Abstract

Background: The aim of this study was to estimate the magnitude of gender differences in hazardous drinking among middle-aged people and to analyse whether these differences are associated with contextual factors, such as public policies or socioeconomic factors. Methods: Cross-sectional design. The study population included 50- to 64-year-old residents of 16 European countries who participated in the Survey of Health, Ageing and Retirement in Europe project conducted in 2010–12 ( n = 26 017). We estimated gender differences in hazardous drinking in each country. To determine whether different social context or women’s empowerment variables were associated with gender differences in hazardous drinking, we fitted multilevel Poisson regression models adjusted for various individual and country-level variables, which yielded prevalence ratios and their 95% confidence intervals (95% CI). Results: Prevalence of hazardous drinking was significantly higher in men than women [30.2% (95% CI: 29.1–31.4%) and 18.6% (95% CI: 17.7–19.4%), respectively] in most countries, although the extent of these differences varied between countries. Among individuals aged 50–64 years in Europe, risk of becoming a hazardous drinker was 1.69 times higher (95% CI: 1.45–1.97) in men, after controlling for individual and country-level variables. We also found that lower values of the gender empowerment measure and higher unemployment rates were associated with higher gender differences in hazardous drinking. Conclusion: Countries with the greatest gender differences in hazardous drinking were those with the most restrictions on women’s behaviour, and the greatest gender inequalities in daily life. Lower gender differences in hazardous drinking seem to be related to higher consumption among women.

Introduction

Worldwide, men consume alcohol more frequently and in larger amounts than women, and consequently have more alcohol-related problems. 1 These universal gender differences in alcohol consumption have led to the suggestion that biological differences, such as differences in alcohol metabolism, may influence how men and women drink. 2 However, the extent of these differences varies between societies, which implies that social and cultural factors play an important role in modulating the magnitude of these differences in different countries. 3 An adapted conceptual framework, 4,5 which shows how social determinants could affect gender differences in alcohol consumption, is shown in Supplementary figure S1 .

Although more than two-thirds of all alcohol-attributed deaths among people aged 20–64 years occur in the 45–64 years group, 6 most studies on alcohol consumption and related problems focus on adolescents and young adults. Moreover, higher mortality has been observed among people with alcohol dependence than among age-matched people in the general population. 7 Alcohol problems may be more entrenched among middle-aged people, after years of established hazardous drinking or dependence. 8 In this respect, few studies have analysed hazardous drinking in middle-aged people and fewer still have studied gender differences in hazardous drinking. A better understanding of the factors that drive these differences would facilitate the development of policies aimed at reducing hazardous drinking that take into account the differential characteristics of both men and women.

In this sense, several country-level factors could have a role in the relationship between gender and alcohol consumption. First, the political tradition of the government and the culture of a country determine which policies are developed (macroeconomic, social, etc). 9 Country-level policies may have an indirect effect on alcohol consumption through their effects on different axes of inequality and intermediate determinants of alcohol consumption (e.g. poverty, social networks, drug prevention and treatment systems, etc.). Moreover, countries can also develop effective policies directly aimed at preventing or reducing alcohol consumption in the general population (i.e. alcohol control policies). These policies include regulation of alcohol prices and taxes, distribution, availability and advertising, among others, 10 and are influenced by the alcohol industry.

Women’s empowerment increases gender equality and may contribute to changes in gender-based norms in a society. Although an increase in gender equality may improve the health of both men and women, it may also lead women to adopt riskier health behaviours that traditionally were more common among men, 11 such as substance use. 12 Regarding alcohol consumption, gender differences in drinking patterns and related problems seem to be narrowing in younger generations in many countries, 13 as drinking has become more common among women and has decreased or stabilized in men. For this reason, smaller gender differences may lead to greater alcohol-related problems among women. 14 In addition to these changing social conditions, the alcohol industry may be an important modulator of alcohol-related norms among men and women. Alcohol advertising may drive changes in gender-based drinking norms and in the social acceptability of drinking among women, as it contributes to a normalization of alcohol and promotes an increase in consumption. 15

Thus, the aims of this study are to estimate the magnitude of gender differences in hazardous drinking among middle-aged people and to analyse whether these differences could be associated to country-level factors such as public policies or socioeconomic factors in people aged 50–64 years from Europe.

Methods

This study uses a cross-sectional design based on data from surveys conducted between 2010 and 2012 as part of the SHARE project (Survey of Health, Ageing and Retirement in Europe). 16 The study population consisted of 26 017 people aged 50–64 years living in 16 European countries (Austria, Belgium, Czech Republic, Denmark, Estonia, France, Germany, Hungary, Italy, the Netherlands, Poland, Portugal, Slovenia, Spain, Sweden and Switzerland). Each country selected its own probability sample, such that while the type and size of samples was different in each country, the data were comparable. 17 Calibrated individual weights, which depended on the household design weight and the respondent’s calibration variables, were established to deal with problems of unit non-response and sample attrition. Response rates varied from 39% in Belgium to 63% in Hungary. The fraction of subjects excluded from the study due to lack of data for any of the individual variables was 5.5%.

