Subjects
Data from the 2014 Korean Youth Risk Behavior Web-Based Survey (KYRBS) were used. The KYRBS is a government-approved statistical survey that has been performed annually by the Korea Centers for Disease Control and Prevention since 2005 to monitor health-related risk behaviors among Korean adolescents. A stratified multistage cluster sampling design was used to obtain a nationally representative sample of middle- and high-school students. After the survey was fully explained, only participants who provided informed consent completed the anonymous, self-administered web-based survey during a regular class period. The KYRBS did not collect any identifiers, such as students’ names, schools, telephone numbers, home addresses, or social security numbers. A total of 74,167 students from 400 middle schools and 400 high schools participated in the survey, representing a response rate of 97.2 %. The final sample included 72,060 students (36,470 boys and 35,590 girls, mean age 14.94 ± 1.75 years, range 12–18 years). Additional details about the sampling methodology and survey procedure are available elsewhere [33]. The KYRBS was reviewed and approved by the institutional review board of Korea Centers for Disease Control and Prevention (2014-06EXP-02-P-A). The KYRBS data used for the present study is openly available.
Measurements
Socio-demographic variables included sex, age, place of residence (name of city), parents’ level of education, residential type, perceived family economic status, and perceived academic achievement. Places of residence were classified as rural area, small city, or large city. Perceived family economic status and academic achievement were assessed by a 5-point Likert scale (low, low-middle, middle, high-middle, and high).
Lifetime illicit drug use was assessed with the following question: “Have you ever used drugs that are often used nonmedically (e.g., inhalants, glue, stimulants, cannabis, amphetamines, marijuana, codeine, neuroleptics) for mood elevation, hallucinatory experience, or excessive dieting?” Participants who responded “Yes” to this question were asked to choose the drugs they had ever used from a list of 33 commonly used illegal drugs classified into nine categories (for example: inhalants, e.g. butane gas, bond, thinner, varnish, lighter gas; stimulants, e.g. caffeine pills; tranquilizers, e.g. Valium, Ativan, Xanax; a large dose of antitussive, e.g. Luminar, Lubiking, Soma, Zinolta; diet pills, e.g. Lasix, Prion; marijuana; Hiropon; other hallucinogens, e.g. ketamine, LSD, Ecstasy, 5-MeO-DiPT, Kratom, Ice; opioid drugs, e.g. opiates, morphine, Demerol, Nubaine; cocaine). The 33 drugs listed were accompanied by more common “street” names or brand names. They were also asked to choose the drug they used the first time.
Lifetime smoking was assessed with the following question: “Have you ever smoked even a puff of a cigarette, cigar, or pipe?” Lifetime alcohol use was assessed with the following question: “Have you ever used alcohol?”
Problematic alcohol use during the last 12 months was assessed with the CRAFFT, a 6-item instrument that is used to screen for alcohol use in the adolescent population. It consists of the following yes-no questions: (1) Have you ever ridden in a Car driven by someone (including yourself) who had been using alcohol? (2) Do you ever use alcohol to Relax, feel better about yourself, or fit in? (3) Do you ever use alcohol while you are by yourself or Alone? (4) Do you ever Forget things you did while using alcohol? (5) Do your Family or Friends ever tell you that you should cut down on your drinking? and (6) Have you ever gotten into Trouble while you were using alcohol? Two or more positive responses indicate the potential for a significant alcohol problem.
The level of perceived stress was measured with the following question: “How much stress do you usually feel?” The response options were very little (1), a little (2), an average amount (3), a lot (4), and very much (5). On the basis of the responses, the participants were classified into the following two groups for multivariate logistic regression analyses: (i) ≤ average perceived stress (1–3) and (ii) > average perceived stress (4–5).
Having experienced a depressive mood during the last 12 months was measured with the following question: “In the past 12 months, have you ever felt depression or hopelessness severe enough to compromise your daily activities during 2 weeks or more?” The response options were “yes” or “no”.
Suicidal ideation was assessed with the question: “Have you seriously thought of committing suicide during the past 12 months?” Existence of a suicide plan was assessed with the question: “Have you concretely planned suicide during the past 12 months?” Suicide attempts were assessed with the question: “Have you attempted suicide during the past 12 months?”
Lifetime sexual relations were assessed with the following questions: “Have you ever experienced sexual relations with a partner of the opposite sex?” and “Have you ever experienced sexual relations with a same-sex partner?”
Statistical analysis
Descriptive statistics were used to present the prevalence of substance use. Logistic regression tests were performed to compare the socio-demographic factors between substance users and non-users. Odds ratios (ORs) and 95 % confidence intervals (CIs) were calculated using lifetime alcohol, tobacco, and illicit drug use as the main outcome variable and each socio-demographic variable as the principal predictor. To analyze the association between the adolescents’ substance use and psychological variables, logistic regression tests were performed using lifetime alcohol, tobacco, and illicit drug use as the principal predictors and each psychological variable as the main outcome variables, after controlling for age, sex, area of residence, parental education, residential type, socio-economic status, and level of academic achievement. All analyses took into account the sampling design parameters, weighting, clustering, and stratification factor. The proportion of general subject characteristics was weighted according to the respondent’s probability of being selected for the sex-, grade-, and school type-specific distributions for the region [33]. The finite population correction (fpc) factor was used to avoid the overestimation, when developing variance estimates for population parameters. SAS 9.3 (SAS Institute, Cary, NC) was used to perform all other statistical analyses., and a p-value less than 0.05 was considered significant.