- Email: [email protected]
Self-perceived health in East and West Europe: another European health divide
PII:
Soc. Sci. Med. Vol. 46, No. 10, pp. 1355±1366, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain S0277-9536(97)10093-4 0277-9536/98 $19.00 + 0.00
SELF-PERCEIVED HEALTH IN EAST AND WEST EUROPE: ANOTHER EUROPEAN HEALTH DIVIDE PER CARLSON Department of Sociology, Stockholm University, S-106 91 Stockholm, Sweden AbstractÐThere is a great, and possibly also a growing, dierence in public health between the central/ eastern (CEE) and western European countries. Several suggestions have been put forward as explanations for this health divide. A broader framework than one focusing on medical care systems or behavioural patterns is necessary to examine this dierence. It will be more fruitful to try to identify social and economic factors at large, as well as speci®c explanatory factors. The aim of this study is to ®nd out to what extent ``The East±West Mortality Divide'' was apparent in people's perception of their own health in 1990±1991, as a division in self-perceived health across Europe. If there were indeed dierences, the aim is to examine whether or not they can be explained by speci®c economic and social conditions present in the early 1990s. Data from ``World Values Survey 1990'' reveal a striking east± west divide in self-perceived health among people in the age group 35±64 yr, one of greater size than the gender gap in self-perceived health. The importance of a number of circumstances for people's selfperceived health in the 25 European countries was estimated. The assumption was that any resulting dierence between eastern and western European countries could help to explain the health divide. An attempt was made to estimate how much the east-west health divide would be reduced if some of these circumstances were similar in CEE to those in the west. The results indicate that people's participation in civic activities has a positive eect on their health. This eect is recognised especially on a societal level. This supports theories about civic activities and community performance. In western Europe the tradition of the active citizen is more developed than in eastern Europe. People's life control was important for their self-perceived health in almost every European country, both in the west and the east. In the former communist countries, however, people did not feel that they had the same control over their lives as did people in the west. People's economic satisfaction was the most powerful predictor of self-perceived health, both in the eastern and western parts of Europe. The average level of economic satisfaction in 1990±1991 was considerably lower in CEE. If people's in¯uence and economic resources were the same in the former communist countries as in the west, the health divide, according to my estimations, would decrease by something between 10±30%. # 1998 Elsevier Science Ltd. All rights reserved Key wordsÐself-perceived health, East and West Europe, health divide
INTRODUCTION
The great (and growing?) dierence in health between populations in central/eastern (CEE) and those in western Europe has been recognised by several studies in recent years (see for instance VaÊgeroÈ and Illsley, 1992; Bobak and Marmot, 1996; Hertzman et al., 1996). It has been referred to as ``The European Health Divide'' (VaÊgeroÈ and Illsley, 1992) or ``The East±West Mortality Divide'' (Bobak and Marmot, 1996). Although this is a problem very much of today, a health divide was already present earlier in this century (League of Nations, Health Organization, 1932). A convergent trend dominated into the 1960s, after which a reverse trend has dominated. In other words, the last decades of communism were characterised by a lagging behind or even a deterioration of health. Since 1989 a ``mortality crisis'' has become visible in CEE, especially in Russia. The transition from communism to market economy and democracy seems to be paralleled by a
dramatic deterioration of public health, indicated by a falling life expectancy and a rising mortality, chie¯y among middle-aged men and women, but also among others. For instance, Russian male life expectancy decreased by 6.2 yr between 1990 and 1994, and for women by 3.4 yr (Leon et al., 1997). Up to now, the political response to this has been weak or non-existent in most of the former communist countries in Europe. Several suggestions have been put forward as explanations of the health divide. Often these explanations have focused on various behavioural patterns, such as the heavy smoking or drinking in CEE (Peto et al., 1992; Leon et al., 1997). Insucient resources in the health care system have also been in focus, but this is a far from fully satisfactory explanation (Bobak and Marmot, 1996; Wnuk-Lipinski and Illsley, 1990). Among social scientists, interest is more often concentrated on social stagnation and social disorganisation (Watson, 1995; Shapiro, 1995; VaÊgeroÈ and Illsley, 1992). The latest developments, post-communist,
1355
1356
P. Carlson
have not yet been studied, although such studies are now underway. The aim of this study was to ®nd out to what extent ``The European Health Divide'' or ``The East±West Mortality Divide'' is visible in people's perception of their own health in 1990±1991, as a division in self-perceived health across Europe. If such a division is indeed visible, I want to examine whether or not it can be explained by certain speci®c economic and social conditions which existed in the early 1990s. SOCIAL DISORGANISATION, POWER RESOURCES AND SOCIAL NETWORKS
I explore below the idea that the degree of integration in society could be of vital importance for its health record. A mutual relationship between the individual and the society to which he/she belongs and from which both bene®t, is contrasted with social disorganisation by Durkheim (Durkheim, 1992, p. 209). It is not only the lack of integration that is important. A weak correspondence between collective morality and social/economic structure has a negative eect on people's well-being (anomie) (Durkheim, 1992, p. 241). Both the last decades of the communist system and what followed its collapse can be seen in this light. In his book from 1993 ``Making Democracy Work'', Robert Putnam points to the importance of the civic community in developing successful institutions (Putnam, 1993). One of his main theses is that social and economic ``modernity'' per se is not enough for societal and democratic development. The active participation of citizens in their community is crucial for its functioning and for social cohesion and solidarity. Some social scientists have discussed the relation between civil society and the totalitarian power of the communist system in CEE (Hertzman and Marmot, 1996; Wilkinson, 1996). Totalitarian power undermines civil society. On the other hand, civil society also undermines totalitarian power, as was the case in Poland (Michnick, 1988). From Tocqueville we can learn about the importance of civil society for a working democracy. In ``Democracy in America'' (1840) Tocqueville writes: ``When some view is represented by an association, it must take clearer and more precise shape. It counts its supporters and involves them in its cause; these supporters get to know one another, and numbers increase zeal. An association unites the energies of divergent minds and vigorously directs them toward a clearly indicated goal''. (Tocqueville, 1969, p. 190.) The state's stability and functioning is also dependent on the civil society's perception of its legitimacy. The state cannot solely be coercive, but must provide the citizens with dierent kinds of services. In other words, there is a mutual exchange
between state and civil society (VõÂ ctor-DõÂ az, 1993, p. 59; Weber, 1958, p. 31). This (universalist) approach to ``civil society'' has been criticised by Hann (Hann and Dunn, 1996) among others, who argues for a more relativist approach. According to Hann, civil society has so far been a ``western'' innovation which is not necessarily applicable to societies in e.g. eastern Europe. The civil society is, to a certain extent, created by culture-speci®c and social patterns, unique to each country. It is not just about membership of voluntary associations. Furthermore, civil society diers today from how it was during the era of communism, when associations were more dependent on the state. However, as the Polish case in particular demonstrates, civil organisations and citizens' self-organisation was never entirely destroyed in central and eastern Europe during its communist period. Accordingly, I believe that this concept is useful, especially if put in a context of power resources, life control, social networks and social support. Also, it will be evident in the results presented here if it can bring new light to the situation in 1990±1991. A well developed civil society, encompassing for example a system of local clubs, associations, unions and churches with extensive membership networks, contributes to democracy, social cohesion and societal development. I believe it also will contribute to the improvement of public health. It is assumed that membership in associations was relevant for health in the past and will be relevant for health in the future. One reason for this is that formal and informal networks distribute resources, both material and immaterial, which are useful for avoiding or controlling health risks of various kinds, either by individuals or by local communities in collective action. Collective power resources were analysed by Korpi (Korpi, 1989). His ``power resources approach'' investigates the relationship between welfare state policies and market forces and in what way this relationship determines social policy. The fundamental assumption is that the power resources used in markets and in politics dier in class-related ways. In the market, capital and other economic resources form the fundament of power. These types of resources tend to be unequally distributed among classes and pressure groups in society. In the democratic political sphere, on the other hand, the fundamental power resource is the right to vote and the right to organise. Since these kinds of resources are linked to citizenship, they tend to be more equally distributed in society. Similarly, a strong civil society tends to mobilise these resources to a greater extent, for instance in local and national politics. Thus, the degree of political and trade union organisation has been shown to be signi®cantly related to a country's social policy. Wennemo
Self-perceived health in East and West Europe
demonstrated in her study of 18 OECD countries that the degree of such organisation is also closely related to infant mortality (Wennemo, 1994). Citizens' formal and informal organisation aects public health, directly through local social networks and indirectly through oering power resources which in¯uence a country's social, economic, environmental and other policies. Developments in large parts of CEE over the last few years, have brought market de-regulation, yet the political and democratic control of this de-regulation is fairly weak. This has resulted in a weakening of social integration, with new social divisions opening up, which may have had negative consequences (economic, social and for health) for large groups of the population. It is reasonable to assume that the present public health situation and its determinants (suggested in this article) can be seen in the light of (1) the failure of the former communist system (2) the problems related to its breaking up. There is reason to believe that many of the problems created during communism have been aggravated since 1989. Data presented here basically re¯ect a situation created before 1989, but are nevertheless, to a dierent degree in dierent countries, also in¯uenced by the transition crisis. VaÊgeroÈ and Illsley point to some characteristic features of the (earlier) development of welfare states in western Europe which distinguish them from the present transition in CEE, especially in the dependence of the former on parliamentary politics. `West European countries, operating within the democratic parliamentary tradition, have established an ideology and socio-political system described as ``welfare capitalism''. The development of the welfare state was to a large extent based on the aspirations and in¯uence of labour parties and/ or the trade unions through their participation in national parliamentary politics. This was true for both North-western and Southern Europe. Especially in Catholic Europe, it was in¯uenced by the traditional ideals of Christian charity as expressed by Christian Democratic parties and trade unions. Welfare development was paralleled by improvement in health status for all segments of the population, although not to the same extent for all'. (VaÊgeroÈ and Illsley, 1992, p. 225.) In contrast, post-war developments in CEE were dominated by a weakening of democratic in¯uence. Post-1989 developments represent a strengthening of democratic in¯uence, but not to the same extent as market forces have been strengthened. The balance of power resources, in Korpi's terms, is therefore very unstable in the present transition period. The ongoing transition in CEE can be described as a ``social realignment'', accompanied by increased social disorganisation and for individuals by an increased number of stressful life events. Whenever
1357
new formal and informal networks arise, they can be expected to be socially structured. A large body of epidemiological, medical and sociological research has shown the importance of people's social networks for their state of health. Research has mostly concentrated on explanations at an individual level, i.e. individuals with a well established social network are more protected in stressful life events than others (Berkman, 1986; Cassel, 1976). Sociological research is most concerned with the social context in which social networks are signi®cant for people's health (Brown and Harris, 1978). Cohen and Syme de®ne social support as the resources received from others. Does social support have a direct eect on health, or is the eect indirect (a buer)? They admit that there is evidence for both alternatives, but they consider the indirect eects to be more important (Cohen and Syme, 1985). We know that colleagues, relatives, friends and fellow members of associations or organisations constitute social support in this sense (Hall, 1990; OÈstergren, 1991). We can expect dierences between social and occupational groups in this respect (Berkman, 1986; Cohen and Syme, 1985; Shye et al., 1995). Thus, we expect transition societies of today to be characterised by increased social dierences in health, morbidity and mortality. The mechanisms behind these are not only psychologically mediated by stress, as suggested by Watson (Watson, 1995), but also related to the distribution of economic and political resources, both individually and collectively. It is not exclusively the number of social contacts, or whether or not one is a member of an organisation which is important; it is rather the whole social context in which the individual is embedded and through which he/she receive resources to cope with health problems or health risks. MEASURING HEALTH
One way to measure people's health is to ask them to estimate their own health according to a scale. This is what the ``World Values Survey 1990'' (WVS 1990) (see below) did. Aggregated data on self-perceived health can be used for country comparisons in the same way as mortality rates often are. Both methods have their own speci®c problems, but research based on self-perceived health and mortality nevertheless gives similar results. Self-perceived health predicts mortality in longitudinal follow-up studies of individuals (Kaplan et al., 1996; Welin et al., 1985). Similarly in a European country comparison (based on World Values Survey 1980), several individual attitudes, aggregated into mean values for individual countries, were found to predict national mortality rates from suicide and homicide (Carlson et al., 1994). As a prelude to the present study I analysed the correlation between
1358
P. Carlson
Fig. 1. The relationship between self-perceived health and IHD mortality in 23 European countries, men and women aged 35±64.
