Abstract
Objectives
The life expectancy gap between Central‐Eastern European (CEE) countries, including Hungary, and Western Europe (WE) is mainly attributable to excess cardiovascular (CV) mortality in midlife. This study explores the contribution of socioeconomic, work related, psychosocial, and behavioural variables to explaining variations of middle aged male and female CV mortality across 150 sub‐regions in Hungary.
Design
Cross sectional, ecological analyses.
Setting
150 sub‐regions of Hungary.
Participants and methods
12 643 people were interviewed in Hungarostudy 2002 survey, representing the Hungarian population according to sex, age, and sub‐regions. Independent variables were income, education, control in work, job insecurity, weekend working hours, social support, depression, hostility, anomie, smoking, body mass index, and alcohol misuse.
Main outcome measures
Gender specific standardised premature (45–64 years) total CV, ischaemic heart disease, and cerebrovascular mortality rates in 150 sub‐regions of Hungary.
Results
Low education and income were the most important determinants of mid‐aged CV mortality differences across sub‐regions. High weekend workload, low social support at work, and low control at work account for a large part of variation in male premature CV mortality rates, whereas job insecurity, high weekend workload, and low control at work contribute most noticeably to variations in premature CV mortality rates among women. Low social support from friends, depression, anomie, hostility, alcohol misuse and cigarette smoking can also explain a considerable part of variations of premature CV mortality differences.
Conclusion
Variations in middle aged CV mortality rates in a rapidly changing society in CEE are largely accounted for by distinct unfavourable working and other psychosocial stress conditions.
Keywords: premature cardiovascular mortality, work stress, depression, anomie, alcohol misuse
The gap in life expectancy between Central‐Eastern European (CEE) and Western Europe (WE) continues to be an important public health problem.1,2 This gap increased most dramatically over the past three decades and it is largely attributable to excess cardiovascular (CV) mortality in midlife, and particularly among men. Up to the end of the 1970s, mortality rates in Hungary were lower than those found in Great Britain or Austria. Subsequently, mortality rates declined in WE. But in Hungary and in other CEE countries this trend was reversed, especially among middle aged men. In the 1990, the mortality rates among 45–64 year old men in Hungary rose to and remained at levels that were higher than they were in the 1930s.2 In 1999 the standardised death rates attributable to diseases of the circulatory system were three times higher among Hungarian men and women that the European average, but were 2.8 times higher among men than among women.3 Besides health damaging behaviours, socio economic and psychosocial factors might contribute to this health deterioration in a substantial way.4 Although poverty is a powerful determinant of population health it is unlikely that it provides an important explanation of this change. This is obvious, among others, from the Hungarian case, where the increase of premature mortality occurred at a time when average living standards were raising, and when health care delivery was improved.5 Between 1970 and 1990 an almost threefold increase of gross domestic product was witnessed in Hungary and at the same time, mortality rates increased from 12.5 to 15.5 per thousand in men and from 10.8 to 12.8 in women.2
It may be that the dynamics of sudden socioeconomic changes affect health by less obvious threats and challenges to people's everyday lives, such as the experience of increasing social disparities,6,7 or the weakening and breakdown of sociocultural assets, social networks, and social capital.8,9 The transformation in CEE countries created conditions of loss of control over life, economic deprivation, and social isolation that might undermine the health status of the population. This disproportionately affects the working age population because their work and family support roles rendered them more vulnerable to socioeconomic disruption, which, in the CEE countries, were not buffered by a well functioning civic society.4,8 The factors that might undermine health may differ according to gender and socioeconomic status. Differences in gender roles may make different aspects of life more or less important, or there may be differences in meaningfulness of different factors of the socioeconomic and psychosocial environment.4,10 Working conditions may play an important part in this regard, given the centrality of work as a risk factor in midlife.11,13
Hungary has witnessed an important change in the labour market since 1970. Until the mid‐1970s everybody had to have one working place and there was no possibility for more jobs. Afterwards the first development of the “new economic mechanism” was the initiation of the so called second or third economy, which meant the possibility of better economic situation by accepting or creating second or third jobs by the expenses of considerable overwork. These developments were more frequent among groups in lower socioeconomic strata, as shown by lower educational degree or lower income, because they often dispose of fewer choices of alternatives. It went along with high degree of job instability and associated loss of control.7
This paper aims at exploring the role of adverse working and other psychosocial conditions in association with midlife CV mortality in Hungarian men and women in the frame of an ecological analysis. The mortality rates are available only on ecological level, therefore the relative importance of risk factors can be analysed only on aggregate level.
