I.Introduction

Problematic drinking is the third-leading riskfactor in the world for numerous health and social problems[1]. Worldwide, men have substantially higher risks of excessive alcohol use (16.1%) and death from such excessive use (6.2%) than that of women (4.2% and 1.1%, respectively)[1]. Among the countries in the Organization for Economic Cooperation and Development (OECD), South Korea ranked 13thfor alcohol consumption by adults, and annual alcohol consumption increased two-fold since 1996[2].

The high rates of alcohol use and problematicdrinking among Korean men are a concern. The negative consequences of problematic drinking frequently exacerbate normal age-related changes in the body, such as a weakened immune system, in addition to aggravating multiple illnesses [3]. The most common physical health hazards related to problematic drinking include stroke, liver disease, exacerbated diabetes, and hypertension [4]; these hazards are the main causes of death among middle- and older-aged adults in Korea[5].

Problematic drinking is common amongKorean men because South Korea has an alcohol-permissive culture for men and features strongly in defined gender roles[5]. Men who comply with the cultural values and norms are more likely to consume alcohol than their female counterparts[2]. Drinking is not just considered as a means of coping with stress for men who often face stressful life events such as unemployment, it is also associated with masculinity and social pressures about being a man[1]. Many Korean men also believe that drinking is a way of engaging in a social activity that helps develop good relationships with colleagues [6].

Different age groups have different drinkingbehaviors[4][5]. Alcohol consumption is considered an essential element of many business and social activities for middle-aged men; thus, middle-aged Korean men are more likely to engage in excessive consumption of alcohol with one another[5]. Yet, a study[4] found that about 30% of older Korean men were considered to be heavy drinkers and that they usually drank at home compared to younger Korean men. Because of their different socio-economic status and social activities, the risk factors for problematic drinking among middle- and older-aged Korean men seem to be different from each other.

South Korea has one of the fastest-agingpopulations among all the OECD countries. In 2000, 7% of all Koreans were over 65 years old, but this segment is expected to increase to 14.4% by 2020[7]. Generally, social roles decrease, and undesired leisure time increases as Korean adults age. Older Korean men are likely to prefer inactive leisure activities, such as talking and drinking with friends, than participating in social activities[8].

Several studies have found that importantsocial factors (i.e., socio-economic status, social ties, marital status, employment status, and gender) are associated with problematic drinking in the general population[9][10] and in the Korean population[1]. Typically, social ties are thought to be beneficial to good physical and mental health[11]. However, depending upon the socio-cultural context, social ties can encourage problematic drinking behavior. Having friends who drink is positively related to higher levels of heavy drinking, and it indicates that heavy drinkers tend to maintain a supportive drinking environment[9].

In addition, health related factors are strongpredictors of problematic drinking. A study by Kim and the associates[5] reported that the perception of being unhealthy was positively related to frequent and excessive consumption of alcohol among Korean men. Furthermore, some studies have reported a positive association between problematic drinking and smoking cigarettes[4][12][13] and a positive relationship between problematic drinking and depression[13].

In this study, we examined the relationshipsbetween the psychosocial factors and problematic drinking of Korean men across different age groups. When considering the differential associations of psychosocial factors and problematic drinking between middle- and older-aged men, it is important to investigate psychosocial factors that may have different implications on problematic drinking in men at different ages.

Ⅱ.Methods

1.Design and Study Population

This study performed a secondary analysis ofdata from the survey of Korean Longitudinal Study of Aging (KLoSA), which collected information from August through December 2006 from a nationally representative sample of community-dwelling adults who were 45 years of age and older in South Korea. Multi-staged stratified sampling methods were used to provide national representation. The first stage was stratified using census enumeration districts by location, rural or urban residence, and housing type. The second stage consisted of randomly sampled households within selected enumeration districts. A total of 10,254 respondents completed the interviews. The household response rate was 70.7%, and the individual response rate within households was 75.4%. Weights were assigned to reduce the sampling error introduced by stratified sampling. Weights were computed in three steps: (1) the weight based on two-staged sampling probability, (2)the non-response-adjusted weight, and (3)the benchmark weight reflecting population distribution changes based on demographic changes shown in other nationwide scale surveys.

The aim of this study was to examine thecorrelates of problematic drinking with demographic/socioeconomic dimensions, social support, and health related features among middle- and older-aged Korean men. Thus, women respondents were excluded from the analysis, and our final samples included 3,631 middle- (45 to 64 years) and 1,173 older-aged men (> 65 years). The study was exempt from the approval of Institutional Review Board because it involved secondary data analysis on data which do not include personal identifiers.

2.Measures

1)Problematic drinking

Problematic drinking was measured with theCAGE instrument. The CAGE scale consists of four items with a response of yes (0) or no (1), and scores range from 0 to 4, where a score above 2 indicates problematic drinking. Studies have shown that two or more “yes” answers on the CAGE instrument identify a problematic drinking behavior with a sensitivity of 83-93% and a specificity of 76-90%[14].

