OUP user menu

EPIDEMIOLOGY OF HEAVY ALCOHOL USE IN UKRAINE: FINDINGS FROM THE WORLD MENTAL HEALTH SURVEY

CHARLES P. M. WEBB, EVELYN J. BROMET, SEMYON GLUZMAN, NATHAN L. TINTLE, JOSEPH E. SCHWARTZ, STANISLAV KOSTYUCHENKO, JOHAN M. HAVENAAR
DOI: http://dx.doi.org/10.1093/alcalc/agh152 327-335 First published online: 11 April 2005

Abstract

Aims: To describe the epidemiology of heavy alcohol use in Ukraine, using data from the world mental health (WMH) survey in Ukraine. Methods: The WMH composite international diagnostic interview was administered in 2002 to a national probability sample of Ukrainian adults (n = 4725). An algorithm for classifying heavy use in the past year was developed from self-reports about the quantity and frequency of drinking, and its convergent validity was demonstrated. Prevalence rates and socio-demographic risk factors were examined separately for men and women. Results: The 12-month rates of heavy alcohol use were 38.7% in men and 8.5% in women (22.0% overall). Among heavy alcohol users, 92% of men and 52% of women consumed at least 80 g of ethanol in a typical drinking day on a monthly basis in the year before the interview. The most significant risk factors in men and women were age (26–54 years for men; 18–25 years for women), living in the Southeast region, being in the labour force whether employed or unemployed, and for men, low education and being the father of a young child. A highly significant linear relationship of number of risk factors with heavy alcohol use was found for both sexes. Conclusions: The rates for men were similar to those reported in a Russian national survey with the exception of Southeast Ukraine where the rate was >10% higher. The highest rates were among men who were middle-aged, fathers and unemployed. Future prospective studies are needed to assess the impact of heavy alcohol use on Ukrainian health, mental health and occupational and social functioning.

(Received 30 September 2004; first review notified 10 November 2004; in revised form 2 February 2005; accepted 8 March 2005)

INTRODUCTION

In the decade following 1991, when Ukraine became an independent country, its population declined by >6%, to 48 457 100 (Kuzio, 2003). This change is partially attributable to an increase in the crude death rate from 11.6/1000 in 1989 (Steshenko, 1997) to 15.4/1000 in 2000 (Eurostat, 2002) and to a corresponding decline in life expectancy by 4.5 years for men and 2.3 years for women (World Health Organization, 2003). The increased death rate has disproportionately occurred in men of pre-retirement age (Steshenko, 1997). Similar changes in mortality have occurred in Russia where the leading causes of death among pre-elderly men include cardiovascular disease, accidents, alcohol poisoning and cirrhosis of the liver. Heavy alcohol use is one of the key risk factors associated with such premature mortality (Nemtsov, 1998, Nemtsov, 1999; Chenet et al., 1998, Chenet et al., 2001; Malyutina et al., 2002). As part of the recently conducted Ukraine-world mental health (WMH) survey (Bromet et al., 2004), we assessed the quantity and frequency of alcohol use in a national probability sample. This paper focuses on the prevalence and socio-demographic correlates of heavy alcohol use.

Information about rates of heavy drinking in former Soviet Union and Eastern bloc countries is available from government data on per capita consumption and several national and municipal surveys (Table 1). The rates of monthly heavy alcohol use range from 12% in Krakow (Bobak et al., 2004) to 82% in Udmurtia (Pakriev et al., 1998), and rates of daily heavy use range from 3% in Lithuania (McKee et al., 2000) to 18% in Bulgaria (Balabanova and McKee 1999), reflecting differences in the definitions of heavy consumption, the regions surveyed and the sampling methods employed.

View this table:
Table 1.

