social factors and the gender difference in mortality

Sot. Sci. Med. Vol. 23, No. 6, pp. 605-609, 1986 0277-9536/86 $3.00 + 0.00 Printed in Great Britain. All rights reserved Copyright Q 1986 Pergamon Journals Ltd

SOCIAL FACTORS AND THE GENDER DIFFERENCE IN MORTALITY

MARKKU KOSKENVUO, JAAKKO KAPRIO, JOUKO ~NNQVIST and SEPPO SARNA

Department of public Health Science, University of Helsinki, Haartmaninkatu 3, SF-00290 Helsinki, Finland

Abstract-The effect of social factors on the male/female difference in mortality in Finland was studied by comparing age-adjusted mortality of males and females by social class and marital status. 44,548 death certificates (years 1969-1971) and 1970 census data for 25-64-years olds were analysed. The gender difference was 2.8-fold: 5.3-fold for violent causes and 2.3-fold for natural causes. The greatest gender difference from violent causes was found in accidental poisonings (18.7-fold) and drownings (12.8-fold), and from natural causes in mental disorders (mainly alcoholism; 5.7-fold) and in ischemic heart disease (4.5-fold). The gender difference was most prominent in unskilled workers, divorced and widowed and less prominent in married and upper professionals. The great variation of gender difference of mortality by social class and marital status seems to indicate that mortality difference between males and females is associated to external factors rather than biological differences between men and women. This conclusion is also supported by the progressive increase of gender difference of mortality from 1.4 to 2.8 during the last 80 years in working-aged Finns.

Key words-mortality rates, gender difference, Finland, social factors in mortality

INTRODUCTION

Compared to many industrialized countries, mortality is high in Finland, particularly mortality of middle-aged men. Compared to other Scandinavian countries, Finnish men have the highest mortality for most common causes of death. A particular feature of Finnish mortality is a strikingly large difference between men and women compared to other countries [l-6]. A long-term analysis of male/female mortality ratio from 1861 to 1960 in Finland showed a slow increase between 1910 and 1940, and a clear increase after the Second World War [2]. The same increase began about 20 years earlier in the United States [7]. The change in gender difference of mortality seems also to associate with the still un- explained increase and decline of coronary heart disease in some industrialized countries [8,9]. The gender difference has been explained both by biological and by environmental factors [7,8, 10-121. Gen- der differences in mortality are reflected in differences in incidence rather than in difference in survival [ 111. We studied the variation of the male/female cause- specific mortality ratios by social factors. The aim was to identify the causes of death in which the difference between men and women is the greatest, as well as those causes of death in which the mortality ratio varies most by social class and marital status. Marital status measures, albeit crudely, success in forming a lasting personal relationship, which would be the base for a strong social network. Occupation is the basis for classification of social class and occupational class. While occupational class groups together occupations involved roughly the same type of work, social class groups persons with the same educational level and with the same mental and physical characteristics of work. The former is better

suited for occupational health studies, the latter for studies in social medicine. The great variation of gender difference of mortality by social class and marital status would indicate that external factors play a role in the gender difference of mortality rather than biological differences between men and women. On the contrary a small variation of the difference by social indicators may be caused by two different mechanisms: (1) significant effect of biological factors, or (2) significant effect of external factors, which are not associated with marital status or social class.

MATERIALS AND METHODS

The analysis was based on Finnish mortality data from 1969 to 1971 [described in detail in 131. Death certificates of 44,548 persons aged 25-64 years and population figures from the 1970 census were used to compute mortality rates by S-year age group in men and women by marital status and social class. Rates were age-standardized by the direct method. The standard deviation of morality ratios (RR) was com- puted to illustrate the variation of mortality ratios by marital status or social class.

The marital status and social class classification was based on death certificate data. Data was missing for less than 0.1% for marital status and 2.2% for social class. The unemployed and retired were classified by their former occupation, while house- wives were classified by the occupation of the house- hold head. Classification of social class is based on the classification used by the Central Statistical Office of Finland. The group of upper professionals includes persons with academic education (e.g. economist, lawyer, physician, teacher) or directors or managers of a’company or factory. The group of lower profes-

605

606 MARKKU KOSKENVUO et al.

sionals includes persons with clerical, secreterial or

non-industrial work (clerk, nurse, sales agent, secre- tary) or self-employed on a small-scale (shopkeeper). Skilled workers are trained persons in manual or physical work.