self-perceived health (as measured by WVS 1990 and aggregated into country means) and IHD mortality rates in 23 western and eastern European countries (excluding Denmark and Switzerland). The results, (males aged 35±64: r = ÿ 0.65, females aged 35±64: r = ÿ 0.64, p < 0.001), (Fig. 1), convinced us of the validity and meaningfulness of selfperceived health, as measured in this study. However, in country comparisons of self-perceived health it is likely that respondents' answers are to a certain extent culturally determined. Consequently, there would have to be fairly major dierences in self-perceived health between eastern and western Europe, for them to be entirely convincing.
POPULATIONS UNDER STUDY
World Values Survey (WVS 1990) is based on a large number of national surveys conducted 1990±1991 in 42 countries world-wide. All of these surveys were carried out by means of face to face interviews, with a sampling frame consisting of all adult citizens aged 18 and over. In most cases, strati®ed multi-stage random sampling was used. In most of the countries, and in all the European countries, the samples were full national samples. This study is based on data *Age 35±39: 0.200; age 40±44: 0.186; age 45±49: 0.144; age 50±54: 0.186; age 55±59; 0.147 and age 60±64: 0.149.
from 25 European countries and from respondents aged 35±64 yr. The countries are shown in Table 1. Sampling fractions vary from country to country and countries are of dierent size, which would give some countries a disproportionate impact in any analysis involving pooled samples. A weight factor was introduced, which gives greater weight to the more populous countries than to the less populous ones, so that the pooled analyses more closely approximate European reality. The weight factor was not used in the country-speci®c analyses. Age was controlled for in all the analyses (unless otherwise stated). This was carried out by adding the age variable in all regression models with individual data. When comparing countries by using aggregated data, dierences in age structure are eliminated by a weight corresponding to the proportions of each ®ve year group in the European ``normal population''*. WVS 1990 includes several hundred variables. My selection was carried out in two steps. Firstly I picked out some 30 variables which I thought were generally interesting in the light of my perspective on the east±west health divide as described above. In the next step I excluded those variables which were found not generally to be linked to self-perceived health. I was left with the following variables, which were included in the analyses presented opposite.
Self-perceived health in East and West Europe
1359
Table 1. The 25 investigated European countries with respondents in the ages 35±64 Austria (n = 752) Belarus (n = 529) Belgium (n = 1.363) Bulgaria (n = 529) Czechoslovakia (n = 784) Denmark (n = 490) Estonia (n = 533)
Finland (n = 328) France (n = 468) Great Britain (n = 700) Hungary (n = 567) Ireland (n = 499) Italy (n = 1.003) Latvia (n = 486)
Independent variables Country, age, sex, level of education (age at completion) and occupational group (professionals, routine non-manual workers, skilled manual workers and unskilled manual workers). Membership of non-political association (Yes/ No)*. Life control (1±10) (1 = none and 10 = a great deal). Job satisfaction (1±10) (1 = dissatis®ed and 10 = satis®ed). Freedom to make own decisions at work (1±10) (1 = none and 10 = a great deal). Satisfaction with economic situation of household (1±10) (1 = dissatis®ed and 10 = satis®ed). Political interest (1±4) (1 = very interested and 4 = not at all interested). Importance of family (1±4) (1 = very important and 4 = not at all important). Importance of friends (1±4) (1 = very important and 4 = not at all important). Dependent variable Self-perceived health (1±5) (1 = very good and 5 = very bad). In Lithuania there was no question about satisfaction with household economy. In Belarus, Czechoslovakia, Poland and Switzerland there was no question about association membership.
METHODS OF ANALYSIS
WVS 1990 oers the opportunity to describe selfperceived health and to analyse the relations between some social/economic factors and self-perceived health in eastern and western Europe. First I compare the 25 countries by their mean values of self-perceived health. Secondly, with multiple regressions (OLS), I try to estimate the importance of some variables seen as determinants of individually measured self-perceived health in the 25 European countries. To assess the relative importance of the independent variables I then converted them to a *Membership of social welfare services for elderly, handicapped or deprived; education, arts, music or cultural activities; ecology groups; peace movement; sports or recreation; women's groups; other non-political associations.
Lithuania (n = 479) Netherlands (n = 534) Norway (n = 638) Poland (n = 530) Portugal (n = 546) Romania (n = 573) Russia (n = 1.057)
Spain (n = 1.953) Sweden (n = 533) Switzerland (n = 687) West Germany (n = 996) West (n = 11.490) CEE (n = 6.067)
common scale, that is, to a standardised measure. By multiplying the unstandardised b-values of the OLS regressions by the ratio of the standard deviations of the independent and the dependent variables (i.e. bxy(sx/sy)) a b-value is obtained. This slope coecient shows how many standard deviations the dependent variable will change as the result of a one standard deviation change in the independent variable, controlling for all the other independent variables in the equation (Walsh, 1990, p. 276). I assumed that any general east±west dierence in self-perceived health could result from two sources. Firstly, if variables show dierent eects (i.e. fairly dierent b-values) in eastern and western countries this would be a probable explanation. Secondly, if the average levels of an equally important determinant dier between east and west this would also help explain the dierence in self-perceived health. The models were constructed with one independent variable at a time (10 independent variables altogether) and always with age as a control. Selfperceived health is the dependent variable in all the models and the regressions were performed separately for each sex and for each country. This made a total of 500 regressions (10 2 25). I had theoretical reasons for analysing ``association membership'' on an aggregate level, since, in line with my reasoning above, countries with a high degree of association membership are expected to exercise collective control over health risks. Thus, I carried out two new regressions, one for each sex, with country as the unit (21 observations), and with the aggregated variables ``association membership'' and ``self-perceived health''. Here I will above all discuss those results that in some way can throw light on the east±west dimension in health. Lastly, I will apply the estimated net eects of the most important health determinants in CEE (analysed as a region) to the average west European levels of these determinants. This procedure enables us to predict a new value of self-perceived health, achievable if CEE were raised to the level of western Europe with regard to these particular determinants. Because of potential problems with data reliability and validity and since there is a cross-sectional study design, all results and statements about causality are tentative.
1360
P. Carlson
Fig. 2. Self-perceived poor health in 25 European countries 1990±1991, men and women aged 35±64, (5 = very bad, 1 = very good) (ranked by each country's total mean). RESULTS
Self-perceived health The average level of self-perceived health is generally worse in former communist countries than in western Europe, for both men and women (Fig. 2). The countries with the worst levels are Belarus, Hungary and Russia. The best average level of selfperceived health is to be found in Denmark, Ireland, Sweden and Switzerland. The male east± west gap in the average level of self-perceived health is 0.54 (mwest=2.19 and mCEE=2.73), in other words the average level among CEE men is about 25% poorer than among men in the west. The female gap is 0.66 (mwest=2.28 and mCEE=2.94), which means that CEE women are on a level about 29% poorer than that for women in the west. The east±west divide in self-perceived health is even wider than the sex dierence for the same variable. In Fig. 2 I have no correction for national dierences in age structure. However, the average age and the age distribution in the age span 35±64 are similar in the east and in the west, both for men and women. Membership of an organisation As previously discussed, membership of an organisation of some kind may in¯uence people's health positively in at least two ways. First, on an individual level, where membership is related to social networks and social support which, directly or indirectly, have positive eects on the individual's self-perceived health. Second, on an aggregate level, where people's civic activities positively in¯uence the achievement of the community in terms of public health.
On the individual level, people's membership of non-political organisations showed some weak eect in most of the countries investigated. This means that people who were members of at least one association tend to perceive their health as better. When the data is aggregated the association becomes stronger. At national level, the degree of membership in non-political organisations is fairly closely related to people's self-perceived health (Fig. 3). Countries with a higher degree of membership tend to have better self-perceived (public) health (men r = 0.67, women r = 0.73, p < 0.001). Western European countries, especially the Nordic countries and the Netherlands, have a substantially higher degree of organisation membership. Life control The next step was to estimate the eects of life control on self-perceived health (Table 2). The way in which people perceive their scope for exercising control over their own life showed a similar signi®cant eect on self-perceived health in almost every country, i.e. people with less ``life control'' tend to perceive their health as worse. Dierences between east and west seem to be small. The variation within each region is more striking than that between regions. These results cannot, therefore, explain the health divide. Next, I investigated the extent to which people in the dierent countries can control their own lives. When one compares the means of the ``life control'' variable some interesting dierences are immediately observable (Fig. 4). The west European countries generally report a higher level of life control than the former communist countries. The average level of life control in
Self-perceived health in East and West Europe
1361
Fig. 3. The relationship between the degree of membership in civil (non-political) associations and the proportion reporting good or very good health in 20 European countries 1990±1991, men and women aged 35±64.
the western European countries is m = 6.7 and in the CEE countries m = 6.1. This dierence in means is signi®cant (p < 0.001). In the western parts of Europe, France and the Netherlands had Table 2. Estimated eects of life control on self-perceived health in 25 European countries and in two regions (respondents in the ages 35±64); b-values (controlled for age)
Austria Belgium Denmark Finland France Great Britain Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland W. Germany West Belarus Bulgaria Czechoslovakia Estonia Hungary Latvia Lithuania Poland Romania Russia CEE *p < 0.05. $p < 0.01. %p < 0.001.