The relative importance of work related conditions in comparison with other established or hypothesised psychosocial and behavioural risk factors were also analysed, separately among men and women. It is worthwhile to hypothesise that the established psychosocial and behavioural risk factors, such as depression,14 health damaging behaviour, in particular cigarette smoking,15 body mass index (BMI),15 and alcohol misuse,16 hostility,17 and social support from family and friends18 all contribute to the explanation of regional differences in premature mortality. Additionally, a distinct psychosocial variable, the anomie score19 was included, because in a suddenly changing society the inability of long term planning might contribute substantially to health deterioration.
Adverse working conditions were shown to predict an increased risk of a variety of stress related disease conditions. Up to now, most robust evidence concerns CV morbidity and mortality.12,20,21,22,23,24,25
Gender differences in association with work and midlife CV mortality are also explored. This question is important in view of the well reported gender gap in midlife CV morbidity and mortality1,2,3,26 (table 1). Yet, given the fact that the overall CV mortality rates in Hungary are three times higher than the average in WE countries among women as well,3 an exploration of the high burden of premature CV mortality in Hungarian women is also needed, especially so as larger female CV mortality differences across 150 sub‐regions are reported than among men26 (table 1). As a substantial proportion of women take part in the labour market in Hungary, it is of interest to analyse to what extent adverse working conditions are associated with midlife mortality in women.
Table 1 Descriptives of the variables of the study for the 150 sub‐regions of Hungary. Standardised CV* middle aged (45–64 years) mortality rates and average values for the 150 sub‐regions of Hungary from Hungarostudy 2002 (n = 150).
| Descriptive statistics | Minimum | Maximum | Mean | SE |
|---|---|---|---|---|
| Male total CV* mortality | 408.00 | 939.80 | 636.69 | 9.85 |
| Male IHD† mortality | 91.40 | 541.40 | 319.81 | 6.38 |
| Male cerebrovascular mortality | 48.30 | 316.80 | 160.83 | 4.07 |
| Female CV* mortality | 81.50 | 371.50 | 222.31 | 4.89 |
| Female IHD† mortality | 18.00 | 198.50 | 96.96 | 2.89 |
| Female cerebrovascular diseases | 18.00 | 160.80 | 67.99 | 2.22 |
| Control at work | 0.28 | 1.91 | 1.12 | 2.69E–02 |
| Social support at work | 0.50 | 2.00 | 1.23 | 2.07E–02 |
| Job security | 0.47 | 1.82 | 1.35 | 1.91E–02 |
| Week day work hours | 5.57 | 9.76 | 7.71 | 6.05E–02 |
| Weekend work hours | 4.00 | 8.68 | 6.09 | 6.77E–02 |
| Personal income | 1.11 | 1.94 | 1.44 | 1.40E–02 |
| Education | 2.50 | 4.21 | 3.46 | 3.61–02 |
| Subjective social status | 2.73 | 4.69 | 3.81 | 3.27E–02 |
| Unemployment rate | 1.60 | 20.7 | 5.92 | 0.30 |
| Shift work (%) | 0.00 | 37.2 | 13.5 | 5.21–03 |
| Beck depression score | 2.74 | 14.17 | 8.37 | 0.17 |
| Hostility score | 2.26 | 6.40 | 4.01 | 5.89E–02 |
| Anomie score | 2.49 | 7.28 | 5.08 | 6.33E–02 |
| Social support from friends | 1.10 | 2.40 | 1.73 | 1.96E–02 |
| Social support from parents | 1.00 | 2.65 | 1.77 | 2.85E–02 |
| Social support from partner | 1.35 | 3.00 | 2.27 | 1.87–02 |
| Social support from civic organisations | 0.00 | 1.11 | 0.45 | 1.29–02 |
| Pack/year cigarettes | 25.85 | 148.05 | 83.48 | 1.97 |
| Non‐stop drinking | 0.00 | 0.36 | 3.67E‐02 | 4.72E–03 |
| Morning drinking | 0.00 | 0.30 | 3.97E‐02 | 4.69E–03 |
| Self blame because of drinking | 0.00 | 0.32 | 5.59E‐02 | 4.64E–03 |
| Spirit consumption (0.5 dl) | 0.00 | 1.10 | 0.35 | 1.43–02 |
| BMI | 24.6 | 28.5 | 26.2 | 6.48–02 |
*CV, total cardiovascular mortality rates (ICD 10.I00–I99). †IHD, ischaemic heart disease mortality (ICD 10. I20–I25). Cerebrovascular diseases (ICD 10.I60–I69).
Methods
The Hungarostudy 2002 is a national cross sectional survey representing the adult Hungarian population according to sex, age, and according to 150 sub‐regions.10,27 The sampling methods were described in detail elsewhere.10,27,28 Altogether 12 643 persons were interviewed in their homes. The sample represented 0.16% of the population above age 18. The refusal rate was 17.7%.