Alcohol use surveys in former Soviet Union and Eastern Bloc Countries

CitationSiteStudynAge (years)YearResponse rate (%)Findings
National surveys
    Cockerham, 2000RussiaRLMSb840216–102199580.0Frequent users were males, age 30–39 years
    Bobak et al., 1999RussiaNRBSa159918–65+199666.2Monthly binge: 31% men; 5% women
    Balabanova and McKee, 1999Bulgaria155018+199796.9Daily binge: 18.2% men; 0.8% women
    McKee et al., 2000BalticsEstonia: 201019–641997Estonia: 67.3%Daily binge: 2–9% men; 0.2–0.5% women
Latvia: 2258Latvia: 77.7%
Lithuania: 2139Lithuania: 74.1%
Municipal surveys
    Stack et al., 1994Moscow, RussiaMoscow Oblast survey37416+199268
    Malyutina et al., 2001Novosibirsk, RussiaWHOWave 1: 166125–64Wave 1: 1985–198671.0–73.01994–1995: Monthly binge: 50% men; 5% women.
MONICAcWave 2: 1700Wave 2: 1988–1989
Wave 3: 3050Wave 3: 1994–1995
    Pakriev et al., 1998Udmurtia, Russia85518–65199585.9Monthly binge: 82% men; 11% women.
    Carlson, 2001Taganrog, RussiaTaganrog Household Wave survey219118–70+Wave 1: 1993–1994Wave 1: 91.01998: 160 grams pure ethanol/week: 14% men; 1% women.
Wave 2: 1998Wave 2: 81.0
    Bobak et al., 2004Novosibirsk, Russia Krakow, Poland Karvina-Havirov, Czech RepublicHAPIEEdNovosibirsk: 99045–641999–2000Novosibirsk: 70%Monthly binge: Males: Novosibirsk: 30% Krakow: 2% Karvina-Havirov: 17%
Krakow: 576Krakow: 65%
Karvina-Havirov: 677Koruina–Hauirou: 71%Females: Novosibirsk: 1% Krakow: 2% Karvina-Havirov: 4%
  • a New Russian barometer survey VI.

  • b Russian longitudinal monitoring study.

  • c Multinational monitoring of trends and determinants in cardiovascular disease.

  • d Health, alcohol and psychosocial factors in Eastern Europe.

The correlates of binge or heavy drinking, based on American and Canadian studies, include being male, young (i.e. age 21–25 years), unmarried, lower educational attainment and unemployment (Barnes et al., 1991; U.S. Department of Health and Human Services, 2002; Naimi et al., 2003). In European studies, the correlates are similar except that heavy drinking in men is more prevalent in middle age (30s to 50s) (Bloomfield, 1998; Meyer et al., 2000; Plant and Plant, 2001; Mateos et al., 2002; Schroder et al., 2004). Although the North American and European findings on age at risk differ, similar findings have been reported on the protective effect of parenting status in the US and the Netherlands (Labouvie, 1996; Hajema et al., 1998). Available surveys from the former Soviet Union have consistently found that being male is a risk factor for heavy alcohol use, particularly in the 25–55 age group (Bobak et al., 1999; Cockerham, 2000). None of the studies in the former Soviet Union was conducted in Ukraine where the recent rates of smoking and illicit substance abuse, particularly in the Southeast region, are among the highest in the former Soviet republics (Dehne et al., 2000; Gilmore et al., 2001; Poznyak et al., 2002).

Our study is the first epidemiologic study of alcohol use in Ukraine to be based on a national probability sample. Consistent with the mandate of the World Health Organization (WHO) WMH Consortium, the main interview tool was the WMH-composite international diagnostic interview (WMH-CIDI). This paper focuses on heavy alcohol use determined from questions on frequency of drinking days and the amount typically consumed. We focused on heavy alcohol use, rather than alcoholism per se, for two reasons. First, our goal was to compare the rates of heavy alcohol use in Ukraine with rates reported for other countries in the former Soviet Union. Second, heavy use is widespread in the former Soviet countries and has been linked to illness and premature death. As such, its public health significance extends beyond that of alcohol disorders in the stricter sense. In this paper, the prevalence and correlates of heavy alcohol use are presented separately for men and women.

SUBJECTS AND METHODS

Sample and procedures

In 2002, we conducted a national survey of mental illness, substance disorders and health in Ukraine (Bromet et al., 2004) as part of the WMH initiative (The WHO World Mental Health Survey Consortium, 2004). Ukraine (population ∼48 million) is the second largest country in Eastern Europe (after Russia). Seventy-five percent of the population is ethnic Ukrainian, and 22% is Russian.

