TRENDS IN ALCOHOL INTAKE BY EDUCATION AND MARITAL STATUS IN AN URBAN POPULATION IN RUSSIA BETWEEN THE MID 1980s AND THE MID 1990s

Transcription

1 Alcohol & Alcoholism Vol. 39, No. 1, pp , 2004 doi: /alcalc/agh022, available online at TRENDS IN ALCOHOL INTAKE BY EDUCATION AND MARITAL STATUS IN AN URBAN POPULATION IN RUSSIA BETWEEN THE MID 1980s AND THE MID 1990s SOFIA MALYUTINA, MARTIN BOBAK 1 *, SVETLANA KURILOVITCH, YURI NIKITIN and MICHAEL MARMOT 1 Institute of Internal Medicine, Siberian Branch of the Russian Academy of Medical Sciences, Novosibirsk, Russia and 1 International Centre for Health and Society, Department of Epidemiology and Public Health, University College London, 1 19 Torrington Place, London WC1E 6BT, UK (Received 24 June 2003; first review notified 24 July 2003; in revised form 27 October 2003; accepted 3 November 2003) Abstract Aims: We investigated changes in the distribution of alcohol consumption by education and marital status in Russia during the period of societal transformation after Such changes would indicate the potential role of alcohol in the rising social inequalities in mortality. Methods: We analysed data from three surveys in random population samples conducted in Novosibirsk as part of the WHO MONICA project in 1985/86 (1533 men, 1292 women), 1988/89 (1700 men, no women) and 1994/95 (1526 men, 1510 women), coinciding with the period of societal transformation. Four measures of drinking were examined in relation to education and marital status: prevalence of drinking at least twice a week; the mean intake in the last week; the mean intake per drinking occasion; and the prevalence of binge drinking (>80 g ethanol for men and >60 g for women) at least once a month. Results: Among men, those with university education had the lowest levels of all measures of drinking. Drinking indices increased over time in all educational groups but most sharply in men with high education, thus leading to a smaller education-related difference in the last survey. With respect to marital status, divorced and widowed men tended to drink most, but the pattern was inconsistent, and the difference between divorced and married men also narrowed over time. Among women, alcohol intake increased between the first and last survey. Differences by education and marital status in women were smaller than in men, and binge drinking was inversely related to education. Conclusions: All indices of alcohol consumption in men increased between the mid 1980s and the mid 1990s. The increase in alcohol intake among men was proportionally similar across categories of education and marital status but the absolute differences increased. The contribution of alcohol to the increase in social differentials in mortality in the 1990s was probably modest. INTRODUCTION The dramatic mortality fluctuations in Russia since the mid 1980s attracted considerable attention (Leon et al., 1997; Notzon et al., 1998), and it has been proposed that alcohol was an important proximal cause of these changes (Leon et al., 1997; Shkolnikov and Nemtsov 1997). The hypothesized powerful role of alcohol has been attributed to the binge drinking pattern (Britton and McKee, 2000) which is common in Russia (Bobak et al., 1999; Malyutina et al., 2001). There has been a debate about the extent to which alcohol has been responsible for the Russian mortality crisis (Bobak and Marmot, 1999; Vlassov and Gafarov, 2001) but the fact that alcohol has had an important influence on the health status of the Russian population is undisputed. Paralleling the increase in mortality in the early 1990s, income inequalities in Russia also increased considerably (World Bank, 1996), and there was a marked increase in educational differences in mortality over this period. Mortality rates increased more rapidly in men and women with low education than in those with higher education (Shkolnikov et al., 1998). A cohort study in St Petersburg reported similar findings for men (Plavinski et al., 2003). Given the important role of alcohol in Russia and the increase in social inequalities since the collapse of communism, it is of interest to explore whether the social distribution of alcohol consumption in Russia changed during the period of social transformation. To do so, we used three population surveys conducted by the Novosibirsk part of the WHO MONICA Project (World Health Organization, 1987). The *Author to whom correspondence should be addressed: martinb@ public-health.ucl.ac.uk surveys coincided with the period of major social changes, and thus offered a unique opportunity to assess changes in health behaviours during this period. Novosibirsk is the third largest Russian city and the industrial centre of Western Siberia, and, as far as data are available, it is typical for Russia in terms of mortality, alcohol intake and social development (Nikitin and Gerasimenko, 1995; Nemtsov, 2000; Tchernina, 2000; Malyutina et al., 2002). SUBJECTS AND METHODS The data consist of three cross-sectional surveys conducted in two administrative districts (about inhabitants) of Novosibirsk (1.4 million inhabitants) as part of the WHO MONICA Project (World Health Organization, 1987) in 1985/86, 1988/89 and 1994/1995. The participants were randomly selected from electoral lists. All participants were years old. In both genders, the response rates in individual surveys ranged from 71.2 to 73.1%. Participants were invited for an interview in a clinic. Several measures of alcohol consumption were available. First, the frequency of drinking in the last year, measured on the following scale: never; less than once a month; 1 2 times a month; once a week; 2 4 times a week; 5 6 times a week; and daily. In the present analyses, we examined the prevalence of frequent drinking (at least twice a week). Second, we analysed a continuous measure of alcohol intake at a typical drinking occasion. The amount was reported in units common in Russia (i.e. bottles of beer and grams of wine and spirits) and then recalculated into grams of ethanol. Third, we analysed the mean consumption of beer, wine and spirits during the 7 days preceding the interview. Finally, we examined the prevalence of binge drinking (at least 80 g of ethanol for men and 60 g for 64 Alcohol & Alcoholism Vol. 39, No. 1 Medical Council on Alcohol 2004; all rights reserved