Standardised mortality data
Age standardised midlife (45–64 years) male and female mortality rates for the main causes of death for 10 000 persons in the same sex and age group were computed by the Central Statistical Office of Hungary for each of the 150 Hungarian sub‐regions for the past available years, between 2001 and 2003.26 The standardised premature (45–64 years) total CV (ICD 10, I00–I99), IHD (ICD 10, I20–I25), and cerebrovascular (ICD 10, I60–I69) mortality rates were included as outcome measures in these analyses.
Measures
Work stress
Control at work was assessed by Likert scaled answers (0 to 3) to the item “How much can you influence what happens in your working group?”7Social support at work was measured by answers (0–3) to the item “How much help do you receive from coworkers?”7,27 The number of working hours per week days and on weekend days were recorded. Job security was assessed by Likert scaled answers (0 to 2) to the item “I am happy with my level of job security”.29 Unemployment rates and shift work were also recorded.
Socioeconomic and further psychosocial variables
Education was measured by the highest education grade on a six level score, personal income was assessed as average of personal income on the basis of eight income categories. Subjective social status10 was measured by a 10 rung scale.
To measure the presence and severity of depressive symptoms the shortened Beck depression inventory (BDI) was applied.5,7,27 The BDI is a validated and reliable screening instrument for assessing the severity of depression in population surveys.
Social support outside work was assessed by answers to items on perceived support from partners, parents, friends, and civic or religious organisations separately.7,27 To measure hostility a shortened version of the Cook‐Medley scale was used.7
Anomie was assessed by the four item shortened anomie questionnaire of the Eurobarometer survey.19 The anomie values increased significantly in the past decades in Hungary, and this measure can be an important determinant of health deterioration.19
Information on health damaging behaviour was measured by cigarette smoking (smoking status and average number of pack years per person). BMI was measured on the basis of reported height and weight of the respondents. Alcohol misuse was measured by the AUDIT alcohol abuse score.30 From these score three items seem to be most informative: During the past year how often: (1) have you found that you were not able to stop drinking once you had started? (2) have you needed a first drink in the morning? (3) have you had a feeling of guilt or remorse after drinking? The quantity of alcohol consumption was also recorded.
All analyses were performed based on the SPSS Base 9 programme.
The study was approved by the ethics committee of the Semmelweis University and conforms to the principles embodied in the declaration of Helsinki.
Results
Table 1 shows the descriptive statistics for the variables included into the analyses. There are considerable variations in different forms of premature CV mortality differences and also in work related, other psychosocial and socioeconomic conditions, and behavioural variables among 150 sub‐regions.
Table 2 shows correlations of the four clusters of variables (work related variables; additional psychosocial risk factors; socioeconomic factors; behavioural variables) with premature male and female CV mortality rates, and separately with ischaemic heart disease and cerebrovascular mortality rates. Concerning adverse psychosocial work environment, low control and low social support at work were strongly associated with premature CV mortality rates in both sexes, although there were gender differences. Weekend workload was most closely connected with male, while job insecurity was most closely associated with female CV mortality.
Table 2 Correlation coefficients of associations between work related, other psychosocial, socio‐economic, and behavioural variables and standardised midlife (45–64 years) male and female CV* mortality rates in 2001–2003 in 150 Hungarian sub‐regions (n = 150) (data weighted according to sample size in the sub‐regions).
