The European Journal of Public Health Advance Access originally published online on February 7, 2008
The European Journal of Public Health 2008 18(3):329-334; doi:10.1093/eurpub/ckn005
Miscellaneous |
Religious affiliation and acute coronary syndrome: a population-based case-control study in Tirana, Albania
Genc Burazeri1, Artan Goda2, Enver Roshi1 and Jeremy D. Kark3
1 Department of Public Health, Faculty of Medicine, Tirana, Albania
2 Cardiology Department, University Hospital Centre ‘Mother Teresa’, Tirana, Albania
3 Hebrew University–Hadassah School of Public Health and Community Medicine, Ein Kerem, Jerusalem, Israel
Correspondence: Correspondence:Genc Burazeri MD, PhD, Faculty of Medicine, Rr. ‘Dibres’, No. 371, Tirana, Albania, tel: +355682150535, fax: +3554235903, e-mail: gburazeri{at}yahoo.com
Received May 31, 2007, accepted January 17, 2008
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Background: Our aim was to assess the association of religious affiliation (Muslim versus Christian) with acute coronary syndrome (ACS) in Albania, a predominantly Muslim country in Southeast Europe. Methods: A population-based case-control study was conducted in Tirana, the Albanian capital, in 2003–2006. Of non-fatal consecutive ACS patients, 467 were recruited (370 men aged 59.1 ± 8.7 years and 97 women 63.3 ± 7.1 years, 88% response). The coronary heart disease-free control group comprised 469 men (53.1 ± 10.4 years) and 268 women (54.0 ± 10.9 years) (69% response), 452 and 237 of whom were fully examined. Information collected included sociodemographic, psychosocial and behavioural characteristics by structured interview and anthropometric measurements. Furthermore, data on religious affiliation was available for all but 20 of the non-respondents. Multivariable-adjusted logistic regression was used to assess the association of religious affiliation (Muslim versus Christian) with ACS. Results: Of ACS patients, 77.1% were Muslims compared with 65.8% of the entire control group. Muslims in both sexes were at higher risk of ACS than Christians (age- and sex-adjusted OR = 1.8, 95% CI = 1.4–2.3, P < 0.01). The association persisted (OR = 1.6, 95%CI = 1.1–2.3, P = 0.02) upon further adjustment for marital status, family size, education, income, employment status, social position, emigration of close relatives, financial loss and coronary risk factors. Conclusions: In this transitional country, we found a higher risk of ACS in Muslims than Christians, independent of the socioeconomic circumstances and conventional coronary risk factors assessed. This finding requires replication and the determinants of the excess risk sought.
Keywords: acute coronary syndrome, Albania, Christian, Muslim, religion
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Reports on coronary heart disease (CHD) morbidity and mortality rates in Muslim populations as compared with other religious denominations are inconsistent.1–5 In Jerusalem, higher CHD incidence and mortality was found in Palestinian Arabs, a transitional population, than in Jewish residents.1,2 Small studies of the prevalence of CHD in India4 and South Africa5 showed little or no difference between Muslims and Christians.
A previous hospital-based case-control study of first acute myocardial infarction in Tirana showed no significant difference between Muslims and Christians,6 but both the confidence intervals and the study design did not preclude this possibility.
In Albania, public religious observance was officially prohibited in 1967.7,8 Following the breakdown of the communist regime in 1991, numerous mosques and churches were reopened.7 It has been estimated that 70% of the population are Muslim, 20% are Orthodox Christian and 10% are Roman Catholic.7,9 However, the last official census recording religious affiliation in Albania was conducted in 1938. CHD mortality and morbidity appear to have increased in Albania over the past 10–15 years,10,11 though to a lesser extent than some other countries undergoing rapid transition following collapse of communist rule. The specific rates by religious denomination, however, have not been systematically documented. In a country that is homogenous in terms of ethnicity (about 97% are ethnic Albanians,7) religious affiliation may be an informative cultural factor that distinguishes between population groups and might serve as a risk marker of differential health status.