The Ukraine-WMH survey is based on a nationally representative sample of residents aged 18 and older from the country's 24 oblasts (counties) and the autonomous republic of Crimea. The sampling design had four stages: In the first stage, 170 primary sampling units (PSUs) were selected from the cities, towns and villages with probability proportional to size. The PSUs were drawn such that each oblast (county), and the urban and rural populations in each oblast, were represented proportionally. Second, within the PSUs, postal districts were randomly selected. Third, within each postal district, streets were randomly selected; then buildings within streets; and then, apartments within buildings. Fourth, people 18 years and older were randomly selected within apartments. The response rate was 78.3%. Study participants, compared with those who were unlocated or refusers, tended to be female (61.2% vs 49.5%) and older (48.3 vs 42.9 years).

Face-to-face interviews were carried out with 4725 respondents by the professional interview field staff of the Kiev International Institute of Sociology (KIIS) in collaboration with the Ukrainian Psychiatric Association (UPA). The paper-pencil version of the WMH-CIDI, a fully structured, modularized mental health interview schedule designed for lay interviewers, was administered. The interviewers were trained by a certified CIDI trainer over a period of 1 week.

The recruitment, consent and field procedures were approved by the Human Subjects Committees of University at Stony Brook, KIIS and UPA. Standard forward and back translation procedures were used to translate the instrument into Russian and Ukrainian languages.

Definition of heavy alcohol use

Questions about alcohol consumption in the WMH-CIDI's substance use module addressed the frequency of drinking and the number of grams of liquor (including domestically produced bootleg), wine and beer consumed over the past year on a typical drinking day. Variability was not ascertained. Quantity data were converted to grams of ethanol using suggested equivalents: 0.4 × g of liquor; 0.12 × g of wine; and 0.033 × g of beer (Treml, 1975). For men, heavy alcohol use was defined as consuming ≥80 g of ethanol in a typical drinking day) or consuming either ≥60 g 3–4 days/week or ≥40 g nearly every day. In order to adjust for gender differences in metabolism and body weight (Wechsler and Nelson, 2001), the dose criteria was reduced by 25% for women. Thus, for women, heavy alcohol use was defined as at least monthly consumption of ≥60 g of ethanol in a typical drinking day or consuming either ≥45 g 3–4 days/week or ≥30 g nearly every day. Non-heavy users included ‘lifetime’ abstainers (Table 2).

View this table:
Table 2.

Comparison of heavy alcohol users and non-heavy users on pattern of drinking and selected validation criteria

MalesaFemalesb
Pattern of drinkingHeavy alcohol usersNon-heavy usersHeavy alcohol usersNon-heavy users
Modal frequency1–2 days/week1–3 days/month1–3 days/month1–3 days/month
Median ethanol consumed on typical drinking day (g)120.040.080.024.0
Inter-quartile range (g) for consumption on a typical drinking day77.036.532.023.5
%%χ2%%χ2
Convergent validity
    Lifetime alcoholismc43.615.971.5**16.61.630.2**
    Past-year alcoholismc23.14.361.7**11.20.322.9**
Childhood risk factors
    Drinking < age 1871.861.713.8**64.837.854.7**
    Childhood conduct disorder symptoms37.327.78.6*21.69.212.3*
Behavioral correlates
    Lifetime aggression53.835.640.7**48.324.128.9**
    Current-smoker71.948.995.0**32.69.933.8**
  • a For heavy alcohol users, n = 810. For non-heavy users (because of the structure of the WMH-CIDI), n = 678 for pattern of drinking variables and 1283 for validation criteria variables.

  • b For heavy alcohol users, n = 219. For non-heavy users, n = 610 for pattern of drinking variables and 2359 for validation criteria variables.

  • c DSM-IV criteria for alcohol abuse or alcohol dependence.

  • * P < 0.01;

  • ** P < 0.001.

Demographic and geographic risk factors

Nine socio-demographic variables from the WMH-CIDI were examined: age (categorized as 18–25, 26–34, 35–54, ≥55 years); region (West, North-central including Kiev, Southeast); urbanicity (rural, semi-urban with towns of <200 000 people, urban with cities having >200 000 people); education [primary, secondary, specialized secondary, higher (Shkolnikov et al., 1998)]; financial status (a Ukrainian variable classified as ‘very inadequate’ if there was not enough money for food; ‘inadequate’ if not enough money for clothing; ‘adequate’ if enough money for durables); employment status [out of the labor force (79.0% retired, 9.2% homemakers, 8.5% students and 3.3% disabled), unemployed, employed]; marital status (never married, married before, currently married) and parental status (parent of a child who is under the age of 18 and living at home vs others).