| Male CV* | Male IHD† | Male cerebrovascular | Female CV* | Female IHD† | Female cerebrovascular diseases | |
|---|---|---|---|---|---|---|
| Work related variables: | ||||||
| Control at work | −0.255*** | −0.193* | −0.249** | −0.275** | −0.334*** | −0.188* |
| Social support at work | −0.197* | −0.213** | −0.179* | −0.236** | ||
| Job security | −0.220** | −0.193* | −0.304*** | −0.273** | −0.234** | |
| Weekend work hours | 0.344*** | 0.205* | 0.350*** | 0.225** | 0.200* | |
| Week day work hours | 0.182* | |||||
| Shift work | 0.171* | |||||
| Significant further psychosocial variables: | ||||||
| Hostility | 0.257** | 0.278** | 0.229** | 0.211** | 0.205** | |
| Depression | 0.352*** | 0.198* | 0.364***. | 0.331*** | 0.274** | 0.280** |
| Anomie | 0.340*** | 0.183* | 0.333*** | 0.287*** | 0.225** | 0.171* |
| Social support from parents | −0.222 | −0.236** | ||||
| Social support from friends | −0.372*** | −0.285*** | −0.363*** | −0.345*** | −0.360*** | −0.287*** |
| SE variables | ||||||
| Education | −0.599*** | −0.385*** | −0.592*** | −0.527*** | −0.409*** | −0.460*** |
| Subjective social status | −0.353*** | −0.226** | −0.336*** | −0.303*** | −0.301*** | −0.173* |
| Personal income | −0.512*** | −0.372*** | −0.474*** | −0.402*** | −0.369*** | −0.324*** |
| Unemployment rate | 0.465*** | 0.296*** | 0.491*** | 0.378*** | 0.373*** | 0.271** |
| Behavioural variables: | ||||||
| Pack/year cigarettes | 0.188* | 0.162* | 0.151* | 0.153* | ||
| Non‐stop alcohol | 0.288*** | 0.206* | 0.283*** | 0.313*** | 0.276** | 0.268** |
| Morning alcohol | 0.266** | 0.165* | 0.287*** | 0.224** | 0.229** | 0.223** |
| Self blame because of alcohol | 0.250** | 0.189* | −0.228** | 0.188 | ||
| BMI | 0.192* | NS | 0.195* | 0.205* | NS | 0.241** |
*CV, total cardiovascular mortality rates (ICD 10.I00–I99). †IHD, ischaemic heart disease mortality (ICD 10. I20–I25). *p<0.05; **p<0.01; ***p<0.001.
As anticipated, strong associations of measures of SE status with mortality rates were seen, first of all between low education and premature CV mortality, but low personal income and unemployment rates were also closely connected with CV mortality rates. Moreover, the additional psychosocial risk factors explored in this study showed substantial correlations with mortality rates, in particular lack of social support from friends, anomie, depression, and hostility. Associations with social support from parents, partners, civic and religious organisations were much weaker. The correlation coefficients with cigarette smoking, BMI, and three items of alcohol misuse were significantly connected with premature CV mortality rates, but the reported quantity of spirit consumption showed no significant correlation with CV mortality.
Education and income were significantly and in decreasing order associated with control at work, weekend working hours, job security, social support from friends, depression, anomie, BMI, social support from coworkers, hostility, pack cigarettes per year, and alcohol misuse. Week day working hours were shorter only among the more educated people, but they were not significantly connected with income. Week day and weekend work hours were significantly interrelated (r = 0.608, p = 0.000), which means overwork among the same persons not only at weekends but during the weekdays as well. Shift work correlated with week day working hours, but not with weekend working hours.
Finally, multivariable linear regression analyses were conducted with all those variables that reported significant associations with premature CV mortality rates in bivariate analyses. Because of the high level of colinearity of the variables, there was a need for three separate models. In the first model, the five variables measuring an adverse psychosocial work environment were included and their combined effects were analysed. Among men weekend work hours and low social support from coworkers explained 14.7%, among women low job security and weekend work hours explained 10.9% of the total premature CV mortality variations among sub‐regions. In the case of male cerebrovascular mortality 14.4% of variance could be explained by work related factors such as by weekend work hours and by low control at work. It is important to note that the list of variables entering the models differed between men and women (table 3).
Table 3 Results of multivariable linear regression analysis of work related variables (control at work, social support at work, job security, weekend work hours, and week day work hours) in explaining differences in midlife (45–64 years) male and female total CV*, IHD†,and cerebrovascular disease mortality (those variables which entered into the models).
| Standardised β | t Value | Significance | Adjusted R2 for the final model | |
|---|---|---|---|---|
| Total male CV* mortality | ||||
| Weekend work hours | 0.346 | 4.58 | <0.000 | 0.112 |
| Social support at work | −0.201 | −2.66 | <0.000 | 0.147 |
| Total female CV* mortality | ||||
| Job security | −0.271 | −3.44 | <0.000 | 0.087 |
| Weekend work hours | 0.172 | 2.23 | <0.000 | 0.109 |
| Male IHD† mortality | ||||
| Social support at work | −0.215 | −2.78 | <0.007 | 0.039 |
| Weekend work hours | 0.207 | 2.68 | <0.009 | 0.076 |
| Female IHD† mortality | ||||
| Control at work | −0.334 | −4.31 | <0.000 | 0.106 |
| Male cerebrovascular mortality | ||||
| Weekend work hours | 0.312 | 4.09 | <0.000 | 0.117 |
| Control at work | −0.184 | −2.38 | <0.019 | 0.144 |
| Female cerebrovascular mortality | ||||
| Job security | −0.263 | −3.29 | <0.001 | 0.048 |
| Week day work hours | 0.175 | 2.33 | <0.030 | 0.072 |
*CV, total cardiovascular mortality rates (ICD 10.I00–I99). †IHD, ischaemic heart disease mortality (ICD 10. I20–I25).