The causes of death were classified by the eighth revision of the International Classification of Dis- eases (ICD). The classification followed the main disease categories except for ischaemic heart disease (IHD, ICD 4lG414) and cerebrovascular disease (CVD, ICD 43G438). Categories IV, XI, XII, XIV, XV and XVI as well as some rare external causes of death were excluded from the analyses (a total of 1332 death certificates, 3%). The death certificates did not include information on education. Oc- cupational classification was not used in the analysis, because available census data on occupation included only employed persons.

RESULTS

The male/female mortality ratio was 2.3 for natural and 5.3 for violent causes of death (Table 1). The ratio was smallest for married persons for both natural and violent causes, greater for single persons and greatest for widows and divorced persons, especially for violent causes. By social class, sex mortality ratios were lower for upper professionals, skilled workers and farmers, and higher for lower professionals and unskilled workers. The disease mortality ratios between men and women differed most for mental disorders (RR = 5.7) ischaemic heart disease (IHD; RR = 4.5) and respiratory disease (RR = 3.3; Table 2). For violent causes of the death the following mortality ratios were obtained: poisonings (RR = 18.7), drownings (RR = 12.8), fire accidents (RR = 6.4), traffic accidents (RR = 4.9) homicides (RR = 4.9), suicides (RR = 4.4) and falls (RR = 4.4; Table 3). The male/female mortality ratio varied more by social class and marital status for violent causes than for diseases [see standard devi- ations (SD) in Tables 2 and 31. The variation for disease categories was greatest for mental disorders (SD = 4.0 and 3.2) gastro-intestinal diseases (SD = 1.6 and 0.5), infectious diseases (SD = 1.1 and 1.2) and respiratory diseases (SD = 1.0 and 0.7). The smallest variation was for cerebrovascular diseases (CVD; SD = 0.2 and 0.2) and genitourinary disease

Table 1. Ratios of male and female age-adjusted (25564 years) mortality rates by marital status and social class in Finland

1969-1971

Single Married Widowed Divorced Upper professionals Lower professionals Skilled workers Unskilled workers Farmers Total

All diseases All violent causes

2.4 6.7 2.2 4.3 2.6 9.4 3.0 7.1 1.9 2.4 2.5 4.5 2.0 4.0 2.6 8.5 2.0 5.1 2.3 5.3

(SD = 0.1 and 0.3). For violent causes of death, the variation of mortality ratios by social class and marital status was greatest in poisonings (SD = 15.0 and 7.7) and drownings (SD = 4.5 and 9.5). The smallest variation was in traffic accidents (SD = 2.5 and 1.2).

DISCUSSION

Discussion of materials and methods

The reliability of the cause of death data is dependent on age and the category of death. The variation of proportion of autopsies by social class and marital status was as follows [ 131:

Age

2544 45-54 55-64

Diseases Violent causes

(%) (%)

43-79 75-100 2666 71-96 20-54 7693

The smallest autopsy rate was observed among farmers and the highest among the divorced [ 131. Diagnos- tic errors that are gender dependent will most affect the male/female ratios. The impact of such errors, however, could not be assessed. Marital status was rarely missing from the death certificate (0.1%) and was found to be erroneous in only 0.5% of cases. Social class could be wrongly classified especially for women, who may be classified according to a prior occupation or according to their husband’s occupation. The proportion of women working outside

Table 2. Ratios of male and female age-adjusted (25-64 years) mortality rates by marital status and social class in Finland 1969-1971 for disease categories

Inf Neo Endo Ment New IHD CVD Resp GI G-U

Single Married Widowed Divorced SD

Upper professionals Lower professionals Skilled workers Unskilled workers Farmers SD

Total

3.9 1.5 I.7 8.3 1.4 4.6 1.6 3.9 2.5 0.9 2.4 1.6 0.7 3.8 1.1 4.4 1.2 3.0 1.8 1 .o 4.1 2.1 0.8 9.6 0.6 4.4 1.4 3.4 5.2 0.7 4.8 1.9 2.5 13.6 2.2 4.9 1.6 5.5 4.3 1.0 1.1 0.3 0.8 4.0 0.7 0.2 0.2 I.1 1.6 0.1

3.6 1.1 1.1 1.6 1.0 4.5 1.4 2.0 1.8 0.5 3.6 1.4 1.1 4.1 1.1 5.5 1.6 2.9 2.6 0.9 2.6 1.5 0.9 6.7 1.1 3.7 1.2 3.4 1.6 0.9 4.6 2.1 0.9 8.9 2.0 5.0 1.2 3.8 2.6 1.3 1.4 1.5 0.8 1.7 1.4 3.7 1.2 2.3 2.1 0.8 1.2 0.4 0.1 3.2 0.4 0.8 0.2 0.7 0.5 0.3

2.9 1.7 1.0 5.7 I.2 4.5 1.3 3.3 2.4 1.0

Inf = infectious disease (ICD CtJt&l36). Neo = neoplasms (ICD l&239). Endo = endocrine and metabolic diseases (ICD 24&279). Ment = mental disorders (ICD 290-315). New = diseases of the nervous system (ICD 320-389). IHD = ischemic heart disease (ICD 410414). CVD = cerebrovascular disease (ICD 43&438). Resp = respiratory diseases (ICD 46&519). GI = gastrointestinal diseases (ICD 52&577). G-U = genito-urinary diseases (ICD 58&629).