Men
Women
Western Europe 0.09 ÿ0.08* ÿ0.29% ÿ0.08 ÿ0.20$ ÿ0.22% ÿ0.16$ ÿ0.17% ÿ0.05 ÿ0.07 ÿ0.08 ÿ0.15% ÿ0.19$ ÿ0.10% ÿ0.27%
ÿ0.03 ÿ0.15% ÿ0.25% ÿ0.16 ÿ0.18$ ÿ0.25% ÿ0.19$ ÿ0.17% ÿ0.08 ÿ0.20% ÿ0.10 ÿ0.22% ÿ0.15* ÿ0.07 ÿ0.23%
ÿ0.15%
ÿ0.18%
Central and Eastern Europe ÿ0.05 ÿ0.24% ÿ0.14$ ÿ0.05 ÿ0.07 ÿ0.19$ ÿ0.11 0.10 ÿ0.08 ÿ0.18%
ÿ0.18$ ÿ0.20$ ÿ0.17% ÿ0.23% ÿ0.05 0.00 ÿ0.17$ ÿ0.19$ ÿ0.14* ÿ0.13$
ÿ0.12%
ÿ0.11%
an average level below what might be expected (m1 6.0), and in the eastern parts, the Baltic countries scored better than their eastern neighbours (m1 6.4). The worst situation was reported in Bulgaria and Belarus, with average levels of 5.1 and 5.8 respectively. It appears that even if the dierences between the two regions are not enormous they are nevertheless clear and signi®cant and can help to explain some of the existing health divide. In general, one can say that the dierences in degree of life control between the sexes are small, certainly smaller than national dierences. Household economy People's satisfaction with their household economy had a strong relation to their self-perceived health. This factor could contribute to east±west dierences if either the eects (i.e. the beta coecients) or the average levels of economic satisfaction dier. Economic satisfaction is a very important determinant in almost all countries and for both sexes (Table 3). From the results one can see that greater economic satisfaction is related everywhere to better self-perceived health, except, it seems, in one country (Austria). Figure 5 shows that none of the former communist countries shows a higher degree of economic satisfaction than any of the west European countries! This is true for both men and women. In fact there is an east±west dierence of 1.8, where the mean score of the western European countries is 6.7 and that of the CEE countries is only 4.9 (p < 0.001). Most satis®ed with their household ®nances are the Dutch and the Swiss. Least satis®ed
1362
P. Carlson
Fig. 4. People's perceived control over their lives in 25 European countries 1990±1991, men and women aged 35±64 (10 = a great deal, 1 = none) (ranked by each country's total mean).
are the Bulgarians and the Latvians. This great east±west gap is very interesting since economic satisfaction is a strong predictor of self-perceived health all over Europe. The health divide can Table 3. Estimated eects of economic satisfaction on self-perceived health in 24 European countries and in two regions (respondents in the ages 35±64); b-values (controlled for age)
Austria Belgium Denmark Finland France Great Britain Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland W. Germany West Belarus Bulgaria Czechoslovakia Estonia Hungary Latvia Lithuania Poland Romania Russia CEE *p < 0.05. $p < 0.01. %p < 0.001.
Men
Women
Western Europe 0.10 ÿ0.28% ÿ0.16* ÿ0.16* ÿ0.38% ÿ0.23% ÿ0.14* ÿ0.20% ÿ0.14* ÿ0.16$ ÿ0.17$ ÿ0.21% ÿ0.26% ÿ0.27% ÿ0.33$
0.04 ÿ0.24% ÿ0.22% ÿ0.31% ÿ0.25% ÿ0.26% ÿ0.34% ÿ0.14% ÿ0.12* ÿ0.10 ÿ0.20% ÿ0.25% ÿ0.16* ÿ0.17% ÿ0.25%
ÿ0.23%
ÿ0.20%
Central and Eastern Europe ÿ0.21% ÿ0.23% ÿ0.25% ÿ0.19* ÿ0.12* ÿ0.25% ÿ ÿ0.13* ÿ0.17$ ÿ0.20%
ÿ0.09 ÿ0.24% ÿ0.26% ÿ0.15* ÿ0.11 ÿ0.24% ÿ ÿ0.25% ÿ0.27% ÿ0.16%
ÿ0.19%
ÿ0.17%
accordingly to a certain degree be explained by dierences in economic satisfaction between eastern and western Europe. People's perception of their life control and their degree of economic satisfaction were the two variables that, on the micro level, most clearly demonstrated similar and signi®cant eects in most of the countries, in eastern as well as in western Europe. I will now move on brie¯y to discuss the results of some of the other tested variables. Social interaction Some of the tested variables concerned people's social lives. How people prioritised issues such as family and friends had some eect on self-perceived health, i.e. people who considered family and friends as important tended to have a better selfperceived health. These eects were only visible in a small number of western European countries and this variable cannot be seen as important in explaining the east±west health divide. Class and health Class is traditionally regarded as a good health predictor also in eastern Europe (VaÊgeroÈ and Illsley, 1992). However, the results from this study are not as clear as one might expect. There are dierences in health between dierent educational and occupational groups, but these dierences are here mostly valid for the western European countries. People with a lower educational level and manual workers tend to be more unhealthy than others. In the former communist countries neither occupational group nor educational level demonstrated any signi®cant eect on self-perceived health
Self-perceived health in East and West Europe
1363
Fig. 5. People's satisfaction with their household economy in 24 European countries 1990±1991, men and women aged 35±64 (10 = satis®ed, 1 = dissatis®ed) (ranked by each country's total mean).
(Hungary was one exception). The validity of the information on social class/occupation in the WVS 1990 is not known. There is clearly reason to believe that there are educational and class dierences in health in CEE, but the class pattern may also look somewhat dierent in these countries. Class may have to be conceived or measured dierently from how it was in the WVS 1990 in postcommunist countries. Class, as measured in this survey, could not help us to explain the health divide since it did not demonstrate any signi®cant eect on self-perceived health in the CEE countries.
Work and health Two work-related variables were analysed in this study: job satisfaction and decision-making latitude in one's work and their importance for health. There were no obvious east±west dierences in the eects of these variables. Job satisfaction showed some eect in almost every country, even if the beta coecients were often not signi®cant. People who were more satis®ed with their situation at work tended to have better health and the average level of satisfaction at work was lower in the former communist countries than elsewhere. The freedom to take decisions at work had eects on self-perceived health in several countries, most commonly in western Europe and among men. There was a slight tendency towards better health among people with greater decision-making latitude at work and the average level of this latitude was lower in the former communist countries. Satisfaction at work
and the freedom to make decisions may thus to some extent explain the health divide. Political interest People's political interests had some slight eects on self-perceived health in some, mostly western, countries. People interested in politics tend to perceive their health as better. This variable seems to be more important for women than for men in determining self-perceived health. Political interest did not have any signi®cant in¯uence on the health divide. The importance of membership of organisations, life control and economic satisfaction for the European health divide As I have proposed above, membership of an organisation, life control and economic satisfaction seemed to be important variables determining selfperceived health in the countries studied here. As a thought experiment, I have tried to estimate what the average level of self-perceived health level in the former communist countries would be, if the citizens in those countries were as often organised in civil associations, enjoyed as much life control, and were as satis®ed economically as people in western European countries. From a regression model (applied to the observations from the CEE countries) with the variables age, organisation membership, life control and economic satisfaction we obtain the importance of these variables in this speci®c region. The next step is to put in the average western levels of the three
1364
P. Carlson
last variables. We will now be able to estimate what the average self-perceived health level would be in CEE if these conditions were the same as in the west*. For men, the east±west dierence in self-perceived health averages is 0.54 (2.73CEEÿ2.19west). With our regression equation the estimated new east average for CEE men is 2.67 and the health gap is reduced by 11%. For women, the east±west dierence in self-perceived health averages is 0.66 (2.94±2.28). With our regression equation the new east average for women becomes 2.73, i.e. the health gap is reduced by 32%. Finally, I also estimated some east±west interaction eects from the two most powerful variables at individual level (life control and economic satisfaction separately). Both life control and economic satisfaction were signi®cantly stronger predictors of self-perceived health in western Europe (men and women). However, the interactions were not very strong. The implication of these interactions is that if eastern Europe became more similar to western Europe, increased ``life control'' and ``economic satisfaction'' would result in greater gains in self-perceived health than those estimated above. Finally, one note of caution: we cannot be sure of the direction of the established relations. It might well be true that some people, because of their ill health, perceive that they have less life control, a poorer household economy etc. We must bear this in mind when we interpret the results. DISCUSSION
In this study it is clearly demonstrated that the so-called European health divide, documented for mortality, is also noticeable in self-perceived health. Explaining this divide is hardly a simple task. The variety of social, economic and cultural structures make the picture very complex. The former communist countries, like the western European countries, are a far from homogeneous group. They will also almost certainly develop at dierent paces, and perhaps in dierent directions (Deacon, 1992). Economic performance, ``civil society'' and ``life control'' are all ``system linked''. We are looking *This will be estimated with the equation: Y = a + (age X1) + (ecsat X2) + (lifecont X3) + (civass X4), where Y = predicted average level of self-perceived health in CEE; a = intercept; age = eect of age in CEE; civass = eect of membership of a non-political association in CEE; ecsat = effect of economic satisfaction in CEE; lifecont = eect of life control in CEE; X1=average age in CEE; X2=average economic satisfaction in the west; X3=average life control in the west and X4=average degree of association membership in the west. Men Y = 2.36 + (0.02age48.21X1) ÿ (0.06ecsat6.65X2) ÿ (0.03lifecont6.70X3) ÿ (0.12civass0.47X4) and women Y = 2.33 + (0.02age48.06X1) ÿ (0.06ecsat6.58X2) ÿ (0.02lifecont6.63X3) ÿ (0.08civass0.40X4.