What is already known about this topic
Low education and low income are important determinants of CV mortality
Adverse working conditions predict an increased risk of CV morbidity and mortality
Low social support, depression, hostility, alcohol misuse, and cigarette smoking are cardiovascular risk factors
What this study adds
High weekend workload, low social support at work, and low control at work account for a large part of variation in male premature CV mortality rates
Job insecurity, high weekend workload, and low control at work contribute considerably to variations in premature CV mortality rates among women
Low social support from friends, depression, anomie, hostility, alcohol misuse, and cigarette smoking can explain a considerable part of variations of premature CV mortality differences within Hungary
Model 2 includes depression, social support from friends, anomie, and hostility. Although depression and hostility showed significant correlations with premature mortality rates, in the multiple regression models low social support from friends and in the case of total CV mortality high anomie entered significantly into the models. These two variables explained 18.4% among the male and 15.1% among the female total CV mortality rates. Low social support from friends alone explained 12.4% of female ischaemic heart disease variations and 19% of male cerebrovascular mortality differences (table 4).
Table 4 Results of multivariable linear regression analysis of other psychosocial variables (depression, hostility, anomie, social support from friends) in explaining differences in midlife (45–64 years) male and female total CV*, IHD†, cerebrovascular and other heart disease mortality (n = 150) (those variables that entered in to the models).
| Standardised β | t Value | Significance | Adjusted R2 for the final model | |
|---|---|---|---|---|
| Total male CV* mortality | ||||
| Social support from friends | −0.307 | −.4.05 | <0.000 | 0.190 |
| Anomie | 0.264 | 3.48 | <0.001 | 0.184 |
| Total female CV* mortality | ||||
| Social support from friends | −0.292 | −.3.75 | <0.000 | 0.113 |
| Anomie | 0.216 | 2.77 | <0.006 | 0.151 |
| Male IHD† mortality | ||||
| Social support from friends | −0.285 | −3.62 | <0.000 | 0.075 |
| Female IHD† mortality | ||||
| Social support from friends | −0.360 | −4.69 | <0.000 | 0.124 |
| Male cerebrovascular mortality | ||||
| Social support from friends | −0.299 | −3.92 | <0.000 | 0.190 |
| Female cerebrovascular mortality | ||||
| Social support from friends | −0.204 | −2.35 | <0.020 | 0.076 |
| Depression | 0.190 | 2.20 | <0.029 | 0.100 |
*CV, total cardiovascular mortality rates (ICD 10.I00–I99). †IHD, ischaemic heart disease mortality (ICD 10. I20–I25).
Next, in view of substantial colinearity between socioeconomic and other psychosocial factors, the independent associations of income, education, subjective social status, unemployment, smoking, and alcohol misuse with premature CV mortality were analysed. In agreement with the literature1,2,5,7 the explanatory power of these variables was comparatively strongest, 31.6% of the total male and 25.3% of the total female CV mortality variations could be explained. In both sexes alcohol misuse (not able to stop drinking after started) was a highly significant determinant of total, ischaemic heart disease, and cerebrovascular mortality differences. In regression models unemployment rate explained significantly the female ischaemic and male cerebrovascular mortality variations. According to causes of death these factors were in the closest connection with cerebrovascular mortality differences—that is they explained 31.5% of variance (table 5).
Table 5 Results of multi‐variable linear regression analysis of education, personal income, unemployment rate, smoking, and alcohol misuse in explaining differences in midlife (45–64 years) male and female total CV*, IHD†, cerebrovascular and other heart disease mortality (those variables that entered into the models).
| Standardised β | t Value | Significance | Adjusted R2 for the final model | |
|---|---|---|---|---|
| Total male CV* mortality | ||||
| Income | −0.301 | −2.87 | <0.005 | 0.257 |
| Non‐stop drinking | 0.192 | 2.77 | <0.006 | 0.297 |
| Education | −0.237 | −2.23 | <0.027 | 0.316 |
| Total female CV* mortality | ||||
| Education | −0.415 | −5.74 | <0.000 | 0.209 |
| Non‐stop drinking | 0.226 | 3.18 | <0.002 | 0.253 |
| Male IHD† mortality | ||||
| Income | −0.348 | −4.55 | <0.000 | 0.132 |
| Non‐stop drinking | 0.153 | 2.00 | <0.047 | 0.149 |
| Female IHD† mortality | ||||
| Education | −0.219 | − 2.30 | <0.023 | 0.142 |
| Non‐stop alcohol | 0.193 | 2.56 | <0.011 | 0.177 |
| Unemployment | 0.199 | 2.14 | <0.037 | 0.196 |
| Male cerebrovascular mortality | ||||
| Education | −0.289 | −2.68 | <0.001 | 0.245 |
| Unemployment rate | 0.277 | 2.60 | <0.002 | 0.292 |
| Non‐stop drinking | 0.170 | 2.45 | <0.015 | 0.315 |
| Female cerebrovascular mortality | ||||
| Education | −0.349 | − 4.60 | <0.000 | 0.146 |
| Non‐stop drinking | 0.195 | 2.60 | <0.011 | 0.177 |
*CV, total cardiovascular mortality rates (ICD 10.I00–I99). †IHD, ischaemic heart disease mortality (ICD 10. I20–I25).
