies have similarly adopted variations on the original list proposed by Rutstein et a/.10 However, virtually all the studies of amenable mortality have incorporated the core list of conditions enumerated by Charlton et al. thereby ensuring a degree of comparability. The Elley-Irving scale was used to classify men into different socioeconomic groups.15-16 In this scale, occupations are classified into six levels using an equal weighting of median income and median educational level. The socioeconomic groups represented in the Elley-Irving scale are broadly comparable to other systems of socioeconomic classification, for example the British Registrar General's classification of social class. The Elley-Irving social class 1 consists mainly of professional, technical and related workers, while social class 6 consists of unskilled manual workers.16 Although the scale has come under some criticism—particularly for the arbitrary assignation of equal weights to education and income levels—it nonetheless remains the most frequently used social class scale in New Zealand.17 As in previous studies12-17'" this study was confined to working males aged 15-64 years. Death certificate occupational information in women was considered too unreliable for social class categorization. Even in the case of never-married women, less than 50% could be classified by social class. The denominator data were obtained from 10% samples of the 1976 and 1986 New Zealand Census of Population and Dwellings. The numerator data were obtained from the (Department of Health) Health Statistical Services' registry of
Downloaded from https://academic.oup.com/ije/article-abstract/22/2/255/752473 by guest on 26 November 2017
all deaths in New Zealand during 1975-1977 and 1985-1987. Each individual in the numerator and denominator was categorized according to the ElleyIrving classification on the basis of the 3-digit occupation code of the New Zealand Standard Classification of Occupations (NZSCO)," which is a New Zealand adaptation of the International Standard Classification of Occupations.20 Classification rates by social class were 86% and 81 % for the 1976 and 1986 censusderived denominator data, respectively. As for the numerator data, deaths from all amenable causes totalled 731 among men aged 15-64 between 1975-1977, of which it was possible to classify 85% (623) according to social class; between 1985-1987 there were 567 deaths from all amenable causes, of which it was possible to classify 74% (417) into social classes. The causes of death were determined on the basis of the 3-digit disease code of the 8th Revision (for the years 1975-1977) and the 9th Revision (for the years 1985-1987) of the International Classification of Diseases (fCD).21-22 For each disease grouping, direct AMR per 100000 person-years of risk were calculated for each social class using a computer program described in a previous paper.23 Data were classified into 10-year age groups, and Segi's world population was used as the standard.24-23 Tests for trend with social class were performed using the Mantel-Haenszel extension test.26 The strength of the mortality trend across social class was assessed using a previously described technique.27 This involved performing an age-standardized weighted regression of the class-specific mortality rates for each disease group in order to estimate the slopes of the overall social class mortality trends. The slope for each disease group was then divided by the pooled
257
SOCIAL CLASS DIFTERENCES IN MORTALITY IN NEW ZEALAND
rate for classes 1 -6 to assess the strength of the social class trend.
decline in mortality was particularly noticeable in the categories of hypertensive disease (50% decline), tuberculosis (73% decline), chronic rheumatic heart disease (58% decline), acute respiratory infections (60% decline), and Hodgkin's disease (57% decline) (Tables 2 and 3). The most notable exception to the general trend was mortality from asthma, which increased by 22.5%, from 4.0 to 4.9 per 100000 person-years. However, the New Zealand pattern for asthma mortality is atypical in that an epidemic of asthma deaths began in 1976, and has been linked to the introduction of the beta agonist drug fenoterol in that year.28 Mortality has declined since warnings about the safety of the drug were issued in 1989.29 A strong social class gradient in mortality from amenable causes was observed during both periods, with the lower socioeconomic groups generally experiencing higher mortality rates. Figure 1 presents the AMR ratios for 1975-1977 and 1985-1987, calculated using the AMR for each class, and pooled rates for classes 1-6 in each time period. The gradient across social class was especially marked for the following conditions: hypertensive disease, tuberculosis, asthma, chronic rheumatic heart disease, and acute respiratory infections (P < 0.0001 in all categories). However, between 1975-1977 and 1985-1987 the strength of the social class gradient actually diminished for overall causes of amenable mortality, from 34% (in 1975-1977) to 28% (in 1985-1987). The largest