In this context, we aimed to assess the association of religious affiliation with CHD among residents of Tirana, the Albanian capital. In the absence of previous population representative data, we hypothesized that rates of CHD would be similar between Muslims and Christians, reflecting similar lifestyle habits and behaviours induced by prolonged regime-mandated uniformity and secularism during the period of communist rule.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Study population
A population-based case-control study of acute coronary syndrome (ACS) was conducted in Tirana in 2003–2006. Details of the study design, population, sampling procedures and case definition are described elsewhere.12 Briefly, we recruited 467 consecutive non-fatal ACS patients (370 men aged 59.1 ± 8.7 years, and 97 women 63.3 ± 7.1 years; 88% response), all Tirana residents who were admitted to the University Hospital Centre with unstable angina, Q-wave acute myocardial infarction (MI) or non-Q-wave acute MI. Besides its role as a nationwide tertiary hospital, this is the only facility in Tirana which admits ACS patients. The definition of ACS was based on combinations of clinical signs and symptoms (location, patterns and duration of pain), sequential ECGs (ST-T changes, major Q-wave evolution), echocardiographic criteria (complementary examination for assessment of disorders of segmental kinesis and left ventricular function) and elevated cardiac enzymes in selected patients.12 Of the patients, 301 experienced a first ACS event [72 Christians (of whom 42% had unstable angina and 58% had MI) and 229 Muslims (of whom 38% had unstable angina and 62% had MI)] and 166 had repeat events [35 Christians (of whom 43% had unstable angina and 57% had MI) and 131 Muslims (of whom 53% had unstable angina and 47% had MI)]. The control group comprised a population-based sample of Tirana residents. The sampling frame consisted of the entire population of the municipality of Tirana, as registered in the Albanian census of April 2001. This population was, therefore, considered representative of the source population that generated the ACS cases. Based on a population list received from the National Institute of Statistics, an age- and sex-stratified sample of the adult population was drawn. We sampled 180 men and 120 women in each of four age groups: 35–44, 45–54, 55–64 and 65–74 years, for a total of 1200 individuals, 720 men and 480 women.12 Of 1065 eligible controls (655 men and 410 women), 737 individuals, recruited by home visits, participated in the study (469 men aged 53.1 ± 10.4 years, and 268 women 54.0 ± 10.9 years), for an overall response rate of 69.2%.12
Among the controls sampled, 96 men and 87 women who refused to participate (15 and 21%, respectively, of those eligible) agreed to provide information on educational level, employment status and religious affiliation. For 86% of the 82 men and 63 women who could not be located (12 and 15%, respectively, of the eligible) religion was inferred, after consulting experts on the genesis of names in Albania, from inspection of their names, which in most instances discriminates quite clearly between Muslims and Christians. Of the entire control sample, data on religious affiliation were missing in <2%, 14 men and 6 women. Female non-respondents and respondents were similar with regard to age, religion [OR (Christian versus Muslim) = 1.1, 95% CI = 0.7–1.6], employment status and educational level. Male non-respondents were moderately older than respondents (55.1 ± 10.7 years and 53.1 ± 10.5 years, respectively; P = 0.03) and were significantly more likely to be retired (age-adjusted OR = 3.8, 95% CI = 1.8–7.7), but were similar in regard to religion (OR = 1.2, 95% CI = 0.8–1.7) and educational level.
Data collection
Patients and controls were interviewed and examined in the hospital using identical procedures. A structured questionnaire, including demographic, socioeconomic, psychosocial and classical coronary risk factor characteristics, was administered by the same interviewers to cases and controls. The examination included standardized measurements of weight, height and waist and hip circumferences.12
Participants were asked about their religious affiliation (Muslim, Christian Orthodox, Roman Catholic), which we grouped in the analyses into Muslim and Christian. Socioeconomic characteristics consisted of educational level (0–8 years, 9–12 years and >12 years), employment status (employed, unemployed and retired), relative income (ranked on a 5-point scale relative to the average per capita income in Albania and grouped into three: much lower/lower, about the same and higher/much higher13) and subjective social position [on a 5-rung ladder representing the most well off to the worst off people in Albanian society, modified from Adler et al.,13–15 and grouped into lower (scores 1 and 2), middle and upper (scores 4 and 5)].
Leisure-time exercise was measured by questions tapping the frequency, intensity, and duration of walking (for pleasure, visits, shopping or chores), daily climbing of stairs, and of up to three different sport activities over the past year. Daily energy expenditure was calculated according to Taylor et al.16
The analyses included also data on marital status, number of children and siblings, family type (nuclear versus extended), current smoking, self-reported hypertension and diabetes, family history of CHD, body mass index (BMI), waist-to-hip ratio,12 alcohol intake (number of drinks per week), household financial loss in pyramid savings schemes, which collapsed in Albania in 1997 (yes versus no)17 and emigration of spouse and/or offspring and financial remittance (categorized as no emigrant, emigrants who remit money and emigrants who do not).13
The study was approved by the Albanian Committee of Medical Ethics. Participants gave written consent after being informed about the aims and procedures of the study.