for explanations which allow us to link the macro properties of the two European socio-political systems to individual health risks and health perception. My results indicate that people's participation in civic activities has a positive eect on their health. This eect is strongest on a societal level, which is in line with Putnam's theory about civil activities and community performance. Concepts which, by de®nition, are ecological. In western Europe the tradition of the active citizen is presently stronger than in eastern Europe, a fact which is demonstrated by this study also. The importance of people's life control for their self-perceived health is shown to be very similar among the European countries, both west and east. In 1990±1991, people in the former communist countries did not feel that they had the same control over their lives as did people in the west. The lack of in¯uence among eastern European citizens is chie¯y a problem of democracy, but also a manifestation of a less developed civil society. The concept of civil society is however not totally clear cut. The meaning of it might very well dier from country to country. Here the operational de®nition was restricted to voluntary, non-political and non-religious associations. This de®nition most certainly has limitations. The reason for excluding political organisations and trade unions in the de®nition was the close proximity between them and the state, indeed subordination to the state, in many of the central and eastern European countries. Excluding membership in religious organisations and churches is perhaps more debatable. In Poland, for instance, the church had a central place in civil society during the last years of communism (Buchowski, 1996). The importance of religious organisations and churches and their connection to the state in the countries investigated dier considerably. It was dicult to form an opinion of the extent to which they were important in each country, and therefore church membership was not included in the de®nition of civil society in the study. What was left can thus be seen as the lowest common denominator of civil society. In spite of this limitation, the results discussed above have shown that the operationalisation is useful. This study also demonstrates the close relationship between people's satisfaction with their household economy and their self-perceived health. An unsatisfactory economic situation is most clearly related to a worse self-perceived health in 24 of the 25 investigated countries (Austria was an exception). In 1990±1991, the average level of economic satisfaction was considerably lower in CEE than in the west. This factor thus contributes importantly to the European health divide. The economic growth of the former communist countries is important for the improvement of their public health. Some of these countries had better economic preconditions than others in the early
Self-perceived health in East and West Europe
1990s (Czechoslovakia, Hungary and Poland). It is important also to bear in mind that the distribution of these resources is very important. Since 1990 the disparity in income distribution has grown dramatically in many of these countries and large sections of the populations live under the poverty line (United Nations Children's Fund, 1995). Both The World Bank and the Human Development Report, following Wilkinson (1986), have argued that a more equal income distribution and a more equal distribution of the possession of natural resources create not only a better climate for economic growth (The Independent, 21 July 1996), but also improved health and survival (World Bank, 1993). The results presented in this article re¯ect a situation basically created during the communist era, but many of the problems have been aggravated since then. Political obstacles for citizens in central and eastern Europe have been replaced with economic ones. In the context of the individual's own resources, the result may be very much the same. Those privileged today are often individuals who were privileged yesterday (see for instance Stark, 1996; Hanley et al., 1995). Thus, even if social strati®cation today occasionally has a dierent basis from earlier, where privileged individuals are concerned there is an element of continuity. Finally, some methodological considerations. With this kind of data there are always certain causality problems. For instance, does poor selfperceived health imply perceptions of less control or poorer economy? Negative aectivity, in other words the tendency to rate a number of life circumstances in a negative light, on psychological or personality grounds, is also possible (Watson and Pennbaker, 1989). Those possibilities certainly exist and must be born in mind when interpreting these results. However, during recent decades a large literature indicating similar causal relations as those suggested here has emerged (see for instance Marmot et al., 1997; OÈstergren, 1991; Karasek, 1979). Cultural dierences may also in¯uence perception of health, and there is a possibility of a bias if these perceptions were systematically dierent between east and west Europe. Two things make this less likely as an explanation. First, the level of self-perceived health was systematically linked to the heart disease rate in a country. Secondly, the dierences between east and west were quite substantial, in fact larger than between men and women. Further, it can be argued that the measures used also are somewhat distant from the theoretical concepts suggested in the introduction above. This is true to an extent but still I would argue, that some light has been shed on the question of why there is a European health divide. In sum, in spite of its limitations, the present study indicates that an improvement of public health in the CEE countries might be facilitated by
1365
improvements in household economies, by individuals achieving more control over their everyday lives and by a strengthening of civil society, seen both as a individual resource in situations of hardship and stress, and as a source of power to balance market forces. I suggest that explanations that focus exclusively on individual behaviours, such as smoking and drinking or on supply of medical services, are either one-sided or too limited. AcknowledgementsÐThe author gratefully acknowledges the support from a grant from the Research Council for the Humanities and Social Sciences (grant no F 0915/96) and Professor Denny VaÊgeroÈ at the Department of Sociology, Stockholm University for his generous advice and assistance. An earlier version of this article was presented at the European Society of Medical Sociology, 6th Biennial Conference, August 29±31, 1996, in Budapest, Hungary and at the Nordic Workshop on Health Inequalities, November 21±23, 1996, in Stockholm, Sweden. REFERENCES
Berkman, L. F. (1986) Social networks, support, and health: taking the next step forward. American Journal of Epidemiology 123, 559±562. Bobak, M. and Marmot, M. (1996) East±west mortality divide and its potential explanations: proposed research agenda. British Medical Journal 312, 421±425. Brown, G. and Harris, T. (1978) The Social Origins of Depression. Tavistock, London. Buchowski, M. (1996) The shifting meanings of civil society in Poland. In Civil Society. Challenging Western Models, ed. C. Hann and D. Dunn, pp. 79±98. Routledge, London. Carlson, P., MaÈkinen, I. and VaÊgeroÈ, D. (1994) SjaÈlvmord, mord och kultur. En jaÈmfoÈrelse av tio laÈnder i Europa. [Suicide, homicide and cultural attitudes. An investigation of ten European countries.] Sociologisk Forskning 31, 78±89. Cassel, J. (1976) The contribution of the social environment to host resistance. American Journal of Epidemiology 104, 107±123. Cohen, S. and Syme, L. S. (1985) Social Support and Health. Academic Press, Orlando. Deacon, B. (1992) The New Eastern Europe. Social Policy Past, Present and Future. SAGE, London. Durkheim, E. (1992) Suicide. A Study in Sociology. Routledge, London. Hall, E. (1990) Women's Work: An Inquiry into the Health Eects of Invisible and Visible Labor. Dissertation, Karolinska Institute, Solna. Hanley, E., Yershova, N. and Anderson, R. (1995) Russia ÐOld wine in a new bottle? The circulation and reproduction of Russian elites, 1983±1993 Theory and Society 24, 639±668. Hann, C. and Dunn, D. ed. (1996) Civil Society. Challenging Western Models. Routledge, London. Hertzman, C., Kelly, S. and Bobak, M. ed. (1996) East± West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants. Kluwer, Dordrecht. Hertzman, C. and Marmot, M. (1996) The leading hypothesis and its discontents: a synthesis of evidence and outstanding issues regarding the east±west life expectancy gap. In East±West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants, ed. C. Hertzman, S. Kelly and M. Bobak, pp. 211±219. Kluwer, Dordrecht. The Independent, 21 July 1996.
1366
P. Carlson
Kaplan, G., Goldberg, D., Everson, S., Cohen, R. D., Salonen, R., Tuomilehto, J. and Salonen, J. (1996) Perceived health status and morbidity and mortality: evidence from the Kuopio ischaemic heart disease risk factor study. International Journal of Epidemiology 25, 259±265. Karasek, R. A. (1979) Job demands, job decision latitude, and mental strain: implications for job redesign. Administrative Science Quarterly 24, 285±307. Korpi, W. (1989) Power, politics, and state autonomy in the development of social citizenship: social rights during sickness in eighteen OECD countries since 1930. American Sociological Review 54, 309±328. League of Nations, Health Organization (1932) Statistics of Noticeable Diseases for the Year 1930. Health Section, Geneve. Leon, D. A., Chenet, L., Shkolnikov, V. M., Zakharov, S., Shapiro, J., Rakhmanova, G., Vassin, S. and McKee, M. (1997) Huge variation in Russian mortality rates 1984±1994: artefact, alcohol, or what? Lancet 350, 383±388. Marmot, M. G., Bosma, H., Hemingway, H., Brunner, E. and Stansfeld, S. (1997) Contribution of job control and other risk factors to social variations in coranary heart incidence. Lancet 350, 235±239. Michnick, A. (1988) En lektion i vaÈrdighet. EssaÈer fraÊn det polska motstaÊndet 1973±1988. [A lecture in dignity. Essays from the Polish resistance 1973±1988.] Ordfronts foÈrlag, Stockholm. OÈstergren, P.-O. (1991) Psychosocial Resources and Health. With special reference to social network, social support and cardiovascular disease. Dissertation, Studentlitteratur, MalmoÈ. Peto, R., Lopez, A., Boreham, J., Thun, M. and Heath, C. (1992) Mortality from tobacco in developed countries: indirect estimation from national vital statistics. Lancet 339, 1268±1278. Putnam, R. (1993) Making Democracy Work. Civic Traditions in Modern Italy. Princeton University Press, Princeton. Shapiro, J. (1995) The Russian mortality crisis and its causes. In Russian Economic Reform at Risk, ed. A. AÊslund. Pinter, London. Shye, D., Mullooly, J. P., Freeborn, D. K. and Pope, C. R. (1995) Gender dierences in the relationship between social network support and mortality: a longitudinal study of an elderly cohort. Social Science and Medicine 41, 935±947.
Stark, D. (1996) Recombinant property in east European capitalism. American Journal of Sociology 101, 993± 1027. Tocqueville, A. (1969) Democracy in America. Anchor Books, Garden City, New York. (Cited by Putnam, 1993 p. 90.). United Nations Children's Fund (1995) Central and eastern Europe in transition. Public policy and social conditions. Poverty, children and policy: responses for a brighter future. Economies in Transition Studies. Regional Monitoring Report 3. VõÂ ctor-DõÂ az, V. M. (1993) The Return of Civil Society. The Emergence of Democratic Spain. Harvard University Press, London. VaÊgeroÈ, D. and Illsley, R. (1992) Inequality, health and policy in east and west Europe. International Journal of health Sciences 3, 225±239. Walsh, A. (1990) Statistics for the Social Sciences. With Computer Applications. Harper and Row, New York. Watson, D. and Pennbaker, J. (1989) Health complaints, stress, and distress: exploring the central role of negative aectivity. Psychological Review 96, 234. Watson, P. (1995) Explaining rising mortality among men in eastern Europe. Social Science and Medicine 41, 923± 934. Weber, M. (1958). Economy and Society, Vol. 2. University of California Press, Berkeley. Welin, L., Tibblin, G., Svardsudd, K., Tibblin, B., AnderReciva, S., Larsson, B. and Wilhelmsen, L. (1985) Prospective study of social in¯uences on mortality. The study of men born in 1913 and 1923. Lancet 20, 915± 918. Wennemo, I. (1994) Sharing the Costs of Children. Studies on the Development of Family Support in the OECD Countries. Dissertation, Institute for social research, Stockholm. Wilkinson, R. (1986) Income and mortality. In Class and Health, ed. R. Wilkinson, pp. 88±114. Tavistock, London. Wilkinson, R. (1996) Health and civic society in eastern Europe before 1989. In East±West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants, ed. C. Hertzman, S. Kelly and M. Bobak, pp. 195±209. Kluwer, Dordrecht. Wnuk-Lipinski, E. and Illsley, R. (1990) Non-market economies in health. Social Science and Medicine 31, 833±836. World Bank (1993) World Development Report. Investing in Health. Oxford University Press, New York.