Data analysis

Comparison of the unweighted distribution of the sample with the 2001 population census of Ukraine showed that the sample over-represented women, people >55 years of age, and those living in semi-urban settings. These biases were corrected by weighting the sample to the approximate gender, age, urbanicity and regional distributions of the 2001 census. Analyses were conducted using SUDAAN (2003). All analyses were conducted separately for men and women. Logistic regression was used to explore the relationships of the risk factors to heavy alcohol use. Tests for trend (Wald F statistic with 1 degree of freedom for the numerator) were performed for age group, urbanicity, education and financial status. We first examined each risk factor separately, and then did a multivariate analysis, where age was kept in the model irrespective of significance, while other risk factors were eliminated in a stepwise fashion in order of least significance (modified backwards elimination). The explanatory power of the model was calculated as: Math Lastly, we summed the risk factors that were significant in the multivariate model to examine its relationship to heavy alcohol use. Statistical significance was set at P < 0.05.

RESULTS

Demographic characteristics are presented in Table 3. The sample contained relatively more women than men, particularly those ≥55 years. There was also a high rate of poverty, with approximately three-quarters of respondents unable to meet basic needs and only 50% employed.

View this table:
Table 3.

Weighted frequencies of demographic and geographic risk factors

Populationa %TotalMalesFemales
Risk factorsn%n%n%
Genderb
    Male45.0212545.0
    Female55.0260055.0
Ageb(years)
    18–2515.171515.137317.534313.2
    26–3415.571415.135416.736013.8
    35–5436.6175937.283139.192835.7
    55+32.7153732.556826.796937.3
Regionc
    West24.6116824.751824.465025.0
    North-central33.2157433.369432.688033.9
    Southeast42.2198342.091443.0106941.1
Urbanicityc
    Rural32.2152132.268632.383432.1
    Semi-urban32.7164934.975235.489734.5
    Urban35.1155532.968732.386833.4
Educationc
    Primary6.64619.81316.133112.7
    Secondary43.1217946.1108150.9109842.2
    Specialized secondary32.0128127.155326.172828.0
    Higher18.280217.035916.944317.1
Financial statusc
    Very inadequate23.9143630.849323.694336.6
    Inadequate47.5238451.1112954.0125548.7
    Adequate27.684818.247022.434814.7
Employment statusc
    Out of the labor force42.0178937.960828.7118245.5
    Unemployed12.657012.131514.82569.8
    Employed44.2236150.0119956.5116244.7
Marital statusc
    Never married14.272815.445021.227810.7
    Married before25.6117224.830214.287033.5
    Married60.1282559.8137364.6145255.8
Parental status
    Not parent of child329369.9153172.4176267.8
    Parent of child142130.158327.683732.2
  • a Total population = 48 457 100 based on 2001 Ukrainian census.

  • b From 2001 Ukrainian census.

  • c From KIIS.

Prevalence

Patterns of use for heavy and non-heavy alcohol users are presented in Table 2. We found that 1.0% of men and 4.7% of women were lifetime abstainers. To examine convergent validity, we compared heavy users with non-heavy users on lifetime and past-year DSM-IV alcohol abuse and dependence (Table 2). As expected, heavy alcohol users had significantly higher rates. We next approached validity by testing the differences between the groups on established childhood risk factors and behavioral correlates (Pulkkinen and Pitkanen, 1994; Dewit et al., 2000; Wells et al., 2000; Warner and White, 2003) assessed in the WMH-CIDI. As Table 2 shows, the measure of heavy alcohol use distinguished between groups on all measures.

The 12-month prevalence of heavy alcohol use was 38.7% for men and 8.5% for women (OR = 6.8: 95% CI = 5.7–8.2, p < 0.001). The overall rate was 22.0%. Figure 1 shows that the rates were the highest in men aged 26–34 years (49.7%) and 35–54 years (45.5%). In women, the highest rate occurred in the 18–25 age group (16.2%).

Fig. 1.

Relationship of gender and age to heavy alcohol use in Ukraine.

Ninety-two percent of the male heavy alcohol users consumed ≥80 g ethanol in 1 day at least once per month in the past year. In contrast, 50.4% of female heavy alcohol users consumed this amount at least monthly.