Statistical analysis
Of the 737 control respondents, 54 with evidence of pre-existing CHD (based on a structured questionnaire and ECG) were excluded from the analysis. A further 48 controls (17 men and 31 women) underwent a partial interview without anthropometric measurements,12 leaving for analysis 635 controls (415 men and 220 women) who were both interviewed and examined.
Sex-specific binary logistic regression was used to assess the association of religious affiliation (Muslim versus Christian) with covariates (sociodemographic characteristics and conventional risk factors for CHD) and with ACS events. Age-adjusted or age- and sex-adjusted odds ratios (ORs), their 95% confidence intervals (CIs) and P-values were calculated. Subsequently, sociodemographic factors (marital status, family size, educational level, employment status, relative income and social position), financial loss and emigrant status, smoking, leisure-time exercise, hypertension, diabetes, family history of CHD, BMI and waist-to-hip ratio were introduced into the models.
In order to practically eliminate the possibility of selection bias due to non-response in the controls, age-adjusted models and limited multivariable-adjusted models were run with inclusion of the partially interviewed controls (n = 48), the refusals (n = 183), and those not located (n = 125), leaving only 20 individuals (<2%) not accounted for in the entire control group. To compare the associations for first versus repeat ACS events and for MI versus unstable angina, multinomial logistic models were applied. The Hosmer-Lemeshow test was used to assess goodness-of-fit; all analyses met the criterion.18 The Statistical Package for Social Sciences (SPSS for windows, version 11.0, Chicago, IL, USA) was used for all the statistical analyses.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Among ACS patients, 76% of men (n = 280) and 82% of women (n = 80) were Muslims as compared with 67% of men (n = 278) and women (n = 148) in the fully examined control group (not shown in tables). Similar proportions of Muslims (66%) were seen in the entire control group including partial respondents, non-respondents and people not located.
In the control group (table 1), Christians of both sexes were more educated than Muslims (P < 0.01 in men, P = 0.02 in women) and had fewer siblings (P < 0.01 in men, P = 0.01 in women). Muslims were more likely to live in extended families. Muslim men were more likely to be married (P = 0.02) and had more children (P = 0.01). Muslim women were more likely to be unemployed than Christians (P = 0.02) and reported a lower social position (P = 0.02). There were no statistically significant differences in the distribution of conventional CHD risk factors by religion within each sex. Muslims were not more obese. Muslim women tended to smoke less than Christian women, while the reverse was noted for Muslim men (P = 0.05 for interaction). Alcohol intake was similar in Muslims and Christians (table 1). Similar patterns were evident in ACS cases, except for a lower prevalence of hypertension in Christians than Muslims (OR = 0.4, 95% CI = 0.3–0.8) [not shown in tables].
|
Age- and sex-adjusted risk factor associations with ACS are shown in table 2. In both religious groupings, ACS patients were older than controls—a difference which resulted from the mode of selection of controls,12 were more likely to be unemployed and/or retired, had a lower income, reported more financial loss and were more likely to have emigrant relatives who did not remit funds, but did not show an association with educational level. In Christians alone there was an inverse (protective) association with living in an extended family (OR = 0.45, 95% CI = 0.25–0.78). ACS patients of both religions were more obese, including male pattern abdominal obesity, smoked more, reported a higher prevalence of diabetes and were more likely to report a family history of CHD. In Muslims only, there was a positive association of ACS with self-reported hypertension. Protective effects were noted for alcohol intake and exercise (table 2).
|
In age-adjusted analyses (table 3–upper panel), there was evidence of an excess ACS risk in Muslims, which was of similar magnitude in models including all controls (model 1, sex-adjusted OR = 1.8, 95% CI = 1.4–2.3), in those excluding controls not located (model 2, OR = 1.8, 95% CI = 1.3–2.3) and in models restricted to fully examined controls (model 4, OR = 1.6, 95% CI = 1.2–2.2).