Soc. Sci. Med. Vol. 46, No. 10, pp. 1355±1366, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain S0277-9536(97)10093-4 0277-9536/98 $19.00 + 0.00
SELF-PERCEIVED HEALTH IN EAST AND WEST EUROPE: ANOTHER EUROPEAN HEALTH DIVIDE PER CARLSON Department of Sociology, Stockholm University, S-106 91 Stockholm, Sweden AbstractÐThere is a great, and possibly also a growing, dierence in public health between the central/ eastern (CEE) and western European countries. Several suggestions have been put forward as explanations for this health divide. A broader framework than one focusing on medical care systems or behavioural patterns is necessary to examine this dierence. It will be more fruitful to try to identify social and economic factors at large, as well as speci®c explanatory factors. The aim of this study is to ®nd out to what extent ``The East±West Mortality Divide'' was apparent in people's perception of their own health in 1990±1991, as a division in self-perceived health across Europe. If there were indeed dierences, the aim is to examine whether or not they can be explained by speci®c economic and social conditions present in the early 1990s. Data from ``World Values Survey 1990'' reveal a striking east± west divide in self-perceived health among people in the age group 35±64 yr, one of greater size than the gender gap in self-perceived health. The importance of a number of circumstances for people's selfperceived health in the 25 European countries was estimated. The assumption was that any resulting dierence between eastern and western European countries could help to explain the health divide. An attempt was made to estimate how much the east-west health divide would be reduced if some of these circumstances were similar in CEE to those in the west. The results indicate that people's participation in civic activities has a positive eect on their health. This eect is recognised especially on a societal level. This supports theories about civic activities and community performance. In western Europe the tradition of the active citizen is more developed than in eastern Europe. People's life control was important for their self-perceived health in almost every European country, both in the west and the east. In the former communist countries, however, people did not feel that they had the same control over their lives as did people in the west. People's economic satisfaction was the most powerful predictor of self-perceived health, both in the eastern and western parts of Europe. The average level of economic satisfaction in 1990±1991 was considerably lower in CEE. If people's in¯uence and economic resources were the same in the former communist countries as in the west, the health divide, according to my estimations, would decrease by something between 10±30%. # 1998 Elsevier Science Ltd. All rights reserved Key wordsÐself-perceived health, East and West Europe, health divide
INTRODUCTION
The great (and growing?) dierence in health between populations in central/eastern (CEE) and those in western Europe has been recognised by several studies in recent years (see for instance VaÊgeroÈ and Illsley, 1992; Bobak and Marmot, 1996; Hertzman et al., 1996). It has been referred to as ``The European Health Divide'' (VaÊgeroÈ and Illsley, 1992) or ``The East±West Mortality Divide'' (Bobak and Marmot, 1996). Although this is a problem very much of today, a health divide was already present earlier in this century (League of Nations, Health Organization, 1932). A convergent trend dominated into the 1960s, after which a reverse trend has dominated. In other words, the last decades of communism were characterised by a lagging behind or even a deterioration of health. Since 1989 a ``mortality crisis'' has become visible in CEE, especially in Russia. The transition from communism to market economy and democracy seems to be paralleled by a
dramatic deterioration of public health, indicated by a falling life expectancy and a rising mortality, chie¯y among middle-aged men and women, but also among others. For instance, Russian male life expectancy decreased by 6.2 yr between 1990 and 1994, and for women by 3.4 yr (Leon et al., 1997). Up to now, the political response to this has been weak or non-existent in most of the former communist countries in Europe. Several suggestions have been put forward as explanations of the health divide. Often these explanations have focused on various behavioural patterns, such as the heavy smoking or drinking in CEE (Peto et al., 1992; Leon et al., 1997). Insucient resources in the health care system have also been in focus, but this is a far from fully satisfactory explanation (Bobak and Marmot, 1996; Wnuk-Lipinski and Illsley, 1990). Among social scientists, interest is more often concentrated on social stagnation and social disorganisation (Watson, 1995; Shapiro, 1995; VaÊgeroÈ and Illsley, 1992). The latest developments, post-communist,
1355
1356
P. Carlson
have not yet been studied, although such studies are now underway. The aim of this study was to ®nd out to what extent ``The European Health Divide'' or ``The East±West Mortality Divide'' is visible in people's perception of their own health in 1990±1991, as a division in self-perceived health across Europe. If such a division is indeed visible, I want to examine whether or not it can be explained by certain speci®c economic and social conditions which existed in the early 1990s. SOCIAL DISORGANISATION, POWER RESOURCES AND SOCIAL NETWORKS
I explore below the idea that the degree of integration in society could be of vital importance for its health record. A mutual relationship between the individual and the society to which he/she belongs and from which both bene®t, is contrasted with social disorganisation by Durkheim (Durkheim, 1992, p. 209). It is not only the lack of integration that is important. A weak correspondence between collective morality and social/economic structure has a negative eect on people's well-being (anomie) (Durkheim, 1992, p. 241). Both the last decades of the communist system and what followed its collapse can be seen in this light. In his book from 1993 ``Making Democracy Work'', Robert Putnam points to the importance of the civic community in developing successful institutions (Putnam, 1993). One of his main theses is that social and economic ``modernity'' per se is not enough for societal and democratic development. The active participation of citizens in their community is crucial for its functioning and for social cohesion and solidarity. Some social scientists have discussed the relation between civil society and the totalitarian power of the communist system in CEE (Hertzman and Marmot, 1996; Wilkinson, 1996). Totalitarian power undermines civil society. On the other hand, civil society also undermines totalitarian power, as was the case in Poland (Michnick, 1988). From Tocqueville we can learn about the importance of civil society for a working democracy. In ``Democracy in America'' (1840) Tocqueville writes: ``When some view is represented by an association, it must take clearer and more precise shape. It counts its supporters and involves them in its cause; these supporters get to know one another, and numbers increase zeal. An association unites the energies of divergent minds and vigorously directs them toward a clearly indicated goal''. (Tocqueville, 1969, p. 190.) The state's stability and functioning is also dependent on the civil society's perception of its legitimacy. The state cannot solely be coercive, but must provide the citizens with dierent kinds of services. In other words, there is a mutual exchange
between state and civil society (VõÂ ctor-DõÂ az, 1993, p. 59; Weber, 1958, p. 31). This (universalist) approach to ``civil society'' has been criticised by Hann (Hann and Dunn, 1996) among others, who argues for a more relativist approach. According to Hann, civil society has so far been a ``western'' innovation which is not necessarily applicable to societies in e.g. eastern Europe. The civil society is, to a certain extent, created by culture-speci®c and social patterns, unique to each country. It is not just about membership of voluntary associations. Furthermore, civil society diers today from how it was during the era of communism, when associations were more dependent on the state. However, as the Polish case in particular demonstrates, civil organisations and citizens' self-organisation was never entirely destroyed in central and eastern Europe during its communist period. Accordingly, I believe that this concept is useful, especially if put in a context of power resources, life control, social networks and social support. Also, it will be evident in the results presented here if it can bring new light to the situation in 1990±1991. A well developed civil society, encompassing for example a system of local clubs, associations, unions and churches with extensive membership networks, contributes to democracy, social cohesion and societal development. I believe it also will contribute to the improvement of public health. It is assumed that membership in associations was relevant for health in the past and will be relevant for health in the future. One reason for this is that formal and informal networks distribute resources, both material and immaterial, which are useful for avoiding or controlling health risks of various kinds, either by individuals or by local communities in collective action. Collective power resources were analysed by Korpi (Korpi, 1989). His ``power resources approach'' investigates the relationship between welfare state policies and market forces and in what way this relationship determines social policy. The fundamental assumption is that the power resources used in markets and in politics dier in class-related ways. In the market, capital and other economic resources form the fundament of power. These types of resources tend to be unequally distributed among classes and pressure groups in society. In the democratic political sphere, on the other hand, the fundamental power resource is the right to vote and the right to organise. Since these kinds of resources are linked to citizenship, they tend to be more equally distributed in society. Similarly, a strong civil society tends to mobilise these resources to a greater extent, for instance in local and national politics. Thus, the degree of political and trade union organisation has been shown to be signi®cantly related to a country's social policy. Wennemo
Self-perceived health in East and West Europe
demonstrated in her study of 18 OECD countries that the degree of such organisation is also closely related to infant mortality (Wennemo, 1994). Citizens' formal and informal organisation aects public health, directly through local social networks and indirectly through oering power resources which in¯uence a country's social, economic, environmental and other policies. Developments in large parts of CEE over the last few years, have brought market de-regulation, yet the political and democratic control of this de-regulation is fairly weak. This has resulted in a weakening of social integration, with new social divisions opening up, which may have had negative consequences (economic, social and for health) for large groups of the population. It is reasonable to assume that the present public health situation and its determinants (suggested in this article) can be seen in the light of (1) the failure of the former communist system (2) the problems related to its breaking up. There is reason to believe that many of the problems created during communism have been aggravated since 1989. Data presented here basically re¯ect a situation created before 1989, but are nevertheless, to a dierent degree in dierent countries, also in¯uenced by the transition crisis. VaÊgeroÈ and Illsley point to some characteristic features of the (earlier) development of welfare states in western Europe which distinguish them from the present transition in CEE, especially in the dependence of the former on parliamentary politics. `West European countries, operating within the democratic parliamentary tradition, have established an ideology and socio-political system described as ``welfare capitalism''. The development of the welfare state was to a large extent based on the aspirations and in¯uence of labour parties and/ or the trade unions through their participation in national parliamentary politics. This was true for both North-western and Southern Europe. Especially in Catholic Europe, it was in¯uenced by the traditional ideals of Christian charity as expressed by Christian Democratic parties and trade unions. Welfare development was paralleled by improvement in health status for all segments of the population, although not to the same extent for all'. (VaÊgeroÈ and Illsley, 1992, p. 225.) In contrast, post-war developments in CEE were dominated by a weakening of democratic in¯uence. Post-1989 developments represent a strengthening of democratic in¯uence, but not to the same extent as market forces have been strengthened. The balance of power resources, in Korpi's terms, is therefore very unstable in the present transition period. The ongoing transition in CEE can be described as a ``social realignment'', accompanied by increased social disorganisation and for individuals by an increased number of stressful life events. Whenever
1357
new formal and informal networks arise, they can be expected to be socially structured. A large body of epidemiological, medical and sociological research has shown the importance of people's social networks for their state of health. Research has mostly concentrated on explanations at an individual level, i.e. individuals with a well established social network are more protected in stressful life events than others (Berkman, 1986; Cassel, 1976). Sociological research is most concerned with the social context in which social networks are signi®cant for people's health (Brown and Harris, 1978). Cohen and Syme de®ne social support as the resources received from others. Does social support have a direct eect on health, or is the eect indirect (a buer)? They admit that there is evidence for both alternatives, but they consider the indirect eects to be more important (Cohen and Syme, 1985). We know that colleagues, relatives, friends and fellow members of associations or organisations constitute social support in this sense (Hall, 1990; OÈstergren, 1991). We can expect dierences between social and occupational groups in this respect (Berkman, 1986; Cohen and Syme, 1985; Shye et al., 1995). Thus, we expect transition societies of today to be characterised by increased social dierences in health, morbidity and mortality. The mechanisms behind these are not only psychologically mediated by stress, as suggested by Watson (Watson, 1995), but also related to the distribution of economic and political resources, both individually and collectively. It is not exclusively the number of social contacts, or whether or not one is a member of an organisation which is important; it is rather the whole social context in which the individual is embedded and through which he/she receive resources to cope with health problems or health risks. MEASURING HEALTH
One way to measure people's health is to ask them to estimate their own health according to a scale. This is what the ``World Values Survey 1990'' (WVS 1990) (see below) did. Aggregated data on self-perceived health can be used for country comparisons in the same way as mortality rates often are. Both methods have their own speci®c problems, but research based on self-perceived health and mortality nevertheless gives similar results. Self-perceived health predicts mortality in longitudinal follow-up studies of individuals (Kaplan et al., 1996; Welin et al., 1985). Similarly in a European country comparison (based on World Values Survey 1980), several individual attitudes, aggregated into mean values for individual countries, were found to predict national mortality rates from suicide and homicide (Carlson et al., 1994). As a prelude to the present study I analysed the correlation between
1358
P. Carlson
Fig. 1. The relationship between self-perceived health and IHD mortality in 23 European countries, men and women aged 35±64.