Risk factors

The associations of the socio-demographic variables with heavy alcohol use are presented in Tables 4 and 5. The univariate ORs for men (Table 4) showed that the odds of heavy alcohol use were significantly higher in those who were in the age group of 26–34 years and 35–54 years, lived in the Southeast (compared with the West), had a secondary education (i.e. high school), were employed or unemployed (vs out of the labor force), and were the parent of a child under the age of 18 living at home. The final multivariate model showed that heavy alcohol users were more likely to be in the age group of 26–54 years, living in the Southeast, employed or unemployed (vs out of the labor force) and the father of a young child, but the explanatory power was 5.2%.

View this table:
Table 4.

Prevalence of heavy alcohol use by demographic and geographic risk factors in men: weighted proportions

UnadjustedAdjusteda
Risk factors%OR95% CIPOR95% CIP
Age (years)<0.001<0.05
    18–2532.71.01.0
    26–3449.72.0(1.4–2.9)***1.7(1.1–2.4)**
    35–5445.51.7(1.3–2.2)***1.4(1.1–2.0)*
    55+25.80.7(0.5–1.0)*1.0(0.7–1.5)
    F for trend9.3**0.3
Region<0.05<0.05
    West36.01.01.0
    North-central34.50.9(0.7–1.3)1.0(0.7–1.4)
    Southeast43.51.4(1.0–1.8)*1.4(1.1–1.9)*
Urbanicityn.s.a
    Rural41.41.0
    Semi-urban39.70.9(0.6–1.4)
    Urban34.90.8(0.5–1.2)
    F for trend1.8
Education<0.01<0.05
    Primary27.51.01.0
    Secondary43.32.0(1.2–3.3)**1.0(0.6–1.7)
    Specialized secondary37.91.6(1.0–2.7)0.8(0.4–1.3)
    Higher30.41.2(0.7–1.2)0.6(0.3–1.1)
    F for trend4.19.9**
Financial statusn.s.a
    Very inadequate40.41.0
    Inadequate39.41.0(0.8–1.2)
    Adequate36.90.9(0.6–1.2)
    F for trend0.9
Employment status<0.001<0.05
    Out of the labor force25.21.01.0
    Unemployed47.02.6(1.9–3.7)***1.9(1.2–3.1)**
    Employed46.52.3(1.7–3.1)***1.7(1.0–2.7)*
Marital statusn.s.a
    Never married35.61.0
    Married before42.41.3(0.9–2.0)
    Married38.91.2(0.9–1.6)
Parental status<0.001<0.05
    Not parent of a child33.81.01.0
    Parent of child50.72.0(1.5–2.6)***1.5(1.1–2.0)*
  • a Non-significant risk factors were eliminated from the adjusted model in order of least significance.

  • * P < 0.05;

  • ** P < 0.01;

  • *** P < 0.001.

View this table:
Table 5.

Prevalence of heavy alcohol use by demographic and geographic risk factors in women: weighted proportions

UnadjustedAdjusteda
Risk factors%OR95% CIPOR95% CIP
Age (years)<0.001<0.001
    18–2516.21.01.0
    26–3412.50.7(0.5–1.2)0.7(0.4–1.1)
    35–5410.90.6(0.5–0.9)**0.6(0.4–0.8)***
    55+1.90.1(0.1–0.2)***0.1(0.1–0.2)***
    F for trend158.1***70.8***
Region< 0.05<0.05
    West5.61.01.0
    North-central7.31.3(0.9–2.1)1.5(1.0–2.2)
    Southeast11.22.2(1.3–3.6)**2.2(1.4–3.5)**
Urbanicityn.s.a
    Rural6.41.0
    Semi-urban10.41.7(1.0–2.8)*
    Urban8.51.4(0.9–2.1)
    F for trend1.9
Education<0.01a
    Primary2.31.0
    Secondary9.94.7(2.2–9.8)***
    Specialized secondary10.24.8(2.1–11.0)***
    Higher6.93.1(1.5–6.4)**
    F for trend3.0
Financial status<0.001a
    Very inadequate6.91.0
    Inadequate8.21.2(0.8–1.7)
    Adequate13.62.1(1.5–3.0)***
    F for trend13.6***
Employment status<0.001<0.05
    Out of the labor force3.71.01.0
    Unemployed16.35.1(3.1–8.3)***2.2(1.2–3.9)*
    Employed11.63.5(2.5–4.7)***1.6(1.0–2.5)*
Marital status<0.05a
    Never married13.11.0
    Married before7.20.5(0.3–0.8)**
    Married8.40.6(0.4–0.9)*
Parental status<0.001a
    Not parent of a child6.51.0
    Parent of child12.62.1(1.5–2.9)***
  • a Non-significant risk factors were eliminated from the adjusted model in order of least significance.