2 ALCOHOL IN RUSSIA 65 women) at least once a month, estimated by combining the frequency and the typical amount. Alcohol consumption in men was measured in all three surveys for men but was not measured in women in the 1988/89 survey. The overall levels of drinking by survey were published elsewhere (Malyutina et al., 2001). We used two principal socio-economic measures: education and marital status. was categorized into four groups: primary (8 years or less, mean 6.6 years of schooling); secondary (9 12 years, mean 10.1 years of school); higher secondary (an additional 1 3 years of further education, mean 12.6 years of school); and university (a degree, mean 15.7 years of school). was classified into four groups: married; single (never married); divorced; and widowed. We calculated age-adjusted means (standard errors) and proportions of alcohol intake indices by education and marital status for each survey, separately for men and women. Within each survey and sex, we calculated P-values for trend by education, using linear regression for continuous variables and logistic regression for binary variables, and P-values for overall differences (heterogeneity) in means or proportions between categories of marital status, using ANOVA for continuous variables and the change in likelihood ratio in logistic regression (with marital status) for binary variables. We also calculated the statistical significance of the changes in the alcohol variables between 1988/89 and 1994/95 (the period of the main societal changes in Russia) among men. (This was not done for women because alcohol intake in women was not assessed in the second survey.) To allow an easier assessment of the changes in differences between socio-economic groups, we present for each survey the difference (for continuous variables) or ratio (for binary variables) between the extreme groups in terms of all-cause mortality rates; i.e. between the highest and the lowest educational group and between married and divorced subjects (Malyutina et al., 2003). We also examined the interactions between survey number and education/marital status. Where appropriate, the significance is reported in the text. RESULTS Table 1 shows the distribution of the three samples by education and marital status. In both genders, there was a tendency towards increasing education over time, and among males the proportion of unmarried men was slightly increased. More women then men were unmarried, largely due to high proportions of divorced and widowed among women. The mean alcohol intake in the last week before screening increased substantially in both sexes (Table 2). Among men, the intake was lowest in those with university education, and the absolute difference in intake increased over time. Differences by marital status were inconsistent; the intake was highest among divorced men in the first survey but among widowed men in the last survey, but some of these fluctuations are due to small numbers of divorced and widowed men. There were no clear differences in the weekly alcohol intake by education or marital status among women. The proportion of men, and, to a lesser extent, women, who drank at least twice a week increased between the mid 1980s and the mid 1990s (Table 3). In men, there was an inverse educational gradient in prevalence of frequent drinking in each survey. The temporal increase in the proportion of frequent drinkers was somewhat steeper in the highest educational category, and the ratio of primary vs university educational groups slightly declined between 1985 and In analysis of marital status, the prevalence of frequent drinking was highest in divorced men in each survey. However, the ratio of prevalence rate ratio of divorced vs married men was reduced over time (P-value for interaction between marital status and survey was 0.042), mainly because of an increase in frequent drinking among married men. Among women, the proportions of frequent drinking were low, and results for both education and marital status were inconsistent. Table 4 shows analyses of alcohol intake per typical drinking occasion. In men, there was a clear inverse educational gradient in each survey. Between the first and the last survey, the mean intake increased by some 40% in men with primary education, and the difference in dose per occasion between men with primary and university education more than doubled (P-value for interaction between education and survey was 0.004). Divorced men had the highest intake, and the difference between married and divorced men increased slightly between the first and the last surveys. Among women, the mean intake per occasion was inversely related to education, and the gradient became slightly steeper between the first and the last survey, mainly because of a Table 1. Distribution of subjects (percentages) with valid data on alcohol consumption by education and marital status in individual surveys by sex Parameter 1985/ / / / / /95 Primary NA 19.4 Secondary NA 25.6 Intermediate NA 36.7 University NA 19.3 Married (cohabiting) NA 75.0 Single NA 5.0 Divorced NA 12.1 Widowed NA 7.9 Total number NA 1510 NA: not available (alcohol was not measured in women in 1988/89).