self-perceived health (as measured by WVS 1990 and aggregated into country means) and IHD mortality rates in 23 western and eastern European countries (excluding Denmark and Switzerland). The results, (males aged 35±64: r = ÿ 0.65, females aged 35±64: r = ÿ 0.64, p < 0.001), (Fig. 1), convinced us of the validity and meaningfulness of selfperceived health, as measured in this study. However, in country comparisons of self-perceived health it is likely that respondents' answers are to a certain extent culturally determined. Consequently, there would have to be fairly major dierences in self-perceived health between eastern and western Europe, for them to be entirely convincing.
POPULATIONS UNDER STUDY
World Values Survey (WVS 1990) is based on a large number of national surveys conducted 1990±1991 in 42 countries world-wide. All of these surveys were carried out by means of face to face interviews, with a sampling frame consisting of all adult citizens aged 18 and over. In most cases, strati®ed multi-stage random sampling was used. In most of the countries, and in all the European countries, the samples were full national samples. This study is based on data *Age 35±39: 0.200; age 40±44: 0.186; age 45±49: 0.144; age 50±54: 0.186; age 55±59; 0.147 and age 60±64: 0.149.
from 25 European countries and from respondents aged 35±64 yr. The countries are shown in Table 1. Sampling fractions vary from country to country and countries are of dierent size, which would give some countries a disproportionate impact in any analysis involving pooled samples. A weight factor was introduced, which gives greater weight to the more populous countries than to the less populous ones, so that the pooled analyses more closely approximate European reality. The weight factor was not used in the country-speci®c analyses. Age was controlled for in all the analyses (unless otherwise stated). This was carried out by adding the age variable in all regression models with individual data. When comparing countries by using aggregated data, dierences in age structure are eliminated by a weight corresponding to the proportions of each ®ve year group in the European ``normal population''*. WVS 1990 includes several hundred variables. My selection was carried out in two steps. Firstly I picked out some 30 variables which I thought were generally interesting in the light of my perspective on the east±west health divide as described above. In the next step I excluded those variables which were found not generally to be linked to self-perceived health. I was left with the following variables, which were included in the analyses presented opposite.
Self-perceived health in East and West Europe
1359
Table 1. The 25 investigated European countries with respondents in the ages 35±64 Austria (n = 752) Belarus (n = 529) Belgium (n = 1.363) Bulgaria (n = 529) Czechoslovakia (n = 784) Denmark (n = 490) Estonia (n = 533)
Finland (n = 328) France (n = 468) Great Britain (n = 700) Hungary (n = 567) Ireland (n = 499) Italy (n = 1.003) Latvia (n = 486)
Independent variables Country, age, sex, level of education (age at completion) and occupational group (professionals, routine non-manual workers, skilled manual workers and unskilled manual workers). Membership of non-political association (Yes/ No)*. Life control (1±10) (1 = none and 10 = a great deal). Job satisfaction (1±10) (1 = dissatis®ed and 10 = satis®ed). Freedom to make own decisions at work (1±10) (1 = none and 10 = a great deal). Satisfaction with economic situation of household (1±10) (1 = dissatis®ed and 10 = satis®ed). Political interest (1±4) (1 = very interested and 4 = not at all interested). Importance of family (1±4) (1 = very important and 4 = not at all important). Importance of friends (1±4) (1 = very important and 4 = not at all important). Dependent variable Self-perceived health (1±5) (1 = very good and 5 = very bad). In Lithuania there was no question about satisfaction with household economy. In Belarus, Czechoslovakia, Poland and Switzerland there was no question about association membership.
METHODS OF ANALYSIS
WVS 1990 oers the opportunity to describe selfperceived health and to analyse the relations between some social/economic factors and self-perceived health in eastern and western Europe. First I compare the 25 countries by their mean values of self-perceived health. Secondly, with multiple regressions (OLS), I try to estimate the importance of some variables seen as determinants of individually measured self-perceived health in the 25 European countries. To assess the relative importance of the independent variables I then converted them to a *Membership of social welfare services for elderly, handicapped or deprived; education, arts, music or cultural activities; ecology groups; peace movement; sports or recreation; women's groups; other non-political associations.
Lithuania (n = 479) Netherlands (n = 534) Norway (n = 638) Poland (n = 530) Portugal (n = 546) Romania (n = 573) Russia (n = 1.057)
Spain (n = 1.953) Sweden (n = 533) Switzerland (n = 687) West Germany (n = 996) West (n = 11.490) CEE (n = 6.067)
common scale, that is, to a standardised measure. By multiplying the unstandardised b-values of the OLS regressions by the ratio of the standard deviations of the independent and the dependent variables (i.e. bxy(sx/sy)) a b-value is obtained. This slope coecient shows how many standard deviations the dependent variable will change as the result of a one standard deviation change in the independent variable, controlling for all the other independent variables in the equation (Walsh, 1990, p. 276). I assumed that any general east±west dierence in self-perceived health could result from two sources. Firstly, if variables show dierent eects (i.e. fairly dierent b-values) in eastern and western countries this would be a probable explanation. Secondly, if the average levels of an equally important determinant dier between east and west this would also help explain the dierence in self-perceived health. The models were constructed with one independent variable at a time (10 independent variables altogether) and always with age as a control. Selfperceived health is the dependent variable in all the models and the regressions were performed separately for each sex and for each country. This made a total of 500 regressions (10 2 25). I had theoretical reasons for analysing ``association membership'' on an aggregate level, since, in line with my reasoning above, countries with a high degree of association membership are expected to exercise collective control over health risks. Thus, I carried out two new regressions, one for each sex, with country as the unit (21 observations), and with the aggregated variables ``association membership'' and ``self-perceived health''. Here I will above all discuss those results that in some way can throw light on the east±west dimension in health. Lastly, I will apply the estimated net eects of the most important health determinants in CEE (analysed as a region) to the average west European levels of these determinants. This procedure enables us to predict a new value of self-perceived health, achievable if CEE were raised to the level of western Europe with regard to these particular determinants. Because of potential problems with data reliability and validity and since there is a cross-sectional study design, all results and statements about causality are tentative.
1360
P. Carlson
Fig. 2. Self-perceived poor health in 25 European countries 1990±1991, men and women aged 35±64, (5 = very bad, 1 = very good) (ranked by each country's total mean). RESULTS
Self-perceived health The average level of self-perceived health is generally worse in former communist countries than in western Europe, for both men and women (Fig. 2). The countries with the worst levels are Belarus, Hungary and Russia. The best average level of selfperceived health is to be found in Denmark, Ireland, Sweden and Switzerland. The male east± west gap in the average level of self-perceived health is 0.54 (mwest=2.19 and mCEE=2.73), in other words the average level among CEE men is about 25% poorer than among men in the west. The female gap is 0.66 (mwest=2.28 and mCEE=2.94), which means that CEE women are on a level about 29% poorer than that for women in the west. The east±west divide in self-perceived health is even wider than the sex dierence for the same variable. In Fig. 2 I have no correction for national dierences in age structure. However, the average age and the age distribution in the age span 35±64 are similar in the east and in the west, both for men and women. Membership of an organisation As previously discussed, membership of an organisation of some kind may in¯uence people's health positively in at least two ways. First, on an individual level, where membership is related to social networks and social support which, directly or indirectly, have positive eects on the individual's self-perceived health. Second, on an aggregate level, where people's civic activities positively in¯uence the achievement of the community in terms of public health.
On the individual level, people's membership of non-political organisations showed some weak eect in most of the countries investigated. This means that people who were members of at least one association tend to perceive their health as better. When the data is aggregated the association becomes stronger. At national level, the degree of membership in non-political organisations is fairly closely related to people's self-perceived health (Fig. 3). Countries with a higher degree of membership tend to have better self-perceived (public) health (men r = 0.67, women r = 0.73, p < 0.001). Western European countries, especially the Nordic countries and the Netherlands, have a substantially higher degree of organisation membership. Life control The next step was to estimate the eects of life control on self-perceived health (Table 2). The way in which people perceive their scope for exercising control over their own life showed a similar signi®cant eect on self-perceived health in almost every country, i.e. people with less ``life control'' tend to perceive their health as worse. Dierences between east and west seem to be small. The variation within each region is more striking than that between regions. These results cannot, therefore, explain the health divide. Next, I investigated the extent to which people in the dierent countries can control their own lives. When one compares the means of the ``life control'' variable some interesting dierences are immediately observable (Fig. 4). The west European countries generally report a higher level of life control than the former communist countries. The average level of life control in
Self-perceived health in East and West Europe
1361
Fig. 3. The relationship between the degree of membership in civil (non-political) associations and the proportion reporting good or very good health in 20 European countries 1990±1991, men and women aged 35±64.
the western European countries is m = 6.7 and in the CEE countries m = 6.1. This dierence in means is signi®cant (p < 0.001). In the western parts of Europe, France and the Netherlands had Table 2. Estimated eects of life control on self-perceived health in 25 European countries and in two regions (respondents in the ages 35±64); b-values (controlled for age)
Austria Belgium Denmark Finland France Great Britain Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland W. Germany West Belarus Bulgaria Czechoslovakia Estonia Hungary Latvia Lithuania Poland Romania Russia CEE *p < 0.05. $p < 0.01. %p < 0.001.