The dependent variable, hazardous drinking, was calculated using an adaptation of the Alcohol Use Disorders Identification Test Consumption (AUDIT-C), which has been validated and used elsewhere. 18,19 Its construction was based on three questions: two assessing hazardous drinking in terms of frequency and quantity (‘During the last three months, how often did you drink any alcoholic beverages, like beer, cider, wine, spirits or cocktails?’; ‘In the last three months, on the days you drank, about how many drinks do you have?’), and one assessing binge drinking (‘In the last three months, how often did you have six or more drinks on one occasion?’). Each question had five possible answers, which were ranked from 0 to 4 points. A final score was calculated as the sum of scores from each question. Hazardous drinking was constructed as a dichotomous variable: drinking was considered to be hazardous when the score was ≥5 in men, and ≥4 in women. 20 The psychometric properties of AUDIT-C as a dichotomous hazardous drinking screening tool have been reported and are well established. 21

The main individual independent variable was gender. We also included the following individual variables in the analysis: (i) educational level (none, primary-, secondary- or third-level education), as a proxy of socioeconomic position, (ii) age, (iii) migration status (native or immigrant), (iv) employment status (employed, unemployed or other) and (v) self-perceived health (dichotomous: good, very good or excellent; fair or poor).

In addition, we used the following contextual country-level variables, as these may be associated with gender inequalities in a society or to hazardous drinking, according to previous studies 19,22–25 :

  1. ‘Standardized gross domestic product per capita’ (GDP) for 2008, as a proxy for the country’s standard of living [official data from the Organization for Economic Co-operation and Development Programme (OECD)]. 26

  2. ‘Standardized gender empowerment measure’ (GEM) for 2009, as a proxy for gender equality in the society. This variable comes from official data from the United Nations Development Programme, and measures gender inequality in three basic dimensions of empowerment (economic participation, political participation and power over economic resources). 27

  3. ‘Standardized unemployment rate’ for 2010 (all people aged 55–64 years), official data from the OECD. 28

  4. ‘Standardized degree of alcohol advertising restrictions’ for 2008. This index (continuous variable) is based on data reported by each country on alcohol advertising restrictions in nine media (national and cable television, national and local radio, print media, billboards, cinema, internet and point-of-sale) for different types of alcoholic beverages (beer, wine and spirits), available from the WHO alcohol database. 29 A score from 1 to 4 was assigned to every media and type of alcoholic beverage (1 point: no restrictions; 2 points: voluntary restrictions; 3 points: partial restrictions; 4 points: ban), and the final score was computed as the sum of all scores.

  5. ‘Other standardized alcohol control policies’ for 2006. This index (continuous variable) included data on various alcohol policies and is self-reported by every country (control over alcohol production, distribution and availability of alcoholic beverages, permitted blood alcohol levels and alcohol taxes), available from the Bridging the Gap project. 30

  6. Drinking patterns score’ for 2005, based on the scale of risk of drinking patterns in each country, available from the WHO database. 29 We divided participating countries into two groups, those with low-risk drinking patterns (pattern 1 on the WHO scale) and those with risky drinking patterns (patterns 2 and 3 on the WHO scale).

  7. ‘Percentage of immigrants’, based on self-reported data from the SHARE survey 2010–12 for people aged 50–64 years.

Analysis

The prevalence of hazardous drinking and gender differences therein, plus 95% confidence intervals (95% CIs), were estimated for each country. We computed the Pearson coefficient of linear correlation between country-level variables and the prevalence of hazardous drinking among men and women. To estimate the association between country-level variables and gender differences in hazardous drinking, we fitted multilevel Poisson regression models with robust variance, 31,32 which yielded prevalence ratios (PRs) and their 95% CIs. The intercept and the ‘sex’ coefficient were considered as random effects in all models. The final adjusted model included all individual variables as well as the country-level variables that were statistically significant in the multivariate analysis. Statistical analyses were conducted using STATA 11.0 and HLM6.

Results

Gender differences in hazardous drinking

Table 1 shows the prevalence of hazardous drinking for each individual-level variable, country and sex. The prevalence of hazardous drinking in our sample was significantly higher among men than women [30.2% (95% CI = 29.1–31.4%) vs. 18.6% (95% CI = 17.7–19.4%), respectively]. Gender differences in hazardous drinking were observed in most countries. Figure 1 shows the extent of gender differences in hazardous drinking for each country, which were higher in Eastern European countries. No gender differences in hazardous drinking were observed in Italy, the Netherlands, Sweden and Switzerland, after controlling for other variables.

Figure 1

Scatter plot of prevalence of hazardous drinking among men and women from 50 to 64 years. Adjusted PR of hazardous drinking in men compared to women, by country (50- to 64-year-olds). European countries included in the SHARE project, 2010–12

Figure 1

Scatter plot of prevalence of hazardous drinking among men and women from 50 to 64 years. Adjusted PR of hazardous drinking in men compared to women, by country (50- to 64-year-olds). European countries included in the SHARE project, 2010–12

Table 1

Distribution of the prevalence of hazardous drinking for the independent variables studied, by sex. European countries included in the SHARE project 2010–12