Discussion
This study found strong associations of variables defining an adverse psychosocial work environment and other socioeconomic and psychosocial variables with variations of premature CV mortality rates by 150 sub‐regions among middle aged men and women in Hungary. Although based on ecological analysis, results are of interest in view of a substantial increase in premature male and female CV mortality in recent years in Hungary and in view of sudden socioeconomic changes. Hungary had witnessed higher income inequality during these years compared with neighbouring CEE countries.4 This means stronger socioeconomic deprivation in lower socioeconomic strata. During the past decades there was a considerable decrease in perceived control in work and in perceived social support from coworkers and friends, more so in the lowest socioeconomic strata of the society.7 According to these results adverse working conditions, low social support from friends, depression, hostility, anomie, smoking, BMI, and alcohol misuse were significantly connected with low education and income.
In the middle aged population work and working conditions are central in determining health.31 An independent predictive role of work stress on male and female CV mortality, based on individual level data, was found in a recent prospective epidemiological investigation in Finland.23 Similarly, incident fatal and non‐fatal coronary heart disease, the major contributor to CV mortality, was predicted by work related stress in other epidemiological studies.22 These results are interpreted as support of an important role of long term psychosocial stress in the development of CV diseases.5,31,32,33
Policy implications
In the population of CEE countries exposed to far reaching socioeconomic changes a cluster of stressful, work related, and other psychosocial conditions accounts for substantially increased premature CV morbidity and mortality rates.
An integrated public health approach is needed to identify and reduce the impact of adverse bio‐psycho‐social conditions on premature CV mortality.
Given gender differences in psychosocial stress at work, preventive programmes should take these into account.
Improved occupational mental health services are needed to prevent a growing health burden in CEE countries.
In this study associations of work related stress with variation in CV mortality were found for both sexes although the contribution of single components differed by gender. Although weekend workload seems to be an important risk factor in both sexes, the pressure to work on the weekends without pause, matters more for Hungarian men,10 while low job security seems to be more important for women. Although these gender specific constellations might reflect the particularly stressful situation of the rapid socioeconomic changes that occurred in CEE it should be mentioned that considerable inconsistency exists in international research on gender variation in associations between psychosocial stress at work and CV morbidity or mortality.34,35
An additional research question concerns the relative importance of work related conditions, compared with other psychosocial variables, in accounting for CV mortality differences. Social support outside work, depression and hostility are three well established psychosocial factors explaining increased CV morbidity and mortality.36 In our study anomie, the inability for long term planning in combination with feelings of societal demoralisation seems to be an important CV risk factor as well. In bivariate analysis, each one of these factors exhibits strong associations with mortality variation across sub‐regions. This finding is consistent with a growing body of evidence linking depression to CV morbidity and mortality.5,14 Given close associations between socioeconomic, work related, and additional psychosocial factors, it is tempting to hypothesise that these latter conditions act as mediators of the link between low socioeconomic status and premature CV mortality. It is through high effort—that is, overwork at the weekends as well—low personal income and low job security, low social support from coworkers and friends, low control at work, anomie, and hostility that low socioeconomic status, as measured by education, exerts its “toxic” effects on health. These conditions go along with depressive symptoms that aggravate the health burden of chronic psychosocial stress.5
Among established behavioural risk factors alcohol misuse, cigarette smoking, and BMI were significantly associated with premature CV mortality variations in our study as well. In agreement with the literature,2 the reported quantity of alcohol consumption was not significantly associated with premature CV mortality. Inability to stop drinking after starting, as a measure of alcohol misuse, seems to be particularly relevant in our sample. Alcohol misuse is a highly prevalent public health problem in Hungary. The alcohol related hepatic cirrhosis mortality rates increased dramatically in the past decades in Hungary, and it is currently the highest one in Europe.37
While smoking and alcohol misuse are established CV risk factors,38 their valid and specific measurement is still debated. More elaborate individual data based analyses might also find that an adverse psychosocial work environment results in an increase of unhealthy behaviours, and that both direct and indirect effects on CV morbidity and mortality are expected.36,38,39 A recent prospective study showed that effort‐reward imbalance at work predicts alcohol dependence in men.40
Despite the limitations of the ecological study of a population exposed to far reaching socioeconomic change we show that a cluster of stressful, work related, and other psychosocial conditions account for a substantial part of variation in middle aged male and female premature CV mortality differences across Hungarian sub‐regions. An integrated public health approach is needed to understand and prevent the bio‐psycho‐social determinants of premature mortality crisis in CEE.