Men
Women
Western Europe 0.09 ÿ0.08* ÿ0.29% ÿ0.08 ÿ0.20$ ÿ0.22% ÿ0.16$ ÿ0.17% ÿ0.05 ÿ0.07 ÿ0.08 ÿ0.15% ÿ0.19$ ÿ0.10% ÿ0.27%
ÿ0.03 ÿ0.15% ÿ0.25% ÿ0.16 ÿ0.18$ ÿ0.25% ÿ0.19$ ÿ0.17% ÿ0.08 ÿ0.20% ÿ0.10 ÿ0.22% ÿ0.15* ÿ0.07 ÿ0.23%
ÿ0.15%
ÿ0.18%
Central and Eastern Europe ÿ0.05 ÿ0.24% ÿ0.14$ ÿ0.05 ÿ0.07 ÿ0.19$ ÿ0.11 0.10 ÿ0.08 ÿ0.18%
ÿ0.18$ ÿ0.20$ ÿ0.17% ÿ0.23% ÿ0.05 0.00 ÿ0.17$ ÿ0.19$ ÿ0.14* ÿ0.13$
ÿ0.12%
ÿ0.11%
an average level below what might be expected (m1 6.0), and in the eastern parts, the Baltic countries scored better than their eastern neighbours (m1 6.4). The worst situation was reported in Bulgaria and Belarus, with average levels of 5.1 and 5.8 respectively. It appears that even if the dierences between the two regions are not enormous they are nevertheless clear and signi®cant and can help to explain some of the existing health divide. In general, one can say that the dierences in degree of life control between the sexes are small, certainly smaller than national dierences. Household economy People's satisfaction with their household economy had a strong relation to their self-perceived health. This factor could contribute to east±west dierences if either the eects (i.e. the beta coecients) or the average levels of economic satisfaction dier. Economic satisfaction is a very important determinant in almost all countries and for both sexes (Table 3). From the results one can see that greater economic satisfaction is related everywhere to better self-perceived health, except, it seems, in one country (Austria). Figure 5 shows that none of the former communist countries shows a higher degree of economic satisfaction than any of the west European countries! This is true for both men and women. In fact there is an east±west dierence of 1.8, where the mean score of the western European countries is 6.7 and that of the CEE countries is only 4.9 (p < 0.001). Most satis®ed with their household ®nances are the Dutch and the Swiss. Least satis®ed
1362
P. Carlson
Fig. 4. People's perceived control over their lives in 25 European countries 1990±1991, men and women aged 35±64 (10 = a great deal, 1 = none) (ranked by each country's total mean).
are the Bulgarians and the Latvians. This great east±west gap is very interesting since economic satisfaction is a strong predictor of self-perceived health all over Europe. The health divide can Table 3. Estimated eects of economic satisfaction on self-perceived health in 24 European countries and in two regions (respondents in the ages 35±64); b-values (controlled for age)
Austria Belgium Denmark Finland France Great Britain Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland W. Germany West Belarus Bulgaria Czechoslovakia Estonia Hungary Latvia Lithuania Poland Romania Russia CEE *p < 0.05. $p < 0.01. %p < 0.001.
Men
Women
Western Europe 0.10 ÿ0.28% ÿ0.16* ÿ0.16* ÿ0.38% ÿ0.23% ÿ0.14* ÿ0.20% ÿ0.14* ÿ0.16$ ÿ0.17$ ÿ0.21% ÿ0.26% ÿ0.27% ÿ0.33$
0.04 ÿ0.24% ÿ0.22% ÿ0.31% ÿ0.25% ÿ0.26% ÿ0.34% ÿ0.14% ÿ0.12* ÿ0.10 ÿ0.20% ÿ0.25% ÿ0.16* ÿ0.17% ÿ0.25%
ÿ0.23%
ÿ0.20%
Central and Eastern Europe ÿ0.21% ÿ0.23% ÿ0.25% ÿ0.19* ÿ0.12* ÿ0.25% ÿ ÿ0.13* ÿ0.17$ ÿ0.20%
ÿ0.09 ÿ0.24% ÿ0.26% ÿ0.15* ÿ0.11 ÿ0.24% ÿ ÿ0.25% ÿ0.27% ÿ0.16%
ÿ0.19%
ÿ0.17%
accordingly to a certain degree be explained by dierences in economic satisfaction between eastern and western Europe. People's perception of their life control and their degree of economic satisfaction were the two variables that, on the micro level, most clearly demonstrated similar and signi®cant eects in most of the countries, in eastern as well as in western Europe. I will now move on brie¯y to discuss the results of some of the other tested variables. Social interaction Some of the tested variables concerned people's social lives. How people prioritised issues such as family and friends had some eect on self-perceived health, i.e. people who considered family and friends as important tended to have a better selfperceived health. These eects were only visible in a small number of western European countries and this variable cannot be seen as important in explaining the east±west health divide. Class and health Class is traditionally regarded as a good health predictor also in eastern Europe (VaÊgeroÈ and Illsley, 1992). However, the results from this study are not as clear as one might expect. There are dierences in health between dierent educational and occupational groups, but these dierences are here mostly valid for the western European countries. People with a lower educational level and manual workers tend to be more unhealthy than others. In the former communist countries neither occupational group nor educational level demonstrated any signi®cant eect on self-perceived health
Self-perceived health in East and West Europe
1363
Fig. 5. People's satisfaction with their household economy in 24 European countries 1990±1991, men and women aged 35±64 (10 = satis®ed, 1 = dissatis®ed) (ranked by each country's total mean).
(Hungary was one exception). The validity of the information on social class/occupation in the WVS 1990 is not known. There is clearly reason to believe that there are educational and class dierences in health in CEE, but the class pattern may also look somewhat dierent in these countries. Class may have to be conceived or measured dierently from how it was in the WVS 1990 in postcommunist countries. Class, as measured in this survey, could not help us to explain the health divide since it did not demonstrate any signi®cant eect on self-perceived health in the CEE countries.
Work and health Two work-related variables were analysed in this study: job satisfaction and decision-making latitude in one's work and their importance for health. There were no obvious east±west dierences in the eects of these variables. Job satisfaction showed some eect in almost every country, even if the beta coecients were often not signi®cant. People who were more satis®ed with their situation at work tended to have better health and the average level of satisfaction at work was lower in the former communist countries than elsewhere. The freedom to take decisions at work had eects on self-perceived health in several countries, most commonly in western Europe and among men. There was a slight tendency towards better health among people with greater decision-making latitude at work and the average level of this latitude was lower in the former communist countries. Satisfaction at work
and the freedom to make decisions may thus to some extent explain the health divide. Political interest People's political interests had some slight eects on self-perceived health in some, mostly western, countries. People interested in politics tend to perceive their health as better. This variable seems to be more important for women than for men in determining self-perceived health. Political interest did not have any signi®cant in¯uence on the health divide. The importance of membership of organisations, life control and economic satisfaction for the European health divide As I have proposed above, membership of an organisation, life control and economic satisfaction seemed to be important variables determining selfperceived health in the countries studied here. As a thought experiment, I have tried to estimate what the average level of self-perceived health level in the former communist countries would be, if the citizens in those countries were as often organised in civil associations, enjoyed as much life control, and were as satis®ed economically as people in western European countries. From a regression model (applied to the observations from the CEE countries) with the variables age, organisation membership, life control and economic satisfaction we obtain the importance of these variables in this speci®c region. The next step is to put in the average western levels of the three
1364
P. Carlson
last variables. We will now be able to estimate what the average self-perceived health level would be in CEE if these conditions were the same as in the west*. For men, the east±west dierence in self-perceived health averages is 0.54 (2.73CEEÿ2.19west). With our regression equation the estimated new east average for CEE men is 2.67 and the health gap is reduced by 11%. For women, the east±west dierence in self-perceived health averages is 0.66 (2.94±2.28). With our regression equation the new east average for women becomes 2.73, i.e. the health gap is reduced by 32%. Finally, I also estimated some east±west interaction eects from the two most powerful variables at individual level (life control and economic satisfaction separately). Both life control and economic satisfaction were signi®cantly stronger predictors of self-perceived health in western Europe (men and women). However, the interactions were not very strong. The implication of these interactions is that if eastern Europe became more similar to western Europe, increased ``life control'' and ``economic satisfaction'' would result in greater gains in self-perceived health than those estimated above. Finally, one note of caution: we cannot be sure of the direction of the established relations. It might well be true that some people, because of their ill health, perceive that they have less life control, a poorer household economy etc. We must bear this in mind when we interpret the results. DISCUSSION
In this study it is clearly demonstrated that the so-called European health divide, documented for mortality, is also noticeable in self-perceived health. Explaining this divide is hardly a simple task. The variety of social, economic and cultural structures make the picture very complex. The former communist countries, like the western European countries, are a far from homogeneous group. They will also almost certainly develop at dierent paces, and perhaps in dierent directions (Deacon, 1992). Economic performance, ``civil society'' and ``life control'' are all ``system linked''. We are looking *This will be estimated with the equation: Y = a + (age X1) + (ecsat X2) + (lifecont X3) + (civass X4), where Y = predicted average level of self-perceived health in CEE; a = intercept; age = eect of age in CEE; civass = eect of membership of a non-political association in CEE; ecsat = effect of economic satisfaction in CEE; lifecont = eect of life control in CEE; X1=average age in CEE; X2=average economic satisfaction in the west; X3=average life control in the west and X4=average degree of association membership in the west. Men Y = 2.36 + (0.02age48.21X1) ÿ (0.06ecsat6.65X2) ÿ (0.03lifecont6.70X3) ÿ (0.12civass0.47X4) and women Y = 2.33 + (0.02age48.06X1) ÿ (0.06ecsat6.58X2) ÿ (0.02lifecont6.63X3) ÿ (0.08civass0.40X4.