RESULTS Over the 10-year study period, the AMR from all causes of amenable mortality, pooled for all men irrespective of social class categorization, declined by 29% from 25.5 per 100000 person-years (in 1975-1977) to 18.0 per 100000 person-years (in 1985-1987). It is difficult to make valid comparisons of the pooled mortality rates for classes 1-6 (i.e. excluding the men who could not be classified by the EUey-Irving scale) with the overall mortality rates pooled for all men age 15-64 (i.e. including the men who could not be classified). As previously mentioned, a higher proportion of the denominator census data were found to be classifiable compared to the numerator data (deaths). Consequently, the nominal death rate for the unclassifiable category was comparatively high; and in both time periods, the overall mortality rates pooled for all men aged 15-64 years were slightly higher than the pooled rates for men in classes 1-6. For this reason, all social class analyses were restricted to comparing the death rate in each class with the pooled rate for classes 1-6, and the data for the unclassifiable category were excluded from further analysis. The pooled mortality rate for men in Elley-Irving classes 1-6 declined by a similar amount (30%) from 26.0 to 18.1 deaths per 100000 person-years. The
TABLE 2 Age-standardized mortality rates (per 1000O0 person-years) for causes of death amenable to medical intervention in employed New Zealand males aged 15-64 years, 1975-1977, by Elley-Irving social class
EUey- Irving social class scale Pooled rate (classes 1-6)
Pooled rate (all men aged 15-64)
Slope
Vo of pooled rate
% of total amenable rate
Cause of death
1
2
3
4
5
Hypertensive disease Tuberculosis Asthma Rheumatic heart disease Appendicitis Acute respiratory disease Bacterial infections Hodglcin's disease Abdominal hernias Cholecystitis Deficiency anaemias Pneumonia and bronchitis
1.8 1.5 6.6 3.0 0.8 0.0 0.0 3.8 0.0 0.8 0.0 0.0
2.1 1.5 1.3 4.4 0.0 0.6 1.0 1.8 1.0 0.4 0.0 1.2
4.5 1.4 2.1 3.8 0.2 0.5 0.7 1.3 0.1 0.6 0.0 1.8
4.9 1.5 4.2 5.2 0.0 0.6 0.5 0.8 0.4 0.4 0.0 3.7
7.4 2.6 6.0 7.8 0.2 1.1 0.7 1.5 0.4 0.4 0.2 5.3
12.1 8.1 9.3 16.1 0.6 5.1 2.2 0.9 1.6 1.6 0.0 12.1
5.4 2.2 4.3 6.0 0.2 1.0 0.8 1.4 0.5 0.5 0.0 3.8
4.9 2.3 4.0 5.4 0.2 1.2 0.9 1.3 0.4 0.5 0.0 4.5
1.78 0.89 1.27 2.01 0.03 0.65 0.20 -0.28 0.11 0.07 0.03 2.04
33.0 40.3 29.6 33.6 15.0 65.2 24.4 -20.0 22.8 14.8 0.0 53.6
20.2 10.1 14.4 22.9 0.3 7.4 2.2 -3.2 1.3 0.9 0.4 23.1
All amenable All non-amenable
18.3 433.1
15.4 444.0
17.0 479.9
22.1 515.0
33.7 623.6
69.8 793.8
26.0 541.8
25.5 479.1
8.80 67.46
33.9 12.5
100.0 -
All causes
451.4
459.4
496.9
537.1
657.3
963.6
567.8
504.6
76.27
13.4
6
Downloaded from https://academic.oup.com/ije/article-abstract/22/2/255/752473 by guest on 26 November 2017
258
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
TABLE 3 Age-standardized mortality rates (per 100000 person-years) for causes of death amenable to medical intervention in employed New Zealand males aged 15-64 years, 1985-1987, by Elley-Irving social class
Elley-Irving social class scale Pooled rate (classes 1-6)
Pooled rate (all men aged 15-64)
Slope
ft of pooled rate
ft of total amenable rate
Cause of death
1
2
3
4
5
6
Hypertensive disease Tuberculosis Asthma Rheumatic heart disease Appendicitis Acute respiratory disease Bacterial infections Hodgkin's disease Abdominal hernias Cholecystitis Deficiency anaemias Pneumonia and bronchitis
0.0 0.0 4.3 0.6 0.0 0.0 0.6 0.0 0.6 0.0 0.0 2.8
2.6 0.5 1.7 1.0 0.0 0.0 0.6 0.8 0.0 0.6 0.0 2.0
2.0 0.2 4.2 2.5 0.2 0.5 0.5 0.5 0.0 0.2 0.0 3.1
2.4 0.2 5.2 2.1 0.2 0.2 0.8 1.2 0.8 0.3 0.0 4.2
4.8 0.8 6.9 4.3 0.5 0.8 0.7 0.6 0.5 0.5 0.0 9.5
3.9 3.5 7.0 5.1 0.0 1.2 1.7 0.0 1.2 0.0 0.0 7.4
2.7 0.6 4.9 2.5 0.2 0.4 0.7 0.6 0.4 0.3 0.0 4.7
2.5 0.7 4.9 2.6 0.1 0.4 0.9 0.6 0.3 0.2 0.0 4.7
0.70 0.37 0.97 0.89 0.06 0.19 0.14 0.04 0.16 0.01 0.00 1.50
25.9 61.2 19.9 35.4 32.5 46.3 20.6 7.0 41.0 1.7 0.0 32.0
13.9 7.3 19.3 17.6 1.3 3.7 2.9 0.8 3.3 0.1 0.0 29.9
All amenable All non-amenable
9.0 291.8
10.0 399.1
13.7 371.2
17.7 486.3
30.1 593.0
31.1 583.2
18.1 460.4
18.0 410.9
5.04 65.05
28.0 14.1
100.0 -
All causes
300.8
409.1
384.9
504.0
623.1
614.3
478.5
428.9
70.08
14.6
-
3.00
0.00
Elley-Irving social class FIGURE 1 New Zealand male age-standardized rate ratios for causes of death amenable to medical intervention, by social class
Downloaded from https://academic.oup.com/ije/article-abstract/22/2/255/752473 by guest on 26 November 2017
SOCIAL CLASS DIFFERENCES IN MORTALITY IN NEW ZEALAND