  Men
 
  Women
 
  Total
 
 
 N  % HD a 95% CI N  % HD a 95% CI N  % HD a 95% CI 
Individual-level variables 11 457 30.2 (29.1–31.4) 14 560 18.6 (17.7–19.4) 26 017 24.1 (23.4–24.9) 
Age 
    50–54 years 2751 31.3 (29.1–33.6) 3739 18.7 (17.1–20.5) 6490 24.7 (23.3–26.1) 
    55–59 years 4213 30.5 (28.6–32.6) 5210 18.6 (17.2–20.1) 9423 24.5 (23.3–25.8) 
    60–64 years 4493 29.1 (27.3–31.0) 5611 18.4 (17.1–19.8) 10 104 23.3 (22.2–24.5) 
Educational level 
    None 124 27.4 (18.2–38.9) 229 9.3 (5.6–15.0) 353 16.8 (12.3–22.6) 
    Primary studies 1114 32.4 (28.8–36.1) 1898 15.0 (13.0–17.4) 3012 22.0 (20.0–24.1) 
    Secondary studies 7236 30.1 (28.7–31.6) 8697 17.9 (16.8–19.0) 15 933 24.0 (23.1–24.9) 
    Tertiary studies 2983 29.8 (27.5–32.3) 3736 22.2 (20.4–24.1) 6719 25.8 (24.3–27.3) 
Migration status 
    Native 1046 23.0 (20.1–26.2) 1363 14.6 (12.6–16.9) 2409 18.5 (16.7–20.4) 
    Immigrant 10 411 31.0 (29.7–32.2) 13 197 19.0 (18.1–19.9) 23 608 24.7 (24.0–25.5) 
Employment status 
    Employed 6494 31.1 (29.6–32.7) 6750 20.7 (19.4–22.0) 13 244 26.2 (25.2–27.3) 
    Unemployed 881 35.3 (31.3–39.4) 843 19.1 (15.8–23.0) 1724 28.1 (25.4–31.0) 
    Other 4082 27.3 (25.4–29.2) 6967 16.2 (15.0–17.4) 11 049 20.6 (19.6–21.6) 
Self-perceived health 
    Fair or poor 3711 28.1 (26.3–30.1) 4965 12.5 (11.3–13.7) 8676 19.7 (18.6–20.9) 
    Good, very good or excellent 7746 31.2 (29.8–32.7) 9595 21.7 (20.6–22.8) 17 341 26.3 (25.4–27.3) 
Country 
    Austria 1047 27.6 (24.7–30.7) 1342 18.0 (15.7–20.5) 2389 22.7 (20.8–24.6) 
    Belgium 1206 38.8 (34.8–43.0) 1404 31.8 (28.2–35.6) 2610 35.2 (32.5–38.1) 
    Czech Republic 1208 39.7 (35.7–43.7) 1573 18.5 (15.6–21.7) 2781 28.6 (26.1–31.1) 
    Denmark 522 42.6 (38.0–47.3) 604 36.6 (32.5–40.9) 1126 39.5 (36.4–42.7) 
    Estonia 1231 26.0 (23.4–28.8) 1624 6.1 (5.0–7.4) 2855 14.9 (13.5–16.4) 
    France 1213 32.6 (28.7–36.8) 1436 18.5 (15.9–21.5) 2649 25.3 (22.9–27.9) 
    Germany 223 34.3 (26.5–43.1) 333 16.9 (12.6–22.2) 556 25.4 (20.8–30.7) 
    Hungary 713 22.5 (17.9–27.9) 877 6.0 (4.4–8.1) 1590 13.4 (11.2–15.9) 
    Italy 624 24.6 (19.3–30.7) 869 23.6 (19.2–28.7) 1493 24.1 (20.6–27.9) 
    Netherlands 529 42.2 (36.3–48.3) 739 36.1 (31.8–40.5) 1268 39.1 (35.4–42.9) 
    Poland 273 24.1 (18.5–30.6) 403 7.0 (4.3–11.1) 676 15.1 (11.9–18.9) 
    Portugal 417 30.6 (22.3–40.5) 549 17.5 (12–24.8) 966 23.5 (18.5–29.5) 
    Slovenia 622 16.4 (13.5–19.8) 763 9.5 (7.5–11.9) 1385 13.0 (11.2–15.1) 
    Spain 605 26.0 (21.4–31.1) 755 16.9 (13.7–20.8) 1360 21.3 (18.4–24.5) 
    Sweden 238 14.1 (9.3–20.7) 330 13.2 (9.8–17.7) 568 13.6 (10.6–17.4) 
    Switzerland 786 28.7 (25.4–32.2) 959 27.7 (24.7–30.8) 1745 28.2 (26.0–30.5) 
  Men
 