Social factors and the gender difference in mortality

Table 3. Ratios of male and female age-adjusted (25-64 years) mortality rates by marital status and social class in Finland 1969-1971 for violent causes of death

Traf Poi Falls Fire Drown Sui Homi

Single 5.5 44.2 4.2 16.2 9.3 4.4 14.2 Married 4.3 9.1 3.4 3.4 13.0 3.6 2.0 Widowed 9.9 17.4 4.1 1.7 16.7 16.9 1.4 Divorced 5.1 25.8 13.3 6.0 19.7 4.4 4.2 SD 2.5 15.0 4.7 6.5 4.5 6.4 6.0

Upper professionals 2.7 24.0 5.2 2.3 3.2 2.0 0.3 Lower professionals 4.5 5.1 3.9 3.9 9.3 3.9 1.9 Skilled workers 3.7 14.1 4.2 3.1 10.4 3.0 2.3 Unskilled workers 6.0 23. I 3.6 11.6 27.6 6.5 9.7 Farmers 4.4 19.0 4.4 4.7 6.5 4.1 3.9 SD 1.2 7.7 0.6 3.7 9.5 1.7 3.6

Total 4.9 18.7 4.4 6.4 12.8 4.4 4.9

Traf = traffic accidents. Poi = accidental poisonings. Falls = accidental falls. Fire = accidental fires. Drown = drowning accidents. Sui = suicides. Homi = homicides.

607

the home is high in Finland. Among 30-49 years aged in 1970 95% of men and 67% of women belonged to the economically active population [ 141.

A recent work from England indicated that selection by social class did not cause bias in mortality differences in a follow-up after a 5-year period [lS]. It is possible that cross-sectional data is confounded by selection. If etiological factors of a disease affect the determination of marital status or social class, the results may be difficult to interpret. Such factors could be smoking, alcohol use, psychiatric dis- turbances, or somatic diseases in childhood and adolescence. Alcohol use is associated with divorce, and thus it is understandable that divorced persons have more alcohol-associated diseases [16]. Because smoking is also associated to alcohol use, the divorced have also more smoking related diseases. On the other hand marital status or social class can affect health-behavior, e.g. smoking, alcohol use, other risk-taking and self-care of own’s health. Social factors and disease undoubtedly form a recursive dy- namic process. It is difficult afterwards to determine which factors affected which and how much. Selective factors are probably not very important as regards widowed and farmers. For these groups the results indicate the health risks associated with them. If men and women are selected by the same mechanism into social class and marital status, the results for these groups indicate true gender differences in health risks.

Discussion of results

The ratio of mortality by sex may have a high score, when male mortality is high, or when female mortality is low. Mortality rates by sex for different causes of death have been published earlier [13]. The interpretation of these results is made difficult because Finnish data on the distribution of risk factors by sex, social class and marital status is sparse. The evaluation of effect of various mental risk factors (e.g. comparison of coping mechanisms of men and women) in gender difference of mortality would be relevant [17-191. Necessary data on Finnish population is not available. Data on smoking, alcohol use, obesity, serum lipids and blood pressure is available only from recent years. Dietary data is much more incomplete. The relationship of social class and marital status to leisure time physical activity and emotional stress levels is not well documented. Inter- action effects with environmental agents cannot be

assessed in this kind of study where data on individ- ual exposures is not available.