2)Socioeconomic variables

Socioeconomic variables include education,household income, and employment. Education was coded into four categories: 1) elementary school or less; 2) middle school graduate; 3) high school graduate; and 4) some college, a college degree, or a graduate degree. Household income was measured with the following two steps. First, total household income was divided by the square root of the number of individuals in the household. Then, this number was divided into quartiles: 1) the highest 25%, 2) the second highest 25%, 3) the third highest 25%, and 4) the lowest 25%. Employment status was coded into four categories: 1) currently employed, 2) retired, 3) unemployed but intending to work, and 4) never employed.

3)Social ties variables

Social ties were measured with fourindicators: marital status, living arrangements, contact with close persons, and social activities. Marital status was coded into two categories: 1) married, 2) widowed, divorced or separated, and never married. Living arrangements were dichotomized into 1) living with others and 2) living alone. Contact with a close person was assessed by the frequency of contact with friends, relatives, or neighbors. The response options for contact with close persons were recoded into three categories: 1) almost every day (more than 4 times per week), 2) less than once a week, and 3) less than once a month or never. Social activities were assessed by the number of participations in religious actives, community activities, etc. and recoded into three categories: none, one, and two or more.

4)Health-related variables

Health-related measures included self-ratedhealth, chronic disease, depression, and smoking status variables. Self-rated health was measured by the question “How would you rate your health in general?” Five response options were provided: 1) very good, 2) good, 3) fair, 4) bad, and 5) very bad. Chronic disease was measured by the total number of chronic diseases diagnosed by a physician and reported by participants, such as hypertension, diabetes, cancer, lung disease, heart problems, stroke, arthritis, and gastrointestinal illnesses. The chronic disease variable did not approach a normal distribution. Thus, the responses were recoded into three categories: 0, 1, and 2 or more. Smoking was coded into three categories: 1) never smoked, 2) former smoker, and 3) current smoker.

Depression was measured with the Short-formof the Center for Epidemiological Studies Short Depression Scale (CES-D10), which consists of 10 items with a 4-point Likert scale (0 to 3). The scores ranged from 0 to 30, with higher scores indicating more severe depressive symptoms. The 10-item CES-D has excellent properties for use as a screening instrument for the identification of major depression in older adults [15]. The Cronbach’s alpha coefficient for this study was 0.79.

3.Statistical analysis

Differences in problematic drinking accordingto each study variable were examined with chi-square tests. Significant differences in problematic drinking between middle- and older-aged men according to each variable were determined by Fisher’s exact tests because of small contingency table cells among the older-aged men. Multiple-classification analysis was used to calculate the age-adjusted prevalence of problematic drinking.

Logistic regression analyses were doneseparately for each age group to assess the relationships between problematic drinking and socioeconomic, social ties, and health-related variables. Two steps were conducted to verify the differences between the groups of middleand older-aged Korean men. First, Wald chi-square statistics was used to test the differences among the coefficients between the two groups. Second, disturbance variance unconstrained models were fitted to assess whether there was significant residual variations across the two age groups. All statistical analyses were done with the SPSS 21.0 and SAS 9.0 computer programs. All analyses were based on weighted data.

Ⅲ.Results

1.Description of samples

The sample characteristics of the two groupsare summarized in <Table 1>. Middle-aged men (66.6%) reported more current drinking than that of the older-aged men (65.7%) (p<0.01). The prevalence of problematic drinking among middle- and older-aged men was 11.4% and 8.8%, respectively.

Older-aged men were less educated than thatof the middle-aged men. Specifically, 50.1% of the older-aged men reported having an elementary school education or less. Middle-aged men were more likely than older-aged men to have higher average monthly incomes and current employment. There were substantial differences between middle- and older-aged men in terms of marital status. Specifically, 8.4% of the older-aged men were widowed compared with 1.9% of the middle-aged men. Older-aged men reported less frequent contacts with close friends, relatives, and/or neighbors and less participation in social activities. More than one-third of the older-aged men (37.8%) reported poor or very poor health status, whereas slightly less than one-sixth of the middle-aged men (15.0%) reported having a poor health status. Older-aged men reported having more chronic diseases than middle-aged men, and older-aged men reported more depressive symptoms (35.9%) than that of middle-aged men (18.5%).

2.Differences in problematic drinking for each variable for each age group

Significant variables associated withproblematic drinking among middle-aged men included educational attainment (p<0.01), household income (p<0.01), employment status (p<0.05), contact with others (p<0.01), self-rated health (p<0.01), chronic diseases (p<0.01), smoking status (p<0.01), and depressive symptoms (p<0.01). More specifically, middle-aged men with lower educational attainment and household income and who were unemployed were more likely to report problematic drinking than their counterparts with higher education and household income and current employment. Those who had more contact with others almost every day were more likely to be problematic drinkers than those who had less contact. In addition, those who reported poor health and more chronic diseases and depressive symptoms were more likely to report problematic drinking.