  Women
 
  Total
 
 
 N  % HD a 95% CI N  % HD a 95% CI N  % HD a 95% CI 
Individual-level variables 11 457 30.2 (29.1–31.4) 14 560 18.6 (17.7–19.4) 26 017 24.1 (23.4–24.9) 
Age 
    50–54 years 2751 31.3 (29.1–33.6) 3739 18.7 (17.1–20.5) 6490 24.7 (23.3–26.1) 
    55–59 years 4213 30.5 (28.6–32.6) 5210 18.6 (17.2–20.1) 9423 24.5 (23.3–25.8) 
    60–64 years 4493 29.1 (27.3–31.0) 5611 18.4 (17.1–19.8) 10 104 23.3 (22.2–24.5) 
Educational level 
    None 124 27.4 (18.2–38.9) 229 9.3 (5.6–15.0) 353 16.8 (12.3–22.6) 
    Primary studies 1114 32.4 (28.8–36.1) 1898 15.0 (13.0–17.4) 3012 22.0 (20.0–24.1) 
    Secondary studies 7236 30.1 (28.7–31.6) 8697 17.9 (16.8–19.0) 15 933 24.0 (23.1–24.9) 
    Tertiary studies 2983 29.8 (27.5–32.3) 3736 22.2 (20.4–24.1) 6719 25.8 (24.3–27.3) 
Migration status 
    Native 1046 23.0 (20.1–26.2) 1363 14.6 (12.6–16.9) 2409 18.5 (16.7–20.4) 
    Immigrant 10 411 31.0 (29.7–32.2) 13 197 19.0 (18.1–19.9) 23 608 24.7 (24.0–25.5) 
Employment status 
    Employed 6494 31.1 (29.6–32.7) 6750 20.7 (19.4–22.0) 13 244 26.2 (25.2–27.3) 
    Unemployed 881 35.3 (31.3–39.4) 843 19.1 (15.8–23.0) 1724 28.1 (25.4–31.0) 
    Other 4082 27.3 (25.4–29.2) 6967 16.2 (15.0–17.4) 11 049 20.6 (19.6–21.6) 
Self-perceived health 
    Fair or poor 3711 28.1 (26.3–30.1) 4965 12.5 (11.3–13.7) 8676 19.7 (18.6–20.9) 
    Good, very good or excellent 7746 31.2 (29.8–32.7) 9595 21.7 (20.6–22.8) 17 341 26.3 (25.4–27.3) 
Country 
    Austria 1047 27.6 (24.7–30.7) 1342 18.0 (15.7–20.5) 2389 22.7 (20.8–24.6) 
    Belgium 1206 38.8 (34.8–43.0) 1404 31.8 (28.2–35.6) 2610 35.2 (32.5–38.1) 
    Czech Republic 1208 39.7 (35.7–43.7) 1573 18.5 (15.6–21.7) 2781 28.6 (26.1–31.1) 
    Denmark 522 42.6 (38.0–47.3) 604 36.6 (32.5–40.9) 1126 39.5 (36.4–42.7) 
    Estonia 1231 26.0 (23.4–28.8) 1624 6.1 (5.0–7.4) 2855 14.9 (13.5–16.4) 
    France 1213 32.6 (28.7–36.8) 1436 18.5 (15.9–21.5) 2649 25.3 (22.9–27.9) 
    Germany 223 34.3 (26.5–43.1) 333 16.9 (12.6–22.2) 556 25.4 (20.8–30.7) 
    Hungary 713 22.5 (17.9–27.9) 877 6.0 (4.4–8.1) 1590 13.4 (11.2–15.9) 
    Italy 624 24.6 (19.3–30.7) 869 23.6 (19.2–28.7) 1493 24.1 (20.6–27.9) 
    Netherlands 529 42.2 (36.3–48.3) 739 36.1 (31.8–40.5) 1268 39.1 (35.4–42.9) 
    Poland 273 24.1 (18.5–30.6) 403 7.0 (4.3–11.1) 676 15.1 (11.9–18.9) 
    Portugal 417 30.6 (22.3–40.5) 549 17.5 (12–24.8) 966 23.5 (18.5–29.5) 
    Slovenia 622 16.4 (13.5–19.8) 763 9.5 (7.5–11.9) 1385 13.0 (11.2–15.1) 
    Spain 605 26.0 (21.4–31.1) 755 16.9 (13.7–20.8) 1360 21.3 (18.4–24.5) 
    Sweden 238 14.1 (9.3–20.7) 330 13.2 (9.8–17.7) 568 13.6 (10.6–17.4) 
    Switzerland 786 28.7 (25.4–32.2) 959 27.7 (24.7–30.8) 1745 28.2 (26.0–30.5) 

a: % HD, prevalence of hazardous drinkers calculated using the AUDIT-C test.

Associations between gender differences in hazardous drinking and contextual factors

As shown in figure 2 , the correlation between hazardous drinking and GDP was statistically significant in women ( r = 0.74; P < 0.05) but not in men ( r = 0.33; P = 0.22). We also observed a significant correlation between GEM and the prevalence of hazardous drinking among women ( r = 0.73; P < 0.05) but not among men ( r = 0.39; P = 0.14). A significant linear correlation between the degree of regulation of alcohol advertising and the prevalence of hazardous drinking was found among men ( r = −0.68; P < 0.05) but not among women ( r = −0.45; P = 0.08). No statistically significant linear correlations were found between unemployment rate, other alcohol control policies or percentage of immigrants and hazardous drinking. GDP and GEM were highly correlated ( r = 0.81; P < 0.05).

Figure 2

Correlation between country-level continuous variables and prevalence of hazardous drinking in men and women aged 50–64 years. Countries from the SHARE Project, 2010–12

Figure 2

Correlation between country-level continuous variables and prevalence of hazardous drinking in men and women aged 50–64 years. Countries from the SHARE Project, 2010–12

Overall, risk of becoming a hazardous drinker in people aged 50–64 years was 1.69 times higher in men than women (95% CI = 1.45–1.97) after controlling for individual and country-level variables ( table 2 ). In the multilevel adjusted analysis, the prevalence of hazardous drinking was influenced by the degree of restriction on alcohol advertising in the country (PR = 0.78; 95% CI = 0.69–0.88). Furthermore, gender differences in hazardous drinking were affected by GEM and unemployment rate. Countries with greater gender equality (i.e. higher GEM values) had lower gender differences in hazardous drinking (PR = 0.27; 95% CI = 0.11–0.65), mainly due to the higher prevalence of hazardous drinking among women, and countries with higher unemployment rates had higher gender differences in hazardous drinking (PR = 1.15; 95% CI = 1.04–1.27). We found no association between gender differences in hazardous drinking and other contextual variables (GDP, other alcohol control policies and percentage of immigrants). Thus, the final model did not include these variables. The final model explained 64% of the variance in gender differences between countries ( table 2 ).