declines in the strength of the social class gradient were observed for hypertensive diseases (33-26"%), asthma (30-20%), acute respiratory infections (65-46%), and acute and chronic cholecystitis (15-2%). By comparison, the strength of the social class gradient actually increased for tuberculosis (40% in 1975-1977, to 61% in 1985-1987), appendicitis (15-33%), and abdominal hernias (23-41%). In the case of mortality from Hodgkin's disease, a negative social class slope was observed in. 1975-1977, which was subsequently reversed by 1985-1987 to become a slightly positive social class gradient. During the study period, there was a marked decline in amenable death rates among men in Elley-Irving classes 1-6. The decline in the death rate from amenable causes (30%) was twice as large as the decline in the death rate due to non-amenable causes (15%). As a result, amenable mortality accounted for a smaller proportion of the all-causes mortality slope in 1985-1987 (7%), than it had in 1975-1977 (12%). However, the strength of the overall gradient due to amenable mortality (34%) was noticeably stronger than the strength of the gradient for all-causes mortality (13%) in 1975-1977, and this pattern was still apparent in 1985-1987 (28% for amenable causes versus 15% for all causes). This indicates that, despite the encouraging decline in the gradient for amenable mortality, there remains considerable potential for continued reduction in deaths from amenable causes in New Zealand.
DISCUSSION The present study showed that amenable causes of mortality declined by 29% in New Zealand men aged 15-64 years over the 10-year study period. This rate of decline was more than double that of non-amenable causes of mortality which declined by 14% over the same period. The more rapid decline in amenable causes of death is consistent with reports from other countries,7"9 where the differences in the rate of decline between amenable and non-amenable mortality rates have been attributed to improvements in health care. However, few studies have examined the social class differences in amenable causes of mortality. Three findings from the present study should be emphasized. Firstly, social class inequalities exist for both amenable as well as non-amenable causes of death, with the lowest socioeconomic groups experiencing the highest rates of mortality from both causes. Secondly, social class inequalities are more marked for amenable causes of mortality than non-amenable causes, suggesting the importance of access to adequate health
259
care for at least some of the specific causes of amenable mortality. Thirdly, in spite of the fall in the rate of amenable mortality over the 10-year period, social class inequalities in mortality persisted among New Zealand men aged 15-64. Although the degree of social class inequality for amenable mortality diminished somewhat between 1975-1977 and 1985-1987, men in the lowest socioeconomic group continued to experience a death rate from amenable causes of mortality that was nearly 3.5 times higher than men in the highest socioeconomic group (ElleyIrving class 1). It should be stressed, however, that deaths from amenable causes account for only a small proportion of overall mortality (5.0% in 1975-1977 and 4.2% in 1985-1987). Caution must be applied in the interpretation of trends in amenable causes of mortality. These were used by Charlton et al. only as an indicator 'to provide warning signals of possible shortcomings in health care delivery'.3'4 Analyses of geographical variation, either between or within countries, in which amenable mortality was related to the actual provision of health services resources, have generally yielded weak and inconsistent results.6'" Observed changes in amenable mortality may reflect changing patterns of disease incidence rather than the delivery of medical care. Previous investigators have readily acknowledged that for a number of amenable conditions, 'spontaneous' declines in incidence have been observed.9" In studies using aggregated mortality data (including the present study) it is generally not possible to distinguish declining incidence of disease from improved survival, and hence, declining mortality from disease. The decline in amenable mortality should therefore not be solely attributed to improvements in medical care.4'7'8 However, the fact that the amenable mortality rate has declined more rapidly than the non-amenable rate is consistent with a role for improvements in medical care. The absence of a clear relationship between indices of health care and amenable mortality rates does not necessarily invalidate the use of amenable mortality analyses as indicators of the effectiveness of health care services." It has been argued that the health care indices selected to 'explain' the mortality variation in previous analyses were crude measures of the supply or use of health care services. It is possible that the organization, quality and access to health care services are more important for the prevention of amenable deaths than the level of supply per se.n Secondly, the levels of supply may in part reflect the allocation of health services according to regional differences in mortality—for example, in England and Wales, and in
Downloaded from https://academic.oup.com/ije/article-abstract/22/2/255/752473 by guest on 26 November 2017
260
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
New Zealand—where budget allocations to local health authorities and boards are made according to a population-based formula which incorporates an adjustment for 'need' as measured by area-specific standardized mortality ratios." These issues notwithstanding, the findings of the present analysis indicate a cause for concern for the New Zealand health services. In spite of an overall 29