Among the older-aged men, poor self-ratedhealth (p<0.05) and frequent contact with close persons (p<0.05) were associated with problematic drinking. The most significant differences in factors related to problematic drinking between middle- and older-aged men were educational attainment and depressive symptoms <Table 1>. Middle-aged men with lower educational attainment and more severe depressive symptoms were more likely to be problematic drinkers than older-aged men with lower educational attainment and more severe depressive symptoms.

3.Predictors of problematic drinking for middle- and older-aged men

<Table 2> presents the results of a logisticregression on problematic drinking among middleand older-aged men. There were noticeable differences between the two age groups.

Education, employment status, contact withclose persons, depressive symptoms, and smoking status were significantly related to problematic drinking among middle-aged men. Those with an elementary school education level or less were more likely to report problematic drinking than those with a college education or higher (OR=1.96, 95% CI: 1.26-3.05). Those who were unemployed but intended to work were more likely to report problematic drinking than those who were currently employed (OR=1.67, 95% CI: 1.05-2.66). Middle-aged men having frequent contact with close persons were more likely to report problematic drinking (OR=2.35, 95% CI: 1.52-3.65) than those having social contact less than a month or never. Depressive symptoms were associated with a two-fold increase in the risk of problematic drinking. Current smokers were twice more likely to be problematic drinkers than that of non-smokers (OR=2.39, 95%CI: 1.75-3.27).

The sample of older-aged men had novariables that were significantly associated with problematic drinking. The association between poor self-rated health and problematic drinking shown in <Table 1> was not significant in the multiple logistic regression analysis results shown in <Table 2>. In addition, the association between frequency of social contact and problem drinking was also not significant after controlling for other factors in <Table 2>.

Ⅳ.Discussion

This study investigated the factors influencingproblematic drinking among middle- and older-aged men in South Korea. Separate analyses for two different age groups showed the importance of social interaction for problematic drinking in both the middle- and older-aged groups and also enabled the identification of different socioeconomic and health-related factors related to problematic drinking in the different age groups.

Although there is considerable research on thepositive impact of social relationships on health and wellbeing, the influence of the peer network has also been shown to be a strong predictor of alcohol use across one’s life span. For example, Homish and Leonard (2008) found that a larger social network of heavy drinkers was associated with greater levels of heavy drinking [9]. In addition, a previous study [10] showed that a person’s social network affects drinking behavior, which implies that network effects may occur through social norms. Therefore, our findings of positive association between frequent social contacts and problematic drinking in middle-aged men could be partially explained by entrenched drinking customs, such as urging one another to drink and drinking with business partners in Korea and may also support the previous proposition that a drinking culture enhances interpersonal relationships in South Korea[5][11]. Korean adult men are still accepted, or even encouraged to drink during social occasions to enhance friendships and build relationships with working colleagues. Meanwhile, some of previous studies in western countries have also suggested that interactions with others who share similar beliefs about drinking reinforce and encourage problematic drinking[9][10]. However, the KLoSA survey only included information on the frequency of social contacts. No detailed information was available on such factors as the number of drinking buddies in the network, the composition of the social network, the quality of the peer relationships, and changes in the social network. For better understanding of the strong association between the degree of social contact and problematic drinking shown in this study, further studies are needed to examine how such variables could affect problematic drinking.

Although we could not find any significantimpact of frequent social interactions on problematic drinking among older-aged men in the multiple logistic regression models, older-aged men with social interactions at rare intervals showed a higher prevalence of problematic drinking than older-aged men with moderate social relationship in Table 1. This result in our study may partially reflect the fact that social isolation impacts the health and behavioral habits of older adults. The social networks of older adults can impact their health positively through encouragement to adhere to medical treatment or to refrain from negative or risky behaviors. Without the positive influence of social network members, older adults who are socially isolated are at risk for many negative behaviors such as heavy drinking. The sudden disruption in lifestyle caused by retirement and bereavement-which can lead to decreased social activity-is thought to be a major contributory factor among older people who develop a drinking problem, as are isolation and loneliness. In addition, reverse causation could be possible, in that older-aged men with a long history of excessive drinking have very little in the way of meaningful relationships. These findings may help health professionals identify isolated older-aged adults at greater risk for problematic drinking and provide more efficient interventions focused on enhancing the social support network. A strengthening of positive influence from others who are not drinking buddies could lead to more moderate alcohol habits and better health for older adults.