3 66 S. MALYUTINA et al. Table 2. Age-adjusted mean alcohol intake in the last week (in grams of ethanol) by education and marital status in 1985/86, 1988/89 and 1994/95 All subjects Primary < NA 15.9 Secondary NA 19.0 Higher secondary < NA 16.7 University < NA 18.0 P for trend Difference (primary university) Married < NA 17.4 Single NA 11.2 Divorced NA 20.1 Widowed NA 15.7 P for heterogeneity < Difference (divorced married) Table 3. Age-adjusted prevalence of drinking alcohol at least twice a week (percentages) by education and marital status in 1985/86, 1988/89 and 1994/95 All subjects Primary < NA 2.1 Secondary < NA 1.4 Higher secondary < NA 1.2 University < NA 1.5 P for trend Ratio (primary/university) Married < NA 1.3 Single NA 2.2 Divorced NA 1.8 Widowed NA 0.7 P for heterogeneity Ratio (divorced/married) decline among women with university education. Differences by marital status in women were negligible. The prevalence in men of binge drinking at least once a month was inversely associated with education in both genders in each survey (Table 5). Among men, binge drinking became more common in all educational groups between the first and the last survey but, in relative terms, this increase was steepest among university educated men and the ratio of prevalence in men with primary vs university education accordingly decreased. There was interaction between education and survey number (P = 0.015), confirming the differences in the trends in binge drinking over time by education. The prevalence of binge drinking was highest in divorced men in the first survey, but was similar across the four categories in the last survey. Among women, we found an inverse educational gradient and an increase in the prevalence of binge drinking between the first and the last survey. The differences by marital status were not statistically significant. DISCUSSION This study covered an important time in Russia: the societal transformation from the Soviet system to a market economy. The first, second and third surveys corresponded, respectively, to the mid 1980s, reflecting the Soviet period; the end of the 1980s at the height of perestroika; and the early 1990s when the impact of the societal transformation was most intense. Our data confirm previous reports that Russians drink relatively infrequently, but in a high dose per drinking

4 ALCOHOL IN RUSSIA 67 Table 4. Age-adjusted mean alcohol intake per typical occasion (in grams of ethanol) by education and marital status in 1985/86, 1988/89 and 1994/95 All subjects Primary NA 38.4 Secondary NA 36.2 Higher secondary NA 30.7 University NA 23.2 P for trend <0.001 <0.001 < <0.001 Difference (primary university) Married NA 31.7 Single NA 27.1 Divorced NA 34.6 Widowed NA 32.3 P for heterogeneity < < Difference (divorced married) Table 5. Age-adjusted prevalence (percentage) of binge drinking (at least 80 g of ethanol for men and at least 60 g of ethanol for women) at least once a month by education and marital status in 1985/86, 1988/89 and 1994/95 All subjects Primary NA 12.3 Secondary NA 13.1 Higher secondary NA 9.23 University NA 1.46 P for trend <0.001 <0.001 < <0.001 Ratio (primary/university) Married NA 9.3 Single NA 5.6 Divorced NA 12.8 Widowed NA 7.0 P for heterogeneity Ratio (divorced/married) occasion. In all three periods, we found educational differences in all major alcohol consumption indices in men. Differences by marital status were seen mainly between divorced/widowed men and others. There was a tendency towards increased male drinking in the1990s; this led to an increase in absolute differences but to a decrease in relative differences by education and marital status. Because of low levels of drinking among women, the results on females were unstable and inconsistent. Potential limitations of the study Several limitations of this study need to be considered when interpreting the results. First, the study was conducted in Novosibirsk, and is therefore not directly generalizable for the whole of Russia. However, trends in mortality and drinking in Novosibirsk are close to national averages (Nikitin and Gerasimenko, 1995; Nemtsov, 2000; Malyutina et al., 2001, 2002), and Novosibirsk probably represents the social patterns common in Russian urban populations (Tchernina, 2000). The study participants were randomly selected from electoral lists and were interviewed and examined at a clinic. The response rates were high and consistent across surveys. The electoral lists in Russia are regularly updated and are considered complete and reliable. There were no dramatic changes in education or marital status between surveys, although there was a tendency towards increasing educational level and the proportion of married men declined slightly. While these changes seem to reflect genuine social trends,