For disease categories the greatest difference between men and women was for mental disorders. In this diagnostic group alcoholism was the major cause of death (78%). The male/female mortality ratio was highest for the divorced (RR = 13.6), widowed (RR = 9.6) and single persons (RR = 8.3), as well as for unskilled workers (RR = 8.9). Thus alcoholism would seem to be a major gender-associated problem causing increased mortality especially in lonely, poorly educated men. Compared to other disease categories, IHD and CVD mortality ratios showed relatively small variation by social class and marital status. The male/female mortality ratio was 4.5 for IHD and 1.3 for CVD. The male/female ratio for IHD mortality was greatest among the lower professionals (5.5), which has the highest IHD mortality in men [20]. This is also the only social class in which the prevalence of hypertension was higher for men than women [21]. By marital status the sex ratio of prevalence of hypertension was the following [21]: single = 0.69, married = 0.50, widowed = 0.42 and divorced = 0.9 1. The variation of these ratios fit fairly well to the variation of corresponding mortality figures. Gender differences in smoking by marital status and social class are not available for the appropriate years. Sex ratios of proportion of current smokers in 1975 by social class and marital status (The Finnish Twin Cohort Study: unpublished; single = 2.6, married = 2.4, widowed = 1.9, divorced = 1.8, upper professionals = 1.9, lower professionals = 1.7, skilled workers = 2.1, unskilled workers = 3.2, farmers = 5.1) do not correlate to corresponding mortality ratios of IHD. Two prospec- tive studies [IO, 22,231 with multivariate adjustment of various risk factors indicated 2.3-2%fold gender difference in IHD mortality (unadjusted ratios: 2.2-4.8). Differences in known risk factors do not account very well for the IHD-mortality difference between men and women [22-241. In a Russian study, men had in middle-age 1.3-1.5 times more athero- sclerotic changes of a severe type in the right and left coronary arteries, as well as in the abdominal aorta, than women [25]. One would have expected the IHD rate ratio to be smaller, if IHD mortality is highly correlated to the degree of atherosclerosis. Men possibly have higher levels of other IHD risk factors (e.g. arrhythmogenic or thrombogenic factor such as

608 MARKKU KOSKENVUO et d

emotional stress and heavy alcohol use) than those associated with atherosclerosis, or men may be more sensitive to those factors. The role of smoking as a determinant of the gender difference in IHD mortality is not clear [26], even though cigarette smoking is associated with the incidence of IHD independently of genetic factors [27]. The relatively small difference between men and women in CVD mortality rates is probably due to a higher prevalence of hypertension among women [21]. The male/female ratio was also large for respiratory disease mortality (RR = 3.3) and infec?ious disease mortality (RR = 2.9). In both categories the gender difference as largest for the divorced persons and unskilled workers; heavy smoking is common in both these groups among men. Tuber- culosis was the main cause of death for infectious diseases (77% of cases), while for respiratory disease mortality the main causes of death were chronic obstructive lung disease (49%) and pneumonia (24%). Mortality from gastrointestinal disease con- sisted of many causes of death, mainly alcohol related diseases (liver cirrhosis 28%, peptic ulcer 18%, pan- creatitis lS%, gall stone disease 9% and intestinal occlusion 8%). In this disease category men had a 2.4-fold mortality compared to women. The mortality ratio was largest for the widowed (RR = 5.2) and divorced (RR = 4.3). Sex ratios of alcohol use (proportion of persons consuming more than 500 g of absolute alcohol per month) in 1975 by social class and marital status (The Finnish Twin Cohort Study: unpublished; single = 4.6, married = 3.7, widowed = 2.5, divorced = 7.5, upper professionals = 2.6, lower professionals = 2.8, skilled workers = 4.1, unskilled workers = 6.3, farmers = 5.7) did not correlate to corresponding mortality ratios of gastrointestinal diseases. For violent causes, the mortality ratio was greatest for poisonings (RR = 18.7) and accidental drownings (R = 12.8). The variation of mortality ratios by marital status and social class was large in each cause of death category. The greatest gender differences were found in unskilled workers, and in divorced and widowed. For suicides the rate ratio was greatest for the widowed (RR = 16.9). For traffic accidents one would like to know the relative par- ticipation of men and women in different types of transportation modes. The variation of mortality ratios for violent causes by sex was close to the variation for mental disorders (alcoholism). The ex- ception to this was the category of poisonings, for which the male/female ratio was great also for upper professionals (RR = 24).

Low gender difference for most violent causes in upper professionals was explained also by a relatively high mortality rate of women in the group of upper professionals [ 131.

The male/female mortality ratio in Finland has started to increase from 1.4 during 1910-1920 to 2.1 in 1956-1960 [2] and 2.8 in 1969-1971. The increase in gender mortality ratio is probably mainly caused by environmental factors. The excess mortality of Finnish men seems to be related quite clearly to factors associated with marital status and social class. Part of these factors have been identified (as smoking and heavy alcohol use), but part of them have not been well documented (as emotional stress, aggressive behavior and poor coping ability). The variation of

the sex mortality ratio by social factors speaks for external rather than biological factors in causing gender difference of mortality.

1

2

3.

4.

5.

6.

7

8

9

10

11

12

13

14

15.

16.

17.

18.

19.

20.

21.

22.