  • * P < 0.05;

  • ** P < 0.01;

  • *** P < 0.001.

In women (Table 5), the odds of heavy alcohol use were higher for those who were 18–25 years, lived in the Southeast, lived in a semi-urban settings, had more education, had adequate financial status, were employed or unemployed (vs out of the labor force), had never been married and had a child under the age of 18 living at home. The multivariate analysis had an explanatory power of 9.2%, and showed that female heavy alcohol users were more likely to be young, living in the Southeast and employed or unemployed (vs out of the labor force).

Figure 2 shows the significant linear trends for the prevalence of heavy alcohol use with increasing number of risk factors.

Fig. 2.

Relationship of number of risk factors to heavy alcohol use (Tests for linear trend: men (F = 73.6, P < 0.001); women (F = 69.3, P < 0.001).

DISCUSSION

The Ukraine-WMH survey is the first large population-based study of alcohol use in Ukraine. Our survey found that the rate of heavy use was 22.0%. More than 4 out of 5 heavy users consumed >80 g/typical day at least monthly. In both men and women, living in the Southeast and being employed or unemployed (vs out of the labor force) were significant risk factors. Among men, the odds of heavy alcohol use were also increased for those 26–54 years of age, with a high school education and with young children at home. In women, the unique risk factor was being in the 18–25 years age group.

Our national rates of heavy alcohol use (38.7% for men; 8.5% for women) were comparable with the estimates of monthly binge drinking in Russia reported by Bobak et al. (1999). However, there were substantial regional variations within Ukraine, with the highest rate in the Southeast region, which is heavily populated by Russian migrants of the Soviet era. It is interesting to note that two of the three regional surveys in Russia also found particularly high rates (51–82% for men; 5–11% for women) (Pakriev et al., 1998; Malyutina et al., 2001). Clearly, the issue of regional variations is important for targeted public health planning, and more focused research using comparable tools and sampling techniques is needed.

A comparison of our findings with surveys elsewhere suggests that there are differences in risk factors between countries on either side of the Atlantic; i.e. the positive association of male heavy alcohol use with middle age is consistent with surveys in Russia, the Baltics and Bulgaria (Balabanova and McKee, 1999; Bobak et al., 1999; Cockerham, 2000; McKee et al., 2000), as well as Britain, Germany and Spain (Bloomfield, 1998; Meyer et al., 2000; Mateos et al., 2002; Schroder et al., 2004). However, the findings differ from North American surveys in which male heavy drinking declines after age 25 (Layne and Whitehead, 1985; Bachman et al., 2002; Naimi et al., 2003). Our finding that marital status is not significantly related to heavy consumption in men is consistent with those of other surveys in the former Eastern bloc (Balabanova and McKee, 1999; Bobak et al., 1999; Cockerham, 2000; Malyutina et al., 2004), but not with findings from Western epidemiologic research. Most American and European surveys find adult family responsibilities (e.g. being or becoming married or a father) to be negatively correlated with heavy drinking in men (Power and Estaugh, 1990; Miller-Tutzauer et al., 1991; Temple et al., 1991; Kunz and Graham, 1996; Gotham et al., 1997; Hajema and Knibber, 1998; Vik et al., 2003). Not only did our study find marriage unrelated to heavy use, but fatherhood was positively related.

Our study focuses on the prevalence and demographic correlates of heavy alcohol use. Future population-based case–control and longitudinal research is needed to evaluate other risk factors, such as genetic (Wang et al., 2004), environmental (Velleman, 1992; Farrell et al. 1995), alcohol-specific (e.g. drinking norms, peer drinking) (Choquette et al., 1985; Chen et al., 1994; Lintonen and Konu, 2004), and factors associated with historical and concurrent political and economic upheavals in Ukraine. Research focused on economic risk factors is particularly important. To the best of our knowledge, the suggested link between economic stress and heavy drinking in the former Soviet Union (Pridemore, 2002; Walberg et al., 1998) has not been tested with individual-level data.