5 68 S. MALYUTINA et al. we cannot exclude the possibility that the composition of nonresponders with respect to their drinking habits may have changed over time. If heavy drinkers have become more likely to participate in the study, and if the increased participation was related to education or marital status, the results would be unreliable. However, there is no indication that this is the case, and a serious selection bias of this study is therefore unlikely. Second, as with all self-reported measures, it is possible that some measurement bias may have occurred in relation to alcohol consumption. A sub-sample of subjects interviewed in 1994/95 was re-interviewed and provided a blood sample in 2000; both self-reported consumption and serum gammaglutamyl transferase (GGT) correlated well with the original reported alcohol intake (Malyutina et al., 2002; S. Kurilovitch, unpublished). This does not suggest a systematic bias. Random mis-classification, due to imprecise reporting, would tend to attenuate any underlying differences between educational and marital status groups. Male drinking in Russia has not been associated with social stigma, but drinking by women has been considered socially undesirable (though the attitudes towards female drinking and smoking probably started changing in the 1990s). This would lead to under-estimation of levels of drinking among women. Earlier results on GGT and hangover frequency suggested some under-reporting of alcohol intake in women, but the magnitude of such a bias is small (S.Kurilovitch, unpublished). If the social stigma in women differed by education groups or marital status, differences in drinking may be biased but we have no data with which to assess this possibility. Finally, the direction of the association between drinking and marital status is not clear. Usually it is thought that socioeconomic factors determine life styles. However, heavy drinking has a number of negative consequences, including strained family relationships, and it is likely that the high levels of drinking seen in divorced men actually led to divorce (Caces et al., 1999; Leonard and Rothbard, 1999). This problem should not occur with education; this is a stable characteristic which is very unlikely to be changed by alcohol. and drinking Among men, we found an inverse educational gradient in drinking. Previous studies in Russia reported inconsistent results. In a Russian national sample interviewed in 1996, there was a weak and insignificant inverse educational gradient in prevalence of frequent drinking and binge drinking among men and a stronger gradient in women (Bobak et al., 1999). The difference between this and the present study is probably due to the quota sampling approach adopted in the 1996 study, possibly leading to selective non-response in some groups. The educational gradient may also be stronger in urban areas and thus appear weaker in a national sample. Another urban community-based study in the Russian city of Taganrog found a clear inverse gradient in heavy drinking, similar to the present study (Carlson and Vagero, 1998). In the St Petersburg cohort, weekly alcohol intake was also strongly inversely associated with education (Plavinski et al., 2003). In the west, the association between alcohol and socioeconomic position is more complex. While the the average intake is usually higher in higher social groups, alcoholrelated harm is more common in lower socio-economic groups (Crum et al., 1992; Thun et al., 1997; Marmot and Feeney 1999; Ferrieres et al., 2001). This paradox is most likely to be explained by a more risky drinking pattern (e.g. binge or heavy drinking) in lower socio-economic groups (Marmot and Feeney, 1999). The pattern seen in our Russian data is consistent with the general pattern of higher levels of unhealthy behaviours in lower socio-economic groups, with the difference that all and not just some drinking indices in males were inversely associated with education. In women, the lack of a clear relation between education and mean alcohol intake in the last week and prevalence of frequent drinking was probably due to low statistical power. and drinking In western populations, alcohol consumption is, in general, higher in unmarried subjects (Leonard and Rothbard 1999; Power et al., 1999) but the available data from Russia are inconsistent. We found that levels of drinking were, in general, highest in divorced and, particularly in the most recent survey, widowed men. This contrasts with the above mentioned study in a national