|
Upon adjustment for educational level and employment status (table 3, lower panel, model 5) and subsequently for relative income, smoking, hypertension and diabetes (model 6), the excess ACS risk in Muslims persisted undiminished (sex-adjusted OR = 1.8, 95% CI = 1.3–2.4 for both models). Further adjustment (model 7) for marital status, measures of family size, subjective social position and other coronary risk factors (leisure-time exercise, family history of CHD, BMI and waist-to-hip ratio) slightly attenuated the strength of the relationship (OR = 1.6, 95% CI = 1.1–2.3). Subsequent adjustment for alcohol intake, and two important psychosocial predictors of ACS in our sample (financial loss in pyramid schemes17 and emigration of close relatives13) (model 8) did not affect the results in either sex (sex pooled OR = 1.6, 95% CI = 1.1–2.3). In all analyses the association in women appeared to be stronger than in men.
There appeared to be a stronger association of religious affiliation with repeat ACS events than with first events (fully adjusted sex pooled multinomial OR = 1.8, 95% CI = 1.1–3.6 and OR = 1.35, 95% CI = 0.8–1.9) that was not statistically significant (P for interaction = 0.26). The association for unstable angina (fully adjusted sex-pooled OR = 1.6, 95% CI = 1.0–2.7) was similar to that for MI events (Q-wave and non-Q-wave) [OR = 1.4, 95% CI = 0.8–2.0].
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
We found an excess risk of non-fatal ACS in Muslims which was evident in both sexes, possibly more so in women, and persisted upon multivariable adjustment for a wide array of sociodemographic, socioeconomic and psychosocial characteristics and well-established CHD risk factors. It is unlikely that the relationship resulted from a response bias among controls. We had access to data on religious affiliation on 98% of the population representative controls irrespective of response status, thus essentially excluding this source of selection bias. The association was modestly enhanced when all the population representative controls, including non-respondents, were included in the analysis, reflecting our least biased estimate of the risk marker status of religious affiliation. The relatively high response rate in the cases (88%) argues against non-response in the ACS patients causing such bias; the non-respondents would have to have been overwhelmingly Christian to substantially affect the result. This seems unlikely. Among the controls, at least, non-response was more common among Muslims.
In general, Muslims were of a lower socioeconomic status than Christians, being less educated and among women Muslims were more likely to be unemployed and to perceive themselves as occupying a lower social position (table 1). Yet, the excess ACS risk in Muslims persisted after controlling for socioeconomic variables that included educational level, relative income, subjective social position, employment status, marital status, family size, and living in extended families, emigration of first degree family members13 and financial losses during the economic collapse of 1997,17 in addition to adjustment for conventional coronary risk factors. It is possible, but unlikely, that plasma lipids which were not measured in this study, differed between Muslims and Christians to such a degree as to have produced such excess risk in Muslims. This possibility is further reduced given the overall absence of important differences in conventional risk factors between the communities. The determinants of excess risk in Muslims should be sought elsewhere, although residual confounding for socioeconomic variables and/or conventional risk factors including lifestyle factors cannot be entirely excluded.
A previous hospital-based case-control study in Tirana conducted by Roshi et al.6 reported no substantive association between first non-fatal MI and religious affiliation in either sex (in men: OR = 1.20, 95% CI = 0.85–1.70; in women: OR = 0.97, 95% CI = 0.59–1.60). In Roshi's study patients were not restricted to residents of Tirana in contrast with the current study (which may have introduced bias in Roshi's estimates) and controls were hospital-based rather than population-based (which would introduce bias if the sociodemographic and socioeconomic determinants of admission differed between the patients and hospital controls). A further difference lies in the inclusion of first and recurrent cases of ACS in the current study, where the association appeared to be stronger in the recurrent cases, but not significantly so. Restriction of the analysis in the current study to first events of acute MI, adjusting for the same variables as Roshi, modestly weakened the association (OR = 1.41, 95% CI = 0.91–2.17 in men and OR = 1.98, 95% CI = 0.68–5.76 in women) with substantial overlapping of the confidence intervals of the two studies.
A limitation of our study lies in the response rate in the control group, which was 69%. However, the age-adjusted analysis, that included 98% of the entire control group irrespective of participation status, suggests that the association in the fully examined sample might be modestly underestimated. In addition, fatal out-of-hospital events, in-hospital fatalities and non-hospitalized ACS cases were not included in this study. It is possible that there was a preponderance among these cases of Christians, but this is unlikely as the Christians are more educated and thus more likely to seek care when needed. We cannot dismiss the possibility that there is a greater propensity to lethal arrhythmias and out-of-hospital deaths among Christians, although the likelihood of this explanation for our findings is low. In our opinion, it is more likely that ACS events may be underestimated in the less educated Muslim population.