Table 2

PR of hazardous drinking, estimated using multilevel Poisson regression models. A sample of 50- to 64-year-old European residents from 16 countries participating in the SHARE project, 2010–12

  Model 0
 
Model 1
 
Model 2
 
  PR  95% CI  P -value   PR  95% CI  P -value   PR  95% CI  P -value  
Individual variables 
Sex 
    Men  1.682  (1.324–2.137)  <0.001        1.691  (1.454–1.968)  <0.001 
    Women               
Contextual variables (intercept) 
    GDP        1.153  (1.009–1.317)  0.038       
    GEM        2.622  (0.355–19.357)  0.319       
    Unemployment rate        1.005  (0.922–1.096)  0.899       
    Degree of alcohol advertising restrictions        0.804  (0.709–0.913)  0.003  0.778  (0.692–0.875)  <0.001 
    Other alcohol control policies        0.865  (0.742–1.008)  0.062       
    Patterns of drinking score        0.728  (0.527–1.005)  0.053       
    Percentage of immigrants        0.993  (0.973–1.014)  0.509       
Contextual variables (interaction with sex) 
    GDP        0.817  (0.745–0.896)  <0.001       
    GEM        0.271  (0.081–0.906)  0.036  0.267  (0.110–0.651)  0.008 
    Unemployment rate        1.115  (1.000–1.242)  0.049  1.148  (1.040–1.268)  0.010 
    Degree of alcohol advertising restrictions        0.847  (0.746–0.961)  0.014       
    Other alcohol control policies        0.881  (0.773–1.005)  0.057       
    Patterns of drinking score        0.961  (0.709–1.303)  0.785       
    Percentage of immigrants        0.995  (0.972–1.019)  0.670       
Variability 
    Intercept variability (% of change) a  0.310      0.240 (22.5%) 
    Interaction with sex (% of change) a  0.192      0.069 (64.1%) 
  Model 0
 
Model 1
 
Model 2
 
  PR  95% CI  P -value   PR  95% CI  P -value   PR  95% CI  P -value  
Individual variables 
Sex 
    Men  1.682  (1.324–2.137)  <0.001        1.691  (1.454–1.968)  <0.001 
    Women               
Contextual variables (intercept) 
    GDP        1.153  (1.009–1.317)  0.038       
    GEM        2.622  (0.355–19.357)  0.319       
    Unemployment rate        1.005  (0.922–1.096)  0.899       
    Degree of alcohol advertising restrictions        0.804  (0.709–0.913)  0.003  0.778  (0.692–0.875)  <0.001 
    Other alcohol control policies        0.865  (0.742–1.008)  0.062       
    Patterns of drinking score        0.728  (0.527–1.005)  0.053       
    Percentage of immigrants        0.993  (0.973–1.014)  0.509       
Contextual variables (interaction with sex) 
    GDP        0.817  (0.745–0.896)  <0.001       
    GEM        0.271  (0.081–0.906)  0.036  0.267  (0.110–0.651)  0.008 
    Unemployment rate        1.115  (1.000–1.242)  0.049  1.148  (1.040–1.268)  0.010 
    Degree of alcohol advertising restrictions        0.847  (0.746–0.961)  0.014       
    Other alcohol control policies        0.881  (0.773–1.005)  0.057       
    Patterns of drinking score        0.961  (0.709–1.303)  0.785       
    Percentage of immigrants        0.995  (0.972–1.019)  0.670       
Variability 
    Intercept variability (% of change) a  0.310      0.240 (22.5%) 
    Interaction with sex (% of change) a  0.192      0.069 (64.1%) 

a: % of change in variability was calculated using this formula: [(initial variability – final variability)/initial variability] × 100. Model 0: it included hazardous drinking by sex, controlling for all individual variables (age, migration status, educational level, employment status and self-perceived health). Formula: [ forumla *]. Model 1: level 1 was the same as in model 0*, and in level 2 a contextual variable was included separately in each model. Fourteen models were fitted: seven for each contextual factor adjusting the random intercept (formula: [level 2 forumla ]; forumla ), and seven adjusting the random slope for sex (formula: [level 2 forumla ]; forumla )]. Model 2: (final model) it included level 1 mentioned above* and the degree of alcohol advertising restrictions adjusting the variability of intercept among countries ( forumla ) and GEM and the national unemployment rate adjusting the SEX slope coefficient variability among countries ( forumlaforumla ) at level 2; where I  = individuals = 1, … ,26 017; j  = countries = 1, … ,16; X  = individual variables and D contextual variables. Statistically significant variables ( P  < 0.05) are highlighted in bold type.

Discussion

Higher gender equality was associated with lower gender differences in hazardous drinking, mainly due to higher consumption among women. Higher unemployment rates were related to lower prevalence of hazardous drinking in women, and consequently to higher gender differences in hazardous drinking. Restrictions on alcohol advertising have an effect on hazardous drinking, but not by modifying gender differences in alcohol consumption.