sample, in which the prevalence of frequent drinking and binge drinking was similar in divorced and married subjects (Bobak et al., 1999). As with education, the difference between the two studies is most likely to be due to different sampling methods. However, the Taganrog study mentioned above also found only small differences in prevalence of heavy drinking by marital status, although heavy drinking in men was associated with marriage problems (Carlson and Vagero, 1998). It is possible that the relationship between marital status and alcohol in Russia may differ by urbanization, geography or calendar period. It has been proposed that marriage is, in general, associated with better health and that unmarried subjects are more likely to be engaged in unhealthy behaviours, such as smoking, heavy drinking and poor diet (Hajdu et al., 1995; Robles and Kiecolt- Glaser, 2003). With respect to drinking, our data provide only a partial support for this proposition. Divorced men had high rates of frequent drinking and a high mean dose per drinking occasion, but there were no differences in the proportion of binge drinkers by marital status. Data on women were also inconsistent. Overall, we found no overwhelming evidence of a high risk of heavy drinking in unmarried persons. Trends in drinking by education and marital status Among men, the absolute differences by education in all drinking indices increased between the mid 1980s and the mid 1990s. The relative differences remained stable or even decreased. Trends in differences by marital status in men followed a similar pattern the absolute differences between divorced men and the others widened, but the relative differences tended to narrow. Given these results, we were interested in whether alcohol could have contributed to the widening of educational differences in mortality that occurred in Russia in the early 1990s (Shkolnikov et al., 1998). As an illustration, we have calculated the possible effect of binge drinking on the all-cause mortality rate ratio between men with primary vs university education, assuming an approximately double all-cause mortality in binge drinkers compared with non-binge drinkers (an assumption consistent with results of the Novosibirsk

6 ALCOHOL IN RUSSIA 69 cohort; Malyutina et al., 2002), and using the prevalence rates of binge drinking in the first and last survey. Under these assumptions, binge drinking alone would have led to a mortality rate ratio (primary vs university) of 1.14 in the first survey and 1.23 in the last survey. This is a modest contribution, compared with the dramatic increase in Russian mortality differentials reported in the literature (Shkolnikov et al.,1998; Plavinski et al., 2003). The contribution of alcohol to educational differentials could be substantially larger, however. Firstly, the heavy drinking pattern may have become more harmful than what could be captured by the dichotomous criterion of binge drinking (e.g. drinkers seem to consume much larger doses per drinking session). Secondly, heavy drinking may interact with other factors and the population-attributable risk fraction could thus be larger. On the other hand, the estimated modest contribution of alcohol to the increase in educational gradient is consistent with the St Petersburg cohort in which the changes in male all-cause mortality between the 1980s and the 1990s in each educational stratum were similar in men with low and high alcohol intakes (Plavinski et al., 2003). There are no data on trends in mortality differentials by marital status in Russia, though it is commonly assumed that unmarried and especially divorced men in the former communist countries have an increased risk of mortality (Hajdu et al., 1995; Watson 1995). As with education, however, our data suggest that the widening of differences in alcohol intake between unmarried and married men was not large enough to result in dramatic changes in mortality differentials. Acknowledgements Data collection in Novosibirsk was funded by the Siberian Branch of the Russian Academy of Medical Sciences. The analyses were supported by a grant from the Wellcome Trust. M.M. is a recipient of the MRC Research Professorship. REFERENCES Bobak, M. and Marmot, M. (1999) Alcohol and mortality in Russia: is it different than elsewhere? Annals of Epidemiology 9, Bobak, M., McKee, M., Rose, R. and Marmot, M. (1999) Alcohol consumption in a national sample of the Russian population. Addiction 94, Britton, A. and McKee, M. (2000) The relation between alcohol and cardiovascular disease in Eastern Europe: explaining the paradox. Journal of Epidemiology and Community Health 