Acknowledgements
The authors thank the other members of the “Hungarostudy 2002” team (János Réthelyi, Adrienne Stauder, Csilla Csoboth, György Gyukits, János Löke, Andrea Ódor, Katalin Hajdu, Csilla Raduch, András Székely, László Szücs, Sándor Rózsa), the network of district nurses for the home interviews, Professor András Klinger for the sampling procedure, and the National Population Register for the selection of the sample.
Abbreviations
CEE - Central‐Eastern European
WE - Western Europe
CV - cardiovascular
BMI - body mass index
Footnotes
Funding: this study was supported by OTKA TS‐40889 (2002), TS‐40889 (2002), TS049785 (2004) and, NKFP 1/002/2001 and NKFP 1b/020/2004 grants.
Conflicts of interest: none.
References
- 1.Marmot M, Bobak M. Psychosocial and biological mechanisms behind the recent mortality crisis in Central and Eastern Europe. In: Cornia AG, Panicci R, eds. The mortality crisis in transitional economies. Oxford: Oxford University Press, 2001
- 2. Weidner G, Kopp MS, Kristenson M, eds. Heart disease: environment, stress and gender. Amsterdam: IOS Press, 2002
- 3.Gárdos É.Health Status of Population. Budapest: Central Statistical Office, 2002
- 4.Hertzman C, Siddiqi A, Bobak M. The population health context for gender,stress and cardiovascular disease in Central and Eastern europe. In: Weidner G, Kopp MS, Kristenson M, eds. Heart disease: environment, stress and gender. Amsterdam: IOS Press, 200215–25.
- 5.Kopp M S, Réthelyi J. Where psychology meets physiology: chronic stress and premature mortality‐ the Central‐EE health paradox. Brain Res Bull 200462351–367. [DOI] [PubMed] [Google Scholar]
- 6.Rose R. How much does social capital add to individual health? A survey study of Russia. Soc Sci Med 2000511421–1435. [DOI] [PubMed] [Google Scholar]
- 7.Kopp M S, Skrabski Á, Szedmák S. Psychological risk factors, inequality and self‐rated morbidity in a changing society. Soc Sci Med 2000511350–1361. [DOI] [PubMed] [Google Scholar]
- 8.Skrabski A, Kopp M, Kawachi I. Social capital in a changing society: cross sectional associations with middle aged female and male mortality rates. J Epidemiol Community Health 200357114–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Watson P. Explaining rising mortality among men in Eastern Europe. Soc Sci Med 199541923–934. [DOI] [PubMed] [Google Scholar]
- 10.Kopp M S, Skrabski Á, Kawachi I.et al Low socioeconomic status of the opposite gender is a risk factor for middle aged mortality. J Epidemiol Community Health 200559675–678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pikhart H, Bobak M, Siegrist J.et al Psychosocial work characteristics and self‐rated health in four post‐communist countries. J Epidemiol Community Health 200155624–630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Siegrist J. Effort‐reward imbalance at work and health. In: Perrewe P, Gater D, eds. Historical and current perspectives on stress and health. Oxford: Elsevier, 2002261–291.
- 13.Karasek R A, Theorell T.Healthy Work. Stress, productivity, and the reconstruction of working life. New York: Basic Books, 1990
- 14.Rudisch B, Nemoroff C B. Epidemiology of comorbid coronary artery disease and depression. Biol Psychiatry 200354227–240. [DOI] [PubMed] [Google Scholar]
- 15.The World Health Organisation MONICA Project Ecological analysis of the association between mortality and major risk factors of CV disease. Int J Epidemiol 199423505–516. [DOI] [PubMed] [Google Scholar]
- 16.Bobak M, McKee M, Rose R.et al Alcohol consumption in a national sample of the Russian population. Addiction 199994857–866. [DOI] [PubMed] [Google Scholar]
- 17.Williams R B, Schneiderman N. Resolved: psychosocial interventions can improve clinical outcomes in organic disease. Psychosom Med 200264552–557. [DOI] [PubMed] [Google Scholar]
- 18.Kawachi I, Berkman L F. Social cohesion, social capital, and health. In: Berkman LF, Kawachi I, eds. Social epidemiology. New York: Oxford University Press, 2000174–190.