Strengths and limitations

The SHARE survey contains questions that allowed us to use an adaptation of the AUDIT-C test 18,19 which has been validated and is widely used to detect hazardous drinking. As far as we are aware, ours is the first study that uses multilevel statistical methods to assess the association between gender differences in hazardous drinking and individual and contextual-level variables, which is a more robust approach than that used by a previous descriptive study of 18- to 34-year-olds. 23 However, our study also has the following limitations. First, we cannot assume that this population sample reflects the distribution of the full range of alcohol consumers in each country. As samples were selected independently in each country, their size differed, and thus, while the prevalence of hazardous drinking in each country comes from a representative sample of its inhabitants, the overall sample is not representative of the whole European population. In addition, response rates were low in some countries, although individual sample weights were used. Data on advertising restrictions were self-reported, and, as alcohol advertising is not systematically monitored in each country, the degree of compliance with regulations could not be accounted for. 33 Gender differences in hazardous drinking among middle-aged people could have been affected to some extent by different mortality rates among younger hazardous drinkers in each sex in some countries, although we were unable to test this hypothesis. The study consisted of only 16 samples in the second level of our analyses, which may limit our ability to detect statistically significant differences between groups.

Gender differences in hazardous drinking

In line with other studies, 1,2 we found that hazardous drinking was more common among men in almost all participating countries. A moderate to high prevalence of hazardous drinking was observed among men in most countries, while the prevalence in women was more variable (e.g. up to six times higher in some countries than others). Consequently, the extent of gender differences in hazardous drinking varied widely among countries. These country-level differences could be associated with some contextual factors, such as the degree of gender equality in a country or the national unemployment rate, as discussed below.

Associations between gender differences in hazardous drinking and contextual factors

Gender inequalities in a society

Our findings suggest a strong association between GEM and gender differences in hazardous drinking. Greater gender equality in a country appears to be related to higher alcohol consumption in women. Similar results for GEM were found in people aged 18–34 years. 23 In fact, gender differences in alcohol consumption may be related to the degree of women’s empowerment, and to the drinking culture in each country, especially among women. 1 Gender differences in alcohol consumption may be more acute in societies in which gender roles are more polarized (e.g. societies where women are largely confined to the home and to domestic work, while their husbands spend a lot of time working outside the home). 3,34 Traditionally, drinking has been more socially restricted among women than men worldwide, due to the belief that women’s social behaviour and responsibilities may be more adversely affected by drinking. However, relatively high gender equality is expected in societies where drinking is well integrated into social life. 3 These hypotheses are consistent with our results. However, the relationship between prevalence of hazardous drinking and GEM was stronger in women than in men, which suggests that improvement in women’s social conditions might be accompanied by the adoption of risky and unhealthy lifestyles that were traditionally more common in men. 14

Although we observed a strong correlation between GEM and GDP, this does not necessarily mean that women’s empowerment depends only on national wealth, as suggested by Varkey et al. 35 In fact, our results indicate that GEM had a greater influence than GDP on gender differences in hazardous drinking.

The interaction between drinking culture and women’s position in society may also play an important role in the extent of gender differences in hazardous drinking, although we were not able to investigate this. Differences in drinking culture and their interaction with GEM may partly explain the differences in the prevalence and the gender differences in alcohol consumption and hazardous drinking between countries. Moreover, the interests of the alcohol industry could play a role in the relationship between GEM and the drinking culture. A society in which the economic and social position of women is increasing provides a new potential source of customers for the industry. Advertising may reflect the power of industry by allowing it to influence the values of society. However, we have only analysed the degree of advertising restrictions because data on the advertising content were not available. These data may have shown that advertising campaigns that aim to normalize drinking among women could have produced an increase in consumption.

Unemployment rate

In recent years, Europe has been affected by a serious economic recession, in which unemployment increased rapidly and substantially. Economic difficulties have been reported to affect health and health behaviours. 36 In general, alcohol consumption decreases when economic resources decline, 37 but risky-drinking patterns, such as binge drinking, may increase at the same time (i.e. people drink less frequently but drink more on each occasion). 38 In our study, we found an association between higher national unemployment rates and lower prevalence of hazardous drinking among women aged 50–64 years, but not in men. The unemployment rate could partly explain gender differences in hazardous drinking. However, the relationship between unemployment rate and hazardous drinking is complex, 25 and further studies are needed to corroborate the observed association and to verify if national unemployment rates affect men and women differently, as our results suggest.

Other contextual factors

Preventive policies (e.g. alcohol advertising regulations) may help to limit alcohol consumption. Indeed, we found that alcohol advertising regulations are associated to lower prevalence of hazardous drinking in both men and women. But there was not an association with gender differences in hazardous drinking, after adjustment for various country-level variables. A previous study in adolescents hypothesized that the effects of alcohol advertising on girls were greater than on boys. 24 However, the results of our study are not consistent with this hypothesis in middle-aged people. One possible explanation is that women who were not used to consuming alcohol as young adults do not increase their alcohol consumption when they got older, especially those living in Eastern European countries. 39 For this reason, the association between the degree of alcohol advertising restrictions and gender differences could be masked in older adults by the fact that the effects of advertising regulations during adolescence may last into adulthood. 40 Although no statistically significant associations were observed for the other contextual variables analysed, our results could also be explained by other factors that could not be included in the analyses, such as differences between countries in risk behaviour linked to period or cohort effects.

Conclusions

Countries with highest relative gender differences in hazardous drinking were those with the greatest restrictions on women’s behaviour or the greatest gender inequality. For this reason, policies aimed at improving gender equality should be accompanied by specific alcohol control policies to reduce the probability of engaging in alcohol consumption, especially policies with proven effectiveness (e.g. alcohol taxation, reduction of alcohol availability or alcohol advertising regulations). Otherwise gender differences would be reduced but the adverse effects attributable to alcohol consumption would increase in women.

Supplementary data

Supplementary data are available at EURPUB online.