54, Caces, M. F., Harford, T. C., Williams, G. D. and Hanna, E. Z. (1999) Alcohol consumption and divorce rates in the United States. Journal of Studies of Alcohol 60, Carlson, P. and Vagero, D. (1998) The social pattern of heavy drinking in Russia during transition. Evidence from Taganrog European Journal of Public Health 8, Crum, R. M., Bucholz, K. K., Helzer, J. E. and Anthony, J. C. (1992) Risk of alcohol abuse and dependence in adulthood: the association with educational level. American Journal of Epidemiology 135, Ferrieres, J., Ruidavets, J. B., Bataile, V., Arveiler, D., Amouyel, P. and Ducimetiere, P. (2001) Screening of heavy drinkers in a large population based sample. European Heart Journal 22 (Suppl.), 462. Hajdu, P., McKee, M. and Bojan, F. (1995) Changes in premature mortality differentials by marital status in Hungary and in England and Wales. European Journal of Public Health 5, Leon, D. A., Chenet, L., Shkolnikov, V., Zakharov, S., Shapiro, J., Rakhmanova, G., Vassin, S. and McKee, M. (1997) Huge variation in Russian mortality rates : artefact, alcohol, or what? Lancet 350, Leonard, K. E. and Rothbard, J. C. (1999) Alcohol and the marriage effect. Journal of Studies on Alcohol (Suppl), Malyutina, S., Bobak, M., Kurilovitch, S., Gafarov, V., Simonova, G., Nikitin, Y. and Marmot, M. (2002) Relation between heavy and binge drinking and all-cause and cardiovascular mortality in Novosibirsk, Russia: a prospective cohort study. Lancet 360, Malyutina, S., Bobak, M., Kurilovitch, S., Ryizova, E., Nikitin, Y. and Marmot, M. (2001) Alcohol consumption and binge drinking in Novosibirsk, Russia, Addiction 96, Malyutina, S., Bobak, M., Simonova, G., Gafarov, V., Nikitin, Y. and Marmot, M. (2003), marital status and all-cause and cardiovascular mortality in Novosibirsk, Russia: a prospective cohort study. Annals of Epidemiology, in press. Malyutina, S., Simonova, G. and Nikitin, Y. (2002) Coronary heart disease and cardiovascular mortality in the urban Siberian population: gender specific findings from a 10-year cohort study. In Heart disease: environment, stress and gender, G. Weidner, M. Kopp and M. Kristenson eds, pp IOS Press, Amsterdam. Marmot, M. and Feeney, A. (1999) Inequality, alcohol and alcoholassociated harm. In Inequalities in health. A series of seminars held by the Health Authority, S. Waller, A. Crosier and D. McVey, eds, pp Health Authority, London. Nemtsov, A. (2000) Estimates of total alcohol consumption in Russia, Drug and Alcohol Dependence 58, Nikitin, Y. P. and Gerasimenko, N. F. (1995) Health of Siberian population. Siberian Branch of the Russian Academy of Medical Sciences, Novosibirsk (in Russian). Notzon, F. C., Komarov, Y. M., Ermakov, S. P., Sempos, C. T., Marks, J. S. and Sempos, E. V. (1998) Causes of declining life expectancy in Russia. Journal of the American Medical Association 279, Plavinski, S. L., Plavinskaya, S. I. and Klimov, A. N. (2003) Social factors and increase in mortality in Russia in the 1990s: prospective cohort study. British Medical Journal 326, Power, C., Rodgers, B. and Hope, S. (1999) Heavy alcohol consumption and marital status: disentangling the relationship in a national study of young adults. Addiction 94, Robles, T. F. and Kiecolt-Glaser, J. K. (2003) The physiology of marriage: pathways to health. Physiology and Behavior 79, Shkolnikov, V., Leon, D. A., Adamets, S., Andreev, E. and Deev, A. (1998) al level and adult mortality in Russia: an analysis of routine data 1979 to Social Science and Medicine 47, Shkolnikov, V. M. and Nemtsov, A. (1997) The anti-alcohol campaign and variations in Russian mortality. In Premature death in the New Independent States, J. L. Bobadilla, C. A. Costello, and F. Mitchell eds, pp National Academy Press, Washington, DC. Tchernina, N. (2000) Rising unemployment and coping strategies: the case of the Novosibirsk oblast in Russia. In The mortality crisis in transitional economies, G. A. Cornia and R. Paniccia eds, pp Oxford University Press, New York. Thun, M. J., Peto, R., Lopez, A. D., Monaco, J. H., Henley, J., Heath, C. W. and Doll, R. (1997) Alcohol consumption and mortality among middle-aged and elderly US adults. New England Journal of Medicine, 337, Vlassov, V. and Gafarov, V. (2001) Mortality in Russia. Lancet 358, 669. Watson, P. (1995) Explaining rising mortality among men in Eastern Europe. Social Science and Medicine 41, World Bank (1996) World Development Report From plan to market. Oxford University Press, New York. World Health Organization (1987) Multinational monitoring of trends and determinants of cardiovascular diseases MONICA Project. Manual of Operations. Version 1.1. CDV/MNC. December World Health Organization, Geneva.