With respect to middle-aged Korean men, inthis study, no unemployment was associated with problematic drinking. Because of the strong influence of Confucianism in Korean culture, Korean men are expected to be the family breadwinners, and women are expected to be homemakers. Korean men are more likely to be the primary source of income for their families; thus, unemployment has become a significant threat to the financial well-being of families' [16]. Male unemployment can result in the loss of patriarchal power within families, which could result in marital conflict and stress [17]. A previous study [18] reported that family conflict affected heavy drinking among middle-aged Korean men. Marital conflict and stress may be among the primary causes of alcohol consumption for Korean men. Consistent with other studies [19][20], low education among middle-aged men is a risk factor for problematic drinking. This finding could suggest that those with less education may not be informed about the dangers of problematic drinking, including the health and legal risks. It is important to a developing a specific intervention that targets middle-aged men with low educational attainment. Consistent with other studies [4][12], smoking was associated with increased problematic drinking among middle-aged men. Smoking issues should be considered in addressing problematic drinking because drinking and smoking among men are permissive behaviors in Korea. Alcohol and tobacco are the top causes of preventable death, and the dependence on alcohol and tobacco is correlated. Many alcoholics smoke, putting them at high risk for tobacco-related complications including multiple cancers, lung disease, and heart disease (i.e., cardiovascular diseases). Therefore, health professionals working with alcoholics should know that the link between alcohol and tobacco has important implications. Additionally, middle-aged men with more depressive symptoms were more likely to be problem drinkers than their less depressed counterparts. It is a well-known fact that alcohol is a type of self-medication used to relieve psychological distress and depressive symptoms [21]. Alcohol consumption often coincides with a depressed mood which can lead to thoughts of suicide. Because alcohol compromises judgment and makes people impulsive and likely to take risks, health professionals must pay attention to depressed middle aged men who have problematic drinking habits. In addition, we suggest that health professionals assess drinking habits when treating depressed middle-aged men to identify initially the risk of self-harm.

With respect to older-aged men, no significantrelationship among health-related factors and problematic drinking was found. The lack of significance among health-related factors and problematic drinking could partially be explained by the possibility that men who abuse alcohol die earlier. Regardless of our results, older-aged people are at increased risk for harm from alcohol, not only because physiological changes associated with ageing mean that they have a lower tolerance, but also because as people age, they are more likely to have chronic illnesses, take medications, or have functional impairments. The prevalence of risky drinking in older Korean men aged 70 or over is likely to increase (5.1% in 2007, 9.8% in 2008, and 7.3% in 2009)[22]. Furthermore, because the number of older-aged people is increasing, even in the unlikely event that the prevalence of risky drinking remains the same, the absolute number will increase. Excessive alcohol use in older-aged people is associated with adverse consequences. Therefore, geriatric and medical professionals should initiate recommended guidelines to reduce drinking in older-aged people.

The prevalence of problem drinking betweenmiddle- and older-aged men was not statistically different. This finding is not consistent with previous studies that have found that problems with alcohol use decrease with age [22]. One explanation for the inconsistent finding could be partially due to the use of a different measurement for problematic drinking. Most previous research [23] that found a negative relationship between drinking alcohol and age used quantity or volume of alcohol consumption as an indicator of problematic drinking. In our study, problematic drinking was measured with CAGE which did not assess the quantity, frequency, or volume of alcohol use. However, our findings reflect the current drinking habits of the Korean population because the survey of KLoSA is a nationally representative sample of Korean adults. Another possible explanation for the non-significant finding is the increased life expectancy of men (77.9 years in 2012 compared with 61.8 years in 1980) [24]. The increased life expectancy of older-aged men may indicate an increased number of drinking years.

The current study has several limitations.First, the CAGE instrument may underestimate alcohol-related problems in an older population. Several studies have suggested that CAGE is more useful for identifying problematic drinking in younger individuals than in the elderly [25]. Thus, a lack of sensitivity to identifying problematic drinking in older-aged men may account for the failure to find a significant relationship between certain variables and problematic drinking. Second, this study only used one point of a panel dataset and was thus unable to project whether factors associated with problematic drinking behavior changed as age increased. Longitudinal studies are required to examine the life-course of problematic drinking and to better distinguish cohort and time period effects. Lastly, this study did not assess drinking companions in the social-ties variable. Instead, it was assumed that close persons had similar norms, attitudes, and beliefs about drinking.

The strength of the KLoSA data that wereused in this study should be noted. The KLoSA dataset is a large, representative national sample of non-institutionalized men aged 45 and older that is drawn from the general population. This dataset provided sufficient statistical power for the current analyses.

Ⅴ.Conclusion

The findings have important implications inimproving health behaviors among Korean male adults by developing age-specific strategies. As observed in our study, frequent social relationships with close persons increase problematic drinking; health professionals treating Korean men with problematic drinking need to pay attention to middle-aged men actively engaged in social networks.