- 19.Spéder Z S, Paksi B, Elekes Z S. Anomie and stratification at the beginning of the nineties. In: Kolosi T, Tóth IGY, Vukovich GY, eds. Social report 1998. Budapest: Social Research informatics Center, 1999483–505.
- 20.Karasek R, Brisson C, Kawakami N.et al The job content questionnaire: an instrument for internationally comparative assessments of psychosocial job characteristics. J Occup Health Psychol 19983322–355. [DOI] [PubMed] [Google Scholar]
- 21.Siegrist J. Adverse health effects of high effort ‐ low reward conditions at work. J Occup Health Psychol 1996127–43. [DOI] [PubMed] [Google Scholar]
- 22.Marmot M, Theorell T, Siegrist J. Work and coronary heart disease. In: Stansfeld S, Marmot M, eds. Stress and the heart. Psychosocial pathways to coronary heart disease. London: BMJ Books, 200250–71.
- 23.Kivimäki M, Leino‐Arjas P, Luukkonen R.et al Work stress and risk of CV mortality: prospective cohort study of industrial employees. BMJ 2002325857–862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kuper H, Singh‐Manoux A, Siegrist J.et al When reciprocity fails: effort‐reward imbalance in relation to CHD and health functioning within the Whitehall II study. Occup Environ Med 200259777–784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kuper H, Marmot M. Job strain, job demands, decision latitude, and risk of coronary heart disease within the Whitehall II study. J Epidemiol Community Health 200357147–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Klinger A.Age‐standardised midlife (45–64 years) male and female CV mortality rates in the 150 Hungarian sub‐regions in 2001–2003. Budapest: Central Statistical Office, 2004
- 27.Skrabski Á, Kopp M S, Sándor Rózsa.et al Life meaning: an important correlate of health in the Hungarian population. Int J Behav Med 20051278–85. [DOI] [PubMed] [Google Scholar]
- 28.Rózsa S, Réthelyi J, Stauder A.et al The health status of the Hungarian population according to the Hungarostudy 2002: the methods and the characteristics of the representative sample. (In Hungarian). Psychiatria Hungarica 20031883–94. [Google Scholar]
- 29.Rahe R H, Tolles R L. The brief stress and coping inventory: a useful stress management instrument. Int J Stress Management 2002961–70. [Google Scholar]
- 30.Picinelli M, Tessari E, Bartolomasi M.et al Efficacy of the alcohol use disorders identification test as a screening tool for hazardous alcohol intake and related disorders in primary care: a validity study. BMJ 1997314420–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Levi L.Guidance of work related stress: Spice of life or kiss of death. Brussels: European Comission, 2000
- 32.Steptoe A, Willemsen G. Psychophysiological responsivity in coronary heart disease. In: Stansfeld S, Marmot M, eds. Stress and the heart. Psychosocial pathways to coronary heart disease. London: BMJ Books, 2002168–180.
- 33.McEwen B S. Protective and damaging effects of stress mediators. N Engl J Med 1998338171–179. [DOI] [PubMed] [Google Scholar]
- 34.Lee S, Colditz G, Berkman L.et al A prospective study of job strain and coronary heart disease in US women. Int J Epidemiol 2002311147–1153. [DOI] [PubMed] [Google Scholar]
- 35.Peter R, Siegrist J, Hallqvist J.et al Psychosocial work environment and myocardial infarction: Improving risk estimation by combining two alternative job stress models in the SHEEP study. J Epidemiol Community Health 200256294–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Stansfeld S, Marmot M, eds. Stress and the heart. Psychosocial pathways to coronary heart disease. London: BMJ Books, 2001
- 37.Csoboth C S, Birkas E, Buda B.et al Psychosocial correlates of alcohol use and cigarette smoking in Hungary. In: Weidner G, Kopp MS, Kristenson M, eds. Heart disease: environment, stress and gender. Amsterdam: IOS Press, 2002121–129.
- 38.Bobak M. New dietary risk factors: differences between Western and Eastern Europe. In: Weidner G, Kopp MS, Kristenson M, eds. Heart disease: environment, stress and gender. Amsterdam: IOS Press, 2002287–295.
- 39.Siegrist J, Starke D, Chandola T.et al The measurement of effort‐reward imbalance at work. European comparison. Soc Sci Med 200458(special issue)1483–1499. [DOI] [PubMed] [Google Scholar]
- 40.Head J, Stansfeld S A, Siegrist J. The psychosocial work environment and alcohol dependence: a prospective study. Occup Environ Med 200461219–224. [DOI] [PMC free article] [PubMed] [Google Scholar]