Our findings of a higher ACS risk in Muslims in Albania, a country in transition, are compatible with reports from Jerusalem which show higher CHD incidence and mortality in Arab as compared with Jewish residents, with a parallel excess in Muslim women.1,2 There too the Palestinian Muslim population was in transition from a traditional lifestyle to a more Westernized pattern. In our study, however, there were no significant differences between Muslims and Christians with regard to obesity, diabetes or hypertension, which have been suggested to partly account for the excess CHD risk in Muslim populations in Israel1,2 or Qatar.19 In Bosnia, the excess mortality risk in Muslim men was attributed to the particularly high prevalence of smoking20 but, again, in our study we did not find marked differences in smoking rates by religious denomination, and furthermore, the finding persisted after controlling for smoking. It is possible that the stressors accompanying the social, cultural and economic changes of transition are differentially perceived by or differentially affect cultural groups, with the Muslim population possibly undergoing a more stressful and costly transitional process in terms of health outcomes. Muslim affiliation might tap some unfavourable aspects not measured in our study that are related to a sense of autonomy, fulfilment and social participation, particularly among Muslim women. It has been convincingly argued that the lower the individuals are in the social hierarchy, the less likely it is that their fundamental human needs for autonomy and integration into society are met.21 Failure to meet these needs has been suggested to lead to metabolic and endocrine changes that in turn lead to an increased risk of CHD.21 It is conceivable that in comparison with Christians the apparently socially disadvantaged Muslims may cope less well with the stressors of rapid social transition or regional developments (war in the neighbouring countries), increasing their susceptibility to CHD. Furthermore, Muslim affiliation may serve as a sociocultural marker reflecting a cumulative socioeconomic disadvantage at different stages of the life course, which could place individuals of this religious denomination at a higher risk due to lower education, occupation and overall social position,22 which may not have been adequately taken into account in our study.
Based on official mortality statistics, CHD rates have increased following the breakdown of communist regime in Albania, but to a far lesser degree than trends observed in other former communist countries in Central and Eastern Europe and the former Soviet Union.8,20,23 Two models have been suggested to account for the increased morbidity and mortality risk in transitional post-communist countries: one emphasizes the importance of lifestyle/behavioural factors, in particular the role of alcohol as it seems to be the case in Russia, Ukraine or Belarus,23,24 whereas the second focuses on the role of psychosocial stress induced by economic hardship and lack of transparent and fair conditions for everyone,25,26 which may be more applicable for Albania. In the former communist countries of Europe, in particular Russia, socioeconomic inequalities have increased markedly,27 a process that has almost certainly occurred in Albania and may impinge on Muslim–Christian differences.
In conclusion, our study points to a higher ACS risk in Muslims than Christians in Albania in both sexes, irrespective of measured sociodemographic characteristics and recognized CHD risk factors. This finding requires confirmation in Albania and other settings and the determinants of the excess risk in Muslims should then be sought.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
G.B. was a recipient of an Irma Milstein International Doctoral Fellowship at the Hebrew University–Hadassah Braun School of Public Health and Community Medicine, Jerusalem, Israel. The study was supported by the Irma Milstein International Doctoral Program. We thank Mrs Bella Adler, who provided statistical advice, and Mrs Milva Ekonomi, former director of the Institute of Statistics in Tirana, who helped in drawing a population representative control group from the census conducted in Albania in 2001.
Conflicts of interest: None declared.
Key points
|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
1 Kark JD, Fink R, Adler B, et al. The incidence of coronary heart disease among Palestinians and Israelis in Jerusalem. Int J Epidemiol (2006) 35:448–57.