for explanations which allow us to link the macro properties of the two European socio-political systems to individual health risks and health perception. My results indicate that people's participation in civic activities has a positive eect on their health. This eect is strongest on a societal level, which is in line with Putnam's theory about civil activities and community performance. Concepts which, by de®nition, are ecological. In western Europe the tradition of the active citizen is presently stronger than in eastern Europe, a fact which is demonstrated by this study also. The importance of people's life control for their self-perceived health is shown to be very similar among the European countries, both west and east. In 1990±1991, people in the former communist countries did not feel that they had the same control over their lives as did people in the west. The lack of in¯uence among eastern European citizens is chie¯y a problem of democracy, but also a manifestation of a less developed civil society. The concept of civil society is however not totally clear cut. The meaning of it might very well dier from country to country. Here the operational de®nition was restricted to voluntary, non-political and non-religious associations. This de®nition most certainly has limitations. The reason for excluding political organisations and trade unions in the de®nition was the close proximity between them and the state, indeed subordination to the state, in many of the central and eastern European countries. Excluding membership in religious organisations and churches is perhaps more debatable. In Poland, for instance, the church had a central place in civil society during the last years of communism (Buchowski, 1996). The importance of religious organisations and churches and their connection to the state in the countries investigated dier considerably. It was dicult to form an opinion of the extent to which they were important in each country, and therefore church membership was not included in the de®nition of civil society in the study. What was left can thus be seen as the lowest common denominator of civil society. In spite of this limitation, the results discussed above have shown that the operationalisation is useful. This study also demonstrates the close relationship between people's satisfaction with their household economy and their self-perceived health. An unsatisfactory economic situation is most clearly related to a worse self-perceived health in 24 of the 25 investigated countries (Austria was an exception). In 1990±1991, the average level of economic satisfaction was considerably lower in CEE than in the west. This factor thus contributes importantly to the European health divide. The economic growth of the former communist countries is important for the improvement of their public health. Some of these countries had better economic preconditions than others in the early
Self-perceived health in East and West Europe
1990s (Czechoslovakia, Hungary and Poland). It is important also to bear in mind that the distribution of these resources is very important. Since 1990 the disparity in income distribution has grown dramatically in many of these countries and large sections of the populations live under the poverty line (United Nations Children's Fund, 1995). Both The World Bank and the Human Development Report, following Wilkinson (1986), have argued that a more equal income distribution and a more equal distribution of the possession of natural resources create not only a better climate for economic growth (The Independent, 21 July 1996), but also improved health and survival (World Bank, 1993). The results presented in this article re¯ect a situation basically created during the communist era, but many of the problems have been aggravated since then. Political obstacles for citizens in central and eastern Europe have been replaced with economic ones. In the context of the individual's own resources, the result may be very much the same. Those privileged today are often individuals who were privileged yesterday (see for instance Stark, 1996; Hanley et al., 1995). Thus, even if social strati®cation today occasionally has a dierent basis from earlier, where privileged individuals are concerned there is an element of continuity. Finally, some methodological considerations. With this kind of data there are always certain causality problems. For instance, does poor selfperceived health imply perceptions of less control or poorer economy? Negative aectivity, in other words the tendency to rate a number of life circumstances in a negative light, on psychological or personality grounds, is also possible (Watson and Pennbaker, 1989). Those possibilities certainly exist and must be born in mind when interpreting these results. However, during recent decades a large literature indicating similar causal relations as those suggested here has emerged (see for instance Marmot et al., 1997; OÈstergren, 1991; Karasek, 1979). Cultural dierences may also in¯uence perception of health, and there is a possibility of a bias if these perceptions were systematically dierent between east and west Europe. Two things make this less likely as an explanation. First, the level of self-perceived health was systematically linked to the heart disease rate in a country. Secondly, the dierences between east and west were quite substantial, in fact larger than between men and women. Further, it can be argued that the measures used also are somewhat distant from the theoretical concepts suggested in the introduction above. This is true to an extent but still I would argue, that some light has been shed on the question of why there is a European health divide. In sum, in spite of its limitations, the present study indicates that an improvement of public health in the CEE countries might be facilitated by
1365
improvements in household economies, by individuals achieving more control over their everyday lives and by a strengthening of civil society, seen both as a individual resource in situations of hardship and stress, and as a source of power to balance market forces. I suggest that explanations that focus exclusively on individual behaviours, such as smoking and drinking or on supply of medical services, are either one-sided or too limited. AcknowledgementsÐThe author gratefully acknowledges the support from a grant from the Research Council for the Humanities and Social Sciences (grant no F 0915/96) and Professor Denny VaÊgeroÈ at the Department of Sociology, Stockholm University for his generous advice and assistance. An earlier version of this article was presented at the European Society of Medical Sociology, 6th Biennial Conference, August 29±31, 1996, in Budapest, Hungary and at the Nordic Workshop on Health Inequalities, November 21±23, 1996, in Stockholm, Sweden. REFERENCES
Berkman, L. F. (1986) Social networks, support, and health: taking the next step forward. American Journal of Epidemiology 123, 559±562. Bobak, M. and Marmot, M. (1996) East±west mortality divide and its potential explanations: proposed research agenda. British Medical Journal 312, 421±425. Brown, G. and Harris, T. (1978) The Social Origins of Depression. Tavistock, London. Buchowski, M. (1996) The shifting meanings of civil society in Poland. In Civil Society. Challenging Western Models, ed. C. Hann and D. Dunn, pp. 79±98. Routledge, London. Carlson, P., MaÈkinen, I. and VaÊgeroÈ, D. (1994) SjaÈlvmord, mord och kultur. En jaÈmfoÈrelse av tio laÈnder i Europa. [Suicide, homicide and cultural attitudes. An investigation of ten European countries.] Sociologisk Forskning 31, 78±89. Cassel, J. (1976) The contribution of the social environment to host resistance. American Journal of Epidemiology 104, 107±123. Cohen, S. and Syme, L. S. (1985) Social Support and Health. Academic Press, Orlando. Deacon, B. (1992) The New Eastern Europe. Social Policy Past, Present and Future. SAGE, London. Durkheim, E. (1992) Suicide. A Study in Sociology. Routledge, London. Hall, E. (1990) Women's Work: An Inquiry into the Health Eects of Invisible and Visible Labor. Dissertation, Karolinska Institute, Solna. Hanley, E., Yershova, N. and Anderson, R. (1995) Russia ÐOld wine in a new bottle? The circulation and reproduction of Russian elites, 1983±1993 Theory and Society 24, 639±668. Hann, C. and Dunn, D. ed. (1996) Civil Society. Challenging Western Models. Routledge, London. Hertzman, C., Kelly, S. and Bobak, M. ed. (1996) East± West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants. Kluwer, Dordrecht. Hertzman, C. and Marmot, M. (1996) The leading hypothesis and its discontents: a synthesis of evidence and outstanding issues regarding the east±west life expectancy gap. In East±West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants, ed. C. Hertzman, S. Kelly and M. Bobak, pp. 211±219. Kluwer, Dordrecht. The Independent, 21 July 1996.
1366
P. Carlson
Kaplan, G., Goldberg, D., Everson, S., Cohen, R. D., Salonen, R., Tuomilehto, J. and Salonen, J. (1996) Perceived health status and morbidity and mortality: evidence from the Kuopio ischaemic heart disease risk factor study. International Journal of Epidemiology 25, 259±265. Karasek, R. A. (1979) Job demands, job decision latitude, and mental strain: implications for job redesign. Administrative Science Quarterly 24, 285±307. Korpi, W. (1989) Power, politics, and state autonomy in the development of social citizenship: social rights during sickness in eighteen OECD countries since 1930. American Sociological Review 54, 309±328. League of Nations, Health Organization (1932) Statistics of Noticeable Diseases for the Year 1930. Health Section, Geneve. Leon, D. A., Chenet, L., Shkolnikov, V. M., Zakharov, S., Shapiro, J., Rakhmanova, G., Vassin, S. and McKee, M. (1997) Huge variation in Russian mortality rates 1984±1994: artefact, alcohol, or what? Lancet 350, 383±388. Marmot, M. G., Bosma, H., Hemingway, H., Brunner, E. and Stansfeld, S. (1997) Contribution of job control and other risk factors to social variations in coranary heart incidence. Lancet 350, 235±239. Michnick, A. (1988) En lektion i vaÈrdighet. EssaÈer fraÊn det polska motstaÊndet 1973±1988. [A lecture in dignity. Essays from the Polish resistance 1973±1988.] Ordfronts foÈrlag, Stockholm. OÈstergren, P.-O. (1991) Psychosocial Resources and Health. With special reference to social network, social support and cardiovascular disease. Dissertation, Studentlitteratur, MalmoÈ. Peto, R., Lopez, A., Boreham, J., Thun, M. and Heath, C. (1992) Mortality from tobacco in developed countries: indirect estimation from national vital statistics. Lancet 339, 1268±1278. Putnam, R. (1993) Making Democracy Work. Civic Traditions in Modern Italy. Princeton University Press, Princeton. Shapiro, J. (1995) The Russian mortality crisis and its causes. In Russian Economic Reform at Risk, ed. A. AÊslund. Pinter, London. Shye, D., Mullooly, J. P., Freeborn, D. K. and Pope, C. R. (1995) Gender dierences in the relationship between social network support and mortality: a longitudinal study of an elderly cohort. Social Science and Medicine 41, 935±947.
Stark, D. (1996) Recombinant property in east European capitalism. American Journal of Sociology 101, 993± 1027. Tocqueville, A. (1969) Democracy in America. Anchor Books, Garden City, New York. (Cited by Putnam, 1993 p. 90.). United Nations Children's Fund (1995) Central and eastern Europe in transition. Public policy and social conditions. Poverty, children and policy: responses for a brighter future. Economies in Transition Studies. Regional Monitoring Report 3. VõÂ ctor-DõÂ az, V. M. (1993) The Return of Civil Society. The Emergence of Democratic Spain. Harvard University Press, London. VaÊgeroÈ, D. and Illsley, R. (1992) Inequality, health and policy in east and west Europe. International Journal of health Sciences 3, 225±239. Walsh, A. (1990) Statistics for the Social Sciences. With Computer Applications. Harper and Row, New York. Watson, D. and Pennbaker, J. (1989) Health complaints, stress, and distress: exploring the central role of negative aectivity. Psychological Review 96, 234. Watson, P. (1995) Explaining rising mortality among men in eastern Europe. Social Science and Medicine 41, 923± 934. Weber, M. (1958). Economy and Society, Vol. 2. University of California Press, Berkeley. Welin, L., Tibblin, G., Svardsudd, K., Tibblin, B., AnderReciva, S., Larsson, B. and Wilhelmsen, L. (1985) Prospective study of social in¯uences on mortality. The study of men born in 1913 and 1923. Lancet 20, 915± 918. Wennemo, I. (1994) Sharing the Costs of Children. Studies on the Development of Family Support in the OECD Countries. Dissertation, Institute for social research, Stockholm. Wilkinson, R. (1986) Income and mortality. In Class and Health, ed. R. Wilkinson, pp. 88±114. Tavistock, London. Wilkinson, R. (1996) Health and civic society in eastern Europe before 1989. In East±West Life Expectancy Gap in Europe. Environmental and Non-Environmental Determinants, ed. C. Hertzman, S. Kelly and M. Bobak, pp. 195±209. Kluwer, Dordrecht. Wnuk-Lipinski, E. and Illsley, R. (1990) Non-market economies in health. Social Science and Medicine 31, 833±836. World Bank (1993) World Development Report. Investing in Health. Oxford University Press, New York.