Further research is also needed to fully understand the public health significance of our findings. The high prevalence of heavy alcohol use, particularly by Ukrainian men, may mean that associated costs and problems are similarly prevalent. On the other hand, its ubiquity may be a sign that heavy drinking is largely normative, and thus, non-pathological. Indeed, despite high rates of alcoholism in Ukraine (Bromet et al., 2004, World Mental Health Consortium, 2004), only 23.1% of the male and 11.2% of the female heavy users met the diagnostic criteria for an alcohol disorder. Given Ukraine's declining population, a more pressing issue is the impact of heavy alcohol use on mortality. A recent longitudinal study by Malyutina et al. (2002) found that middle-aged, Russian men who binged on an at least monthly basis were at a higher risk for death by injury and that more intense drinkers were at risk for cardiovascular-related death. Future prospectively designed epidemiologic research is needed, which incorporates direct measurement of cardiovascular functioning, and begins with a cohort of men before the age at risk for heavy drinking. The impact of heavy alcohol use on the economy is obviously important to Ukraine during its transitional period. We found that nearly one of two employed men and one of ten employed women are heavy users. Thus, future studies should examine the effect of heavy use on occupational functioning (e.g. productivity, days missed at work).

Past prevention efforts in the former Soviet Union have been criticized for being indiscriminate and overly inclusive (Korolenko et al., 1994). The results of our multivariate analyses indicated that heavy alcohol use is demographically widespread, but that groups with a higher concentration of heavy users can be identified when risk factors are clustered. Almost two-thirds of male heavy alcohol users in our study had three or more risk factors. One preliminary study in Russia (Lukomskaia, 1997) suggests that primary care may serve as an appropriate venue for screening and intervening with patients who have high-risk profiles (e.g. unemployed father with a high school education). Moreover, since half of the heavy alcohol users in Ukraine are employed, work settings provide obvious places to locate intervention programs. Alcohol testing is fairly common in the public and private sectors (Rice and Repo, 2000), but nearly half of Ukraine's oblasts (counties) are without state-funded outpatient programs for prevention or treatment of cases identified in the workplace (Ministry of Health of Ukraine, 2004).

Our study has a number of strengths, including a large national probability sample that reflects the economic conditions of Ukraine (Rhodes et al., 1999; Barnett et al., 2000), the structured assessment, and the use of a valid index for heavy alcohol use. Nevertheless, there are a number of limitations. First, the data are based on self-report, and past studies in former Soviet Union countries have shown that self-reported data underestimates drinking compared with biomedical and sales data (Laatikainen et al., 2002; Nemtsov, 2003). Second, 33 potential respondents could not be interviewed because they were never sober when the interviewer tried to approach them. Third, the diagnostic instrument only determined lifetime abstinence, and hence the non-heavy use group included both users and non-users of alcohol in the past year; unfortunately, it is not possible to estimate what effect this has had on the risk factors presented here. Lastly, the sample excluded people in the military or other institutional settings where heavy drinking is a well-known problem. Thus for men, our estimates may be conservative.

CONCLUSIONS

The Ukraine-WMH survey found that 1 out of every 3 men and 1 out of every 12 women consume alcohol heavily. Future analyses should address the social and economic costs of heavy alcohol use in Ukraine. The associations with age, sex, employment, unemployment and region suggest that interventions should be developed to target high-risk pockets of the population.

Acknowledgments

This study was funded by the National Institute of Mental Health (MH61905, Evelyn Bromet PI). The survey was conducted as part of the WHO World Mental Health Survey consortium directed by Ronald C. Kessler (Harvard University) and T. Bedirhan Ustun (World Health Organization). The authors thank Volodymyr Paniotto, Valeriy Khmelko and Victoria Zakhozha (Kiev International Institute of Sociology) and Julia Pievskaya (Ukrainian Psychiatric Association) for their invaluable assistance in conducting the field work; Inna Korchak, Roxalana Mykhaylyk, Margaret Bloom, and Svetlana Stepukhovich for their excellent work in translating the various components of the study; and Ellen Walters for her continuing statistical support. Special thanks go to the interviewers and participants for their dedication and diligence.

REFERENCES

View Abstract