Four Years of Alcohol Use Trajectories Among Korean Adults: Using Latent Class Growth Analysis
Method: Multiwave longitudinal representative data used in this study is originally from the 2006-2009 Korean Welfare Panel Study (KWPS). The analysis included a sub-sample of 12,403 individuals who were “legal drinkers” at Wave I and those with three or more years of longitudinal information on alcohol use. Both quantity per drinking episode and frequency of drinking per week were included in the analysis to track respondents’ alcohol use, then analyzed via LCGA for classification into different growth trajectories. Unlike standard latent growth modeling which estimates a single growth trajectory, the LCGA identifies distinct subgroups of sample that show different growth trajectories based on individuals’ response patterns over time. A series of multinomial logistic regressions were performed to further explore characteristics of change patterns accounting for age, gender, marital status, and SES (education and median household income).
Results: Four trajectories of drinking quantity were identified: consistently low (n=7,877, 63.5%), consistently heavy (n=1,837, 14.8%), changing from heavy to moderate (n=1,014, 8.2%), and changing from moderate to heavy (n=1,675, 13.5%). Three different trajectory groups of drinking frequency were: consistently low (n=9,902, 81.6%), consistently moderate (n=1,358, 11.2%), and increasing (n=868, 7.2%). [These trajectories will be visually presented in the poster.] The stability in absolute alcohol intake (both by frequency and quantity) level was rather low, and close to two-thirds of the respondents clearly tended to maintain their relative drinking position over time. However, approximately one in seven of the sample continuously held their position as ‘heavy drinkers (>4 drinks a day)’ relative to the group. These heavy drinkers were more likely to be males (OR=46.81), younger (OR=0.94), high income family (above 60% of median household income) (OR=1.48), and have high school (OR=1.35) or less education (OR=1.27). Respondents whose drinking progressed (by frequency) over time were more likely to be males (OR=18.36), married (OR=1.21), and have high school (OR=1.75) or less education (OR=2.56).
Implications: While the majority of the sample showed somewhat a consistently low alcohol use level, yet the results imply that aggressive implementation of prevention programs are needed that target towards individuals whose drinking level gradually increase. The findings also provide demographic profiles of these groups who may be at highest risk thus need practitioners’ clinical assistance for early prevention and treatment of alcohol use disorders.