Funding

This work was supported by the Spanish Network on Addictive Disorders [grant numbers RD06/0001/1018 and RD12/0028/0018]; a Health Strategic Action grant [grant number PI13/00183]; and the SOPHIE project, ‘Evaluating the impact of structural policies on health inequalities and their social determinants and fostering change’ (FP7/2007-2013) [grant number 278173].

Conflicts of interest : None declared.

Key points

  • This study suggests that greater gender equality is associated with lower extent of gender differences in alcohol consumption, mainly due to an increase in women’s consumption.

  • National unemployment rate seems to be related to the prevalence of hazardous drinking in women, but not in men.

  • Countries with higher alcohol advertising regulations have lower prevalence of hazardous drinking, but these results do not seem to affect differently men and women.

Acknowledgements

This article uses data from SHARE wave 4, release 1.1.1, as of 28 March 2013. The SHARE data collection has primarily been funded by the European Commission through the 5th Framework Program (project QLK6-CT-2001-00360 in the Quality of Life thematic program), through the 6th Framework Program (projects SHARE-I3, RII-CT-2006-062193, COMPARE, CIT5-CT-2005-028857 and SHARELIFE, CIT4-CT-2006-028812) and through the 7th Framework Program (SHARE-PREP, No 211909, SHARE-LEAP, No 227822 and SHARE M4, No 261982). Additional funding from the US National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, R21 AG025169, Y1-AG-4553-01, IAG BSR06-11 and OGHA 04-064) and the German Ministry of Education and Research, as well as from various national sources is gratefully acknowledged (see www.share-project.org for a full list of funding institutions). This article forms part of the doctoral dissertation of Marina Bosque-Prous, at the Universitat Pompeu Fabra.

References

1
Mäkelä
P
Gmel
G
Grittner
U
, et al.  . 
Drinking patterns and their gender differences in Europe
Alcohol Alcohol Suppl
 , 
2006
, vol. 
41
 (pg. 
i8
-
18
)
2
Wilsnack
RW
Wilsnack
SC
Kristjanson
AF
, et al.  . 
Gender and alcohol consumption: patterns from the multinational GENACIS project
Addiction
 , 
2009
, vol. 
104
 (pg. 
1487
-
500
)
3
Ahlström
S
Bloomfield
K
Knibbe
R
Gender differences in drinking patterns in nine European countries: descriptive findings
Subst Abus
 , 
2001
, vol. 
22
 (pg. 
69
-
85
)
4
Solar
O
Irwin
A
A Conceptual Framework for Action on the Social Determinants of Health
 , 
2007
Geneva: WHO
Commission on Social Determinants of Health
5
Comisión Para Reducir Las Desigualdades Sociales En Salud En España
[A proposal of policies and interventions to reduce social inequalities in health in Spain. Commission to Reduce Social Inequalities in Health in Spain]
Gac Sanit
 , 
2012
, vol. 
26
 (pg. 
182
-
9
)
6
Rehm
J
Zatonksi
W
Taylor
B
Anderson
P
Epidemiology and alcohol policy in Europe
Addiction
 , 
2011
, vol. 
106
 (pg. 
11
-
9
)
7
Guitart
AM
Espelt
A
Castellano
Y
, et al.  . 
[Impact of alcohol use disorder on mortality: are there age and gender differences?]
Gac Sanit
 , 
2011
, vol. 
25
 (pg. 
385
-
90
)
8
Ferreira
MP
Weems
MK
Alcohol consumption by aging adults in the United States: health benefits and detriments
J Am Diet Assoc
 , 
2008
, vol. 
108
 (pg. 
1668
-
76
)
9
Huber
E
Ragin
C
Stephens
JD
Social democracy, Christian democracy, constitutional structure, and the welfare state
Am J Sociol
 , 
1993
, vol. 
99
 (pg. 
711
-
49
)
10
Anderson
P
Chisholm
D
Fuhr
DC
Effectiveness and cost-effectiveness of policies and programmes to reduce the harm caused by alcohol
Lancet
 , 
2009
, vol. 
373
 (pg. 
2234
-
46
)
11
Sen
G
Östlin
P
George
A
 
Unequal, Unfair, Ineffective and Inefficient. Gender Inequity in Health: Why It Exists and How We Can Change It. Final Report to the WHO Commission on Social Determinants of Health. Women and Gender Equity Knowledge Network. Geneva: World Health Organization, 2007
12
Sánchez-López
MdelP
Cuellar-Flores
I
Dresch
V
The impact of gender roles on health
Women Health
 , 
2012
, vol. 
52
 (pg. 
182
-
96
)
13
Keyes
KM
Li
G
Hasin
DS
Birth cohort effects and gender differences in alcohol epidemiology: a review and synthesis
Alcohol Clin Exp Res
 , 
2011
, vol. 
35
 (pg. 
2101
-
12
)
14
Bond
JC
Roberts
SCM
Greenfield
TK
, et al.  . 
Gender differences in public and private drinking contexts: a multi-level GENACIS analysis
Int J Environ Res Public Health
 , 
2010
, vol. 
7
 (pg. 
2136
-
60
)
15
Anderson
P
de Bruijn
A
Angus
K
, et al.  . 
Impact of alcohol advertising and media exposure on adolescent alcohol use: a systematic review of longitudinal studies
Alcohol Alcohol
 , 
2009
, vol. 
44
 (pg. 
229
-
43
)
16
Börsch-Supan
A
Brandt
M
Hunkler
C
, et al.  . 
Data resource profile: the survey of health, ageing and retirement in Europe (SHARE)
Int J Epidemiol
 , 
2013
, vol. 
42
 (pg. 
992
-
1001
)
17
Malter
F
Börsch-Supan
A
 
