TYPE OF ALCOHOLIC BEVERAGE AND HIGH-RISK DRINKING: HOW RISKY IS BEER DRINKING IN KOREA?

Alcohol & Alcoholism Vol. 39, No. 1, pp. 39–42, 2004 doi:10.1093/alcalc/agh015, available online at www.alcalc.oupjournals.org TYPE OF ALCOHOLIC BEVERAGE AND HIGH-RISK DRINKING: HOW RISKY IS BEER DRINKING IN KOREA? WOOJIN CHUNG Graduate School of Public Health, Yonsei University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120–752, Korea (Received 21 July 2003; first review notified 3 September 2003; in revised form 9 October 2003; accepted 24 October 2003) Abstract — Aims: To examine the association of beverage type with high-risk drinking in Korea. Methods: Data were analysed from the Korea’s 1997 behavioural risk factor surveillance system survey collected by a stratified random sampling method (n = 1045). After alcohol consumption per drinking day was categorized into three risk levels for short-term or ‘acute’ harm on the basis of the World Health Organization’s guidelines, logistic regression analysis adjusting for socio-demographic factors was performed for each beverage type. Results: Using beer drinkers as the reference, those who drank soju (the most widely consumed traditional beverage in Korea) showed an adjusted odds ratio (OR) of 3.62 [95% confidence interval (CI) = 2.28–5.76] for high and very high risk from drinking and those who drank spirits had adjusted OR of 10.65 (95%CI = 4.32–26.27). Conclusions: In Korea, unlike in many western countries, beer is not the highest risk beverage. The relatively high price of beer in Korea is likely to be one influence. It is concluded that country-specific differences are important in studies on associations between alcoholic beverage types and high risk drinking. INTRODUCTION were completed (response rate: 58.1%). Among the 1060 cases, 1045 cases provided full information for our analysis. Together with their socio-demographic characteristics, respondents were asked what types of beverage they had consumed in the month preceding the survey. Questions were also asked concerning the frequency of drinking in the preceding month for each consumed beverage and the amount usually consumed when they drank the beverage. Before including type of beverage in our analyses, we took into consideration two country-specific characteristics related to drinking in Korea. First, two types of alcoholic beverages are traditionally popular in Korea: soju and makkolli. Soju, the most widely used traditional beverage, is a distilled liquor made from fermented potatoes and its alcohol content, 25%, is between that of wine (10–14%) and spirits (more than 35%). Makkolli is the oldest liquor in Korea, made by fermenting non-glutinous and glutinous rice together. The constituents of soju are not similar to those of wine and spirits. Wine is not common in Korea; in terms of alcohol content, wine is more close to makkolli (about 7%) than to beer or soju. On the basis of this information, we first classified a beverage into one of five types (beer, soju, spirits, makkolli, wine). In contrast, for multivariate analyses, we tied makkolli and wine into one category, ‘makkolli or wine.’ At the time of the survey, beer strength varied between products, but was beginning to be standardized at 4% following guidelines from the Korea Alcohol and Liquor Industry Association and Korea’s National Tax Service. Second, the survey revealed that some respondents drank two or more types of beverage in the last month. Some Korean drinkers consume more than one type of beverage in a day. If we include two or more types of beverage for those respondents, findings in a logistic regression analysis will not be independent. To avoid this problem, we selected only one type of beverage from among two or more alcoholic beverages that they drank, in the following way. We first categorized respondents into two groups, drinkers (someone who consumed at least one drink in the past month) and nondrinkers. For each drinker (n = 530), beverage-specific quantity-frequency scores were computed by multiplying the frequency of drinking of each beverage by the usual amount Several studies have suggested that beer is more likely to be associated with high-risk behaviours, such as heavy and excessive drinking and other alcohol-related problems, than are other types of beverage (Lemmens et al., 1992; Jensen et al., 2002). Also, researchers have found that those who preferred beer, compared with those who typically preferred spirits or wine, drank to excessive levels and were more likely to drive after drinking (Berger and Snortum, 1985). However, these findings were obtained from studies in western developed countries. Whether beer is a relatively high-risk beverage in non-western countries is not certain. Using data available in Korea and the World Health Organization’s (WHO) guidelines on high-risk drinking for acute harm, the present study addresses whether beer is a relatively high-risk beverage in Korea. SUBJECTS AND METHODS We analysed data from Korea’s behavioural risk factor surveillance system survey conducted during April and May 1997 by the Ministry of Health and Social Welfare, Republic of Korea, in cooperation with the College of Medicine, Seoul National University. The survey used a stratified random sampling design. The Korean samples were divided into 28 strata based on 120 common socio-demographic characteristics obtained from the National Population and Household Census. Then, households from each stratum were chosen by a proportional allocation method, and interviewed by telephone. Collection of data and procedures were designed in accordance with the methodology recommended for behavioural risk factor surveillance system surveys (Remington et al., 1988). In total, 1826 Koreans aged 15 years or older were interviewed over the telephone, and 1060 cases Correspondence: Graduate School of Public Health, Yonsei University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120–752, Korea. Tel.: +82 2 361 5094; Fax: +82 2 392 7734; E-mail: 39 Alcohol & Alcoholism Vol. 39, No. 1 © Medical Council on Alcohol 2004; all rights reserved 40 W. CHUNG consumed (Kilty, 1990; Smart and Walsh, 1995). After assigning alcohol contents of beverage types [beer 4% (by volume), soju 25%, spirits 40%, makkolli 7% and wine 12%] to the computed quantity-frequency score and assuming that 1 ml consumed alcohol includes 0.79 g of pure alcohol (World Health Organization, 2000), we calculated pure alcohol content which each drinker had consumed in a session through each type of beverage. Finally, we selected the single beverage of which each drinker had consumed the most pure alcohol in the last month, and then finally included that type of beverage in our analyses. Various socio-demographic characteristics previously shown to be related to individual alcohol intake and consumption were assessed as confounding factors. Age was divided by the mean age, 46.37 years. Respondents were classified into male and female. They were grouped according to whether they lived with a partner or not. Respondents repo (...truncated)