REFERENCES

Tracy R. E. Sex difference in coronary disease: two opposing views. J. chron. Dis. 19, 1245,- 1966. Piepponen P. and Ritamies M. Miesten Ylikuolevuus S&&ssa. Vlest(ipoliittisen tutkimuslaitoksen julkaisuja, Sarja B, No. 8, 1966. HBr6 A. S. Kuolleisuus Suomessa ja muissa pohjois- maissa. (Mortality in Finland and in the other Scan- dinavian countries.) Duodecim 82, 1136, 1966. Leppo K., Lindgren J. and Ritamies M. Mortalitv Trends in Finlandin the 1960s. Yearbook of Population Research in Finland, Helsinki, 1971. Occupational Mortality, Decennial supplement, 197G-1972, England and Wales. A publication of the Government Statistical Service. Series DS No. 1. London, 1978. Heller R. F., Williams H. and Sittampalam Y. Social class and ischaemic heart disease: use of the male: female ratio to identify possible occupational haz- ards. J. Epidem. Commun. Hlth 38, 198, 1984. Enterline P. E. Causes of death responsible for recent increases in sex mortality differentials in United States. Milbank Mem. Fund Q. 39, 312, 1961. Patrick C. H., Palesch Y. Y., Feinleib M. and Brody J. A. Sex differences in declining cohort death rates from heart disease. Am. J. publ. Hlih 72, 161, 1982. Levy R. I. Declining mortality in coronary heart disease. Arteriosclerosis 1, 312, 1981. Wingard D. I. The sex differential in mortality rates demographic and behavioral factors. Am. J. Epidem. 115, 205, 1982. Waldron I. Sex differences in illness incidence, progno- sis and mortality: issues and evidence. Sot. Sci. Med. 17, 1107, 1983. Lopez A. D. Sex differentials in mortality. WHO Chron. 38, 217, 1984. Koskenvuo M., Sarna S., Kaprio J. and Lannqvist J. Cause-specific mortality by marital status and social class in Finland during 1969-1971. Sot. Sci. Med. 13A, 691, 1979. Statistical Yearbook of Finland, 1972. Central Statistical Office of Finland, Helsinki, 1973. Fox A. J., Goldblatt P. 0. and Jones D. R. Social class mortality differentials: artefact, selection or life circum- stances? J. Epidem. Commun. Hlth 39, 1, 1985. Poikolainen K. Risk of alcohol-related hospital admis- sion in men as predicted by marital status and social class. J. Stud. Alcohol 44, 986, 1983. McQueen D. V. and Siegrist J. Social factors in the etiology of chronic disease: an overview. Sot. Sci. Med. 16, 353, 1982. Henry J. P. The relation of social to biological processes in disease. Sot. Sci. Med. 16. 369. 1982. Cohen S. and Syme S. L. ‘Issues in the study and application of social support. In Social Support and Healrh (Edited by Cohen S. and Syme S. L.), p. 3. Academic Press, Orlando, 1985. Koskenvuo M., Kaprio J., KesLniemi A. and Sarna S. Differences in mortality from ischemic heart disease by marital status and social class. J. chron. Dis. 33, 95, 1980. Aromaa A. Kohonnut Verenpaine ja sen Kan- santerveydellinen Merkitys Suomessa. (Epidemiology and Public Health Impact of High Blobd Pressure in Finland.) Kansaneliikelaitoksen iulkaisuia AL: 17, Helsinki, 1981. Friedman G. D., Dales L. G. and Ury H. K. Mortality

Social factors and the gender difference in mortality 609

in middle-aged smokers and nonsmokers. New Engl. J. 25. Vihkert A. M. and Zhdanov V. S. The Effects of Various Med. 300, 213, 1979. Diseases on the Development of Atherosclerosis. Per-

23. Wingard D. L., Suarez L. and Barrett-Connor E. The gamon Press, Oxford, 1981. sex differential in mortalitv from all causes and ischemic 26. Kleinman J. C.. Feldman J. J. and Monk M. A. The heart disease. Am. J. E&em. 117, 165, 1983. effects of changes in smoking habits on coronary heart

24. McGill H. C. Atherosclerosis: problems in endpoints disease mortality. Am. J. publ. Hlth 69, 795, 1979. for genetic analysis. In Genetic Analysis of Common 27. Kaprio J. The Incidence of Coronary Heart Disease in Diseases: Applications to Predictive Factors in Coronary Twin Pairs Discordant for Cigarette Smoking. Kan- Disease (Edited by Sing C. F. and Skolnick M.), p. 27. santerveystieteen julkaisuja M, 84, Helsinki, 1984. Alan R. Liss, New York, 1979.

social factors and the gender difference in mortality

Documents