SHARE Wave 4: Innovations & Methodology . Munich: MEA, Max Planck Institute for Social Law and Social Policy, 2013
18
Bush
K
Kivlahan
DR
McDonell
MB
, et al.  . 
The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test
Arch Intern Med
 , 
1998
, vol. 
158
 (pg. 
1789
-
95
)
19
Bosque-Prous
M
Espelt
A
Guitart
AM
, et al.  . 
Association between stricter alcohol advertising regulations and lower hazardous drinking across European countries
Addiction
 , 
2014
, vol. 
109
 (pg. 
1634
-
43
)
20
Gual
A
Segura
L
Contel
M
, et al.  . 
Audit-3 and audit-4: effectiveness of two short forms of the alcohol use disorders identification test
Alcohol Alcohol
 , 
2002
, vol. 
37
 (pg. 
591
-
6
)
21
Meneses-Gaya
C
Zuardi
AW
Loureiro
SR
, et al.  . 
Is the full version of the AUDIT really necessary? Study of the validity and internal construct of its abbreviated versions
Alcohol Clin Exp Res
 , 
2010
, vol. 
34
 (pg. 
1417
-
24
)
22
Grittner
U
Kuntsche
S
Graham
K
Bloomfield
K
Social inequalities and gender differences in the experience of alcohol-related problems
Alcohol Alcohol
 , 
2012
, vol. 
47
 (pg. 
597
-
605
)
23
Rahav
G
Wilsnack
R
Bloomfield
K
, et al.  . 
The influence of societal level factors on men’s and women’s alcohol consumption and alcohol problems
Alcohol Alcohol Suppl
 , 
2006
, vol. 
41
 (pg. 
i47
-
55
)
24
Saffer
H
Dave
D
Alcohol advertising and alcohol consumption by adolescents
Health Econ
 , 
2006
, vol. 
15
 (pg. 
617
-
37
)
25
Henkel
D
Unemployment and substance use: a review of the literature (1990-2010)
Curr Drug Abuse Rev
 , 
2011
, vol. 
4
 (pg. 
4
-
27
)
26
 
Organisation for Economic Co-operation and Development. GDP per head, US $, constant prices, constant PPPs. Available at: http://stats.oecd.org/index.aspx?queryid=559 (5 October 2014, date last accessed)
27
Klugman
J
Rodríguez
F
Kennedy
A
, et al.  . 
Overcoming Barriers: Human Mobility and Development
 , 
2009
New York
United Nations Development Programme
28
 
Organisation for Economic Co-operation and Development. Unemployment rate from 55 to 64 years old. Available at: http://stats.oecd.org/Index.aspx?DataSetCode=LFS_SEXAGE_I_R (5 October 2014, date last accessed)
29
 
World Health Organization. WHO Global Information System on Alcohol and Health. Available at: http://apps.who.int/gho/data/?showonly=GISAH&theme=main (5 October 2014, date last accessed)
30
Karlsson
T
Österberg
E
Scaling alcohol control policies across Europe
Drugs Educ Prev Policy
 , 
2007
, vol. 
14
 (pg. 
499
-
511
)
31
Barros
AJD
Hirakata
VN
Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio
BMC Med Res Methodol
 , 
2003
, vol. 
3
 pg. 
21
 
32
Gelman
A
Hill
J
Data Analysis Using Regression and Multilevel/hierarchical Models
 , 
2006
New York
Cambridge University Press
33
De Bruijn
A
van den Wildenberg
E
van den Broeck
A
Commercial Promotion of Drinking in Europe. Key Findings of Independent Monitoring of Alcohol Marketing in Five European Countries
 , 
2012
Utrecht
European Centre for Monitoring Alcohol Marketing
34
Wilsnack
RW
Vogeltanz
ND
Wilsnack
SC
, et al.  . 
Gender differences in alcohol consumption and adverse drinking consequences: cross-cultural patterns
Addiction
 , 
2000
, vol. 
95
 (pg. 
251
-
65
)
35
Varkey
P
Mbbs
Kureshi
S
Lesnick
T
Empowerment of women and its association with the health of the community
J Womens Health 2002
 , 
2010
, vol. 
19
 (pg. 
71
-
6
)
36
Marmot
MG
Bell
R
How will the financial crisis affect health?
BMJ
 , 
2009
 
338:b1314
37
Ruhm
CJ
Black
WE
Does drinking really decrease in bad times?
J Health Econ
 , 
2002
, vol. 
21
 (pg. 
659
-
78
)
38
Dávalos
ME
Fang
H
French
MT
Easing the pain of an economic downturn: macroeconomic conditions and excessive alcohol consumption
Health Econ
 , 
2012
, vol. 
21
 (pg. 
1318
-
35
)
39
Ahlström
S
Alcohol use and problems among older women and men: a review
Nord Stud Alcohol Drugs
 , 
2008
, vol. 
25
 (pg. 
154
-
61
)
40
Snyder
LB
Milici
FF
Slater
M
, et al.  . 
Effects of alcohol advertising exposure on drinking among youth
Arch Pediatr Adolesc Med
 , 
2006
, vol. 
160
 (pg. 
18
-
24
)

Comments

0 Comments