2 Kark JD, Gordon ES, Haklai Z. Coronary heart disease mortality among Arab and Jewish residents of Jerusalem. Lancet (2000) 356:1410–11.[CrossRef][Web of Science][Medline]
3 Abramson JH, Gofin R. Mortality and its causes among Muslims, Druze and Christians. Isr J Med Sci (1979) 15:965–72.[Medline]
4 Gopinath N, Chadha SL, Jain P, et al. An epidemiological study of coronary heart disease in different ethnic groups in Delhi urban populations. J Assoc Physicians India (1995) 43:30–3.[Medline]
5 Jelsma J, Ferguson G. The determinants of self-reported health-related quality of life in a culturally and socially diverse South African community. Bull World Health Organ (2004) 82:206–12.[Medline]
6 Roshi E, Kamberi A, Goda A, Burazeri G. Myocardial infarction and religion: hospital-based case-control study in Tirana, Albania. Croat Med J (2005) 46:977–83.[Medline]
7 Nuri B, Tragakes E, eds. Health care systems in transition: Albania. (2002) (accessed 30 May 2007). Copenhagen: European Observatory on Health Care Systems. Available at: www.who.dk/document/E80089.pdf.
8 Rechel B, McKee M. Healing the crisis: a prescription for public health action in South Eastern Europe (2003) New York: Open Society Institute Press.
9 United Nations Development Program. Albanian human development report 1996. (1996) Tirana: Albania.
10 Institute of Statistics of Albania. Health indicators for years 1994–1998. (2001) Tirana: INSTAT.
11 Institute of Statistics of Albania. Causes of deaths for the year 2000. (2001) Tirana: INSTAT.
12 Burazeri G, Goda A, Sulo G, et al. Conventional risk factors and acute coronary syndrome during a period of transition: a population-based case-control study in Tirana, Albania. In: Croat Med J (2007) (accessed 30 May 2007). 48:225–33. Available at: http://www.cmj.hr/2007/48/2/17436387.htm.[Web of Science][Medline]
13 Burazeri G, Goda A, Tavanxhi N, et al. The health effects of emigration on those who remain at home. Int J Epidemiol (2007) 36:1265–72.
14 Adler N, Epel ES, Castellazzo G, et al. Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy white women. Health Psychol (2000) 19:586–92.[CrossRef][Web of Science][Medline]
15 Singh-Manoux A, Adler NE, Marmot MG. Subjective social status: its determinants and its association with measures of ill-health in the Whitehall II study. Soc Sci Med (2003) 56:1321–33.[CrossRef][Web of Science][Medline]
16 Taylor HL, Jacobs DR, Schucker B, et al. A questionnaire for the assessment of leisure time physical activities. J Chron Dis (1978) 31:741–55.[CrossRef][Web of Science][Medline]
17 Burazeri G, Goda A, Sulo G, et al. Financial loss in pyramid saving schemes, downward social mobility and acute coronary syndrome in transitional Albania. In: J Epidemiol Community Health. (in press; accepted: 1 October, 2007).
18 Hosmer D, Lemeshow S. Applied logistic regression (1989) New York: Wiley & Sons.
19 Bener A, Al-Suwaidi J, Al-Jaber K, et al. The prevalence of hypertension and its associated risk factors in a newly developed country. Saudi Med J (2004) 25:918–22.[Medline]
20 Rechel B, Schwalbe N, Mckee M. Health in south-eastern Europe: a troubled past, an uncertain future. Bull World Health Organ (2004) 82:539–46.[Medline]
21 Marmot M. Status syndrome: a challenge to Medicine. JAMA (2006) 295:1304–7.
22 Winkleby MA, Jatulis DE, Frank E, Fortmann SP. Socioeconomic status and health: how education, income, and occupation contribute to risk factors for cardiovascular disease. Am J Public Health (1992) 82:816–20.
23 Leon D, Chenet L, Shkolnikov VM, et al. Huge variation in Russian mortality rates 1984–1994: artefact, alcohol, or what? Lancet (1997) 350:383–8.[CrossRef][Web of Science][Medline]
24 Britton A, McKee M. The relation between alcohol and cardiovascular disease in eastern Europe: explaining the paradox. J Epidemiol Community Health (2000) 54:328–32.
25 Bobak M, Pikhart H, Rose R, Hertzman C, et al. Socio-economic factors, material inequalities, and perceived control in self-rated health: cross-sectional data from seven post-communist countries. Soc Sci Med (2000) 51:1343–50.[CrossRef][Web of Science][Medline]
26 Bobak M, Pikhart H, Hertzman C, et al. Socio-economic factors, perceived control and self-reported health in Russia. A cross-sectional survey. Soc Sci Med (1998) 47:269–79.[CrossRef][Web of Science][Medline]
27 Murphy M, Bobak M, Nicholson A, et al. The widening gap in mortality by educational level in the Russian Federation, 1980–2001. Am J Public Health (2006) 96:1293–9.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||