Suicide and Life-Threatening Behavior 45 (4) August 2015 © 2014 The American Association of Suicidology DOI: 10.1111/sltb.12143
477
Psychosocial–Environmental Risk Factors for Suicide Attempts in Adolescents with Suicidal Ideation: Findings from a Sample of 73,238 Adolescents SUN MI KIM, MD, JI HYUN BAEK, MD, DOUG HYUN HAN, MD, PHD, YOUNG SIK LEE, MD, PHD, AND DEBORAH A. YURGELUN-TODD, PHD
We determined risk factors that discriminate between suicide attempt (SA) adolescents and suicidal ideation only (SI only) adolescents using data from the 2010 Korea Youth Risk Behavior Web-based Survey (12–19 years; N = 73,238). In males, heavy alcohol use, drug use, and high perceived sadness/ hopelessness showed significant effects on the presence of SA versus the presence of SI only. In females, along with these variables, low academic achievement, poor perceived health status, high perceived stress, and unhealthy coping strategy were also significantly related to the presence of SA versus SI only. Therefore, clinical interventions targeting adolescents’ psychological distress are warranted to prevent suicide.
Suicide is a major public health problem that merits increased research attention. In 2010, the suicide rate among Korean adolescents was 5.2 per 100,000 people, and suicide was the leading cause of death for Korean adolescents 10 to 19 years of age (Korean Association for Suicide Prevention & Korea Centers for Disease Control & Prevention [KCDC], 2011). The suicide rate of Korean adolesSUN MI KIM, Department of Psychiatry, College of Medicine, Chung-Ang University, Seoul, Korea; JI HYUN BAEK, School of Medicine, Sungkyunkwan University, Seoul, Korea and Bipolar Clinic and Research Program, Massachusetts General Hospital, Boston, MA, USA; DOUG HYUN HAN and YOUNG SIK LEE , Department of Psychiatry, College of Medicine, Chung-Ang University, Seoul, Korea; DEBORAH A. YURGELUNTODD, Brain Institute, University of Utah, Salt Lake City, UT, USA. Addresss correspondence to Young Sik Lee, MD, PhD, Department of Psychiatry, Chung-Ang University Medical Center, 102 Heukseok-ro, Dongjak-gu, Seoul 156-755, Korea; E-mail: [email protected]
cents has been increasing in recent decades, in contrast to other Organization for Economic Cooperation and Development (OECD) countries that have had a decreasing suicide rate during the same period (OECD Social Policy Division, 2011; Statistics Korea, 2013). The number of suicides per 100,000 people aged 10 to 19 years rose 57.2% to 5.58 in 2011 from 3.19 in 2001, according to data from the state-run Korea Health Promotion Foundation. In 2009, the 1-year prevalence of suicidal ideation (SI) was 15.2% in males and 23.5% in females, and the rates of suicide attempts (SAs) were 3.3% and 6.0%, respectively, for Korean youth 12 to 18 years of age (KCDC, 2010). Numerous psychosocial–environmental risk factors associated with adolescent suicidal behaviors have been extensively researched. A number of studies have reported that psychosocial distress, such as feelings of sadness, hopelessness, and loneliness, are related to SI (Peltzer & Pengpid, 2012) and SA (Nock et al., 2008). Suicidal
478 behaviors have also been reported to be associated with low socioeconomic status (SES; Kokkevi, Rotsika, Arapaki, & Richardson, 2012) and various family-related adverse factors, such as family discord and poor parent–child relationships (Borowsky, Resnick, Ireland, & Blum, 1999; Bridge, Goldstein, & Brent, 2006; Canetto, 1997). Several studies of Korean adolescents have found that depressive mood, hostility, low self-esteem, discordance in parent–child relationships, parental divorce, low SES, and body dissatisfaction have positive associations with SI in both males and females (An, Ahn, & Bhang, 2010; Kim & Kim, 2009; Park, Schepp, Jang, & Koo, 2006). Similarly, other studies have found that being involved in school bullying, academic stress, insufficient coping strategies, poor parent–child relationships, depression, substance abuse, and psychosomatic symptoms are risk factors for adolescent SA in both sex groups (Juon, Nam, & Ensminger, 1994; Kim, 2008). Relatively few studies have specifically assessed whether any risk factors show different patterns between individuals with SI only and those who have made attempts. Attempters have been reported to be significantly more likely to report histories of stressful life events, being sexually active, smoking more than one cigarette per day, and having smoked marijuana, compared with participants with SI only (King, Strunk, & Sorter, 2011; Smith, Cukrowicz, Poindexter, Hobson, & Cohen, 2010). In a study on bulimic female adolescents, it was found that, compared with those who had SI only, adolescents with histories of SA were more likely to have been exposed to more frequent violence during their childhoods, to live alone, and to be dissatisfied with their relationships with their partners (Nickel et al., 2006). Therapeutic interventions that decrease rates of SA have failed to achieve similar decreases in SI (Brown et al., 2005; Linehan et al., 2006). Therefore, research efforts identifying the similarities and differences between SI and SA are important to develop interventions to prevent transitioning from ideation to actual attempts. In this
RISK FACTORS
FOR
ADOLESCENT SUICIDE ATTEMPTS
study, we aimed to identify psychosocial– environmental risk factors that can distinguish adolescents who will attempt suicide from those with SI only in a nationally representative sample of Korean adolescents. We also examined the gender differences in those risk factors for SA.
MATERIALS AND METHODS
Participants and Data Collection Data from the 2010 Korea Youth Risk Behavior Web-based Survey (K-YRBWS) were used for the analyses. This survey was conducted by the KCDC and administered to 74,980 participants in a nationally representative sample of Korean adolescents aged 12 to 19 years. The self-report questionnaires were written in Korean, and participants entered their answers to the questions online in the computer classrooms of individual schools. The K-YRBWS questionnaires consisted of 128 questions regarding participants’ demographic characteristics and their status related to their physical and mental health. Every survey questionnaire was required to be completed during the online survey session. Logical fallacies and extreme values were treated as missing data. Consequently, the response rate to the 2010 K-YRBWS was 97.7%. In this study, data from 38,391 male and 34,847 female adolescents (total N = 73,238) were included in the analyses. These data included responses regarding participants’ suicidal behaviors and other family, social, and individual factors known to be related to suicidal behaviors. The institutional review board of Chung-Ang University Medical Center permitted the data analysis without obtaining informed consent because the 2010 K-YRBWS data were collected and analyzed anonymously. Study Procedures Subjects were categorized into one of three mutually exclusive groups depending
KIM
ET AL.
on their recent suicide status: (1) adolescents who had attempted suicide during the previous year (“suicide attempt: SA” group); (2) adolescents who had had suicide ideas, but had not attempted suicide in the previous year (“suicidal ideation only: SI only” group); and (3) those who had never considered or attempted suicide during the previous year (“no suicidality” group). The following psychosocial and environmental characteristics, which have been reported as risk factors for suicidal behaviors in previous studies, were used for the analyses (Table 1). Suicidal Ideation and Attempt. To assess SI, participants were asked whether they had ever seriously wanted to commit suicide during the past year (yes or no). The presence of an SA was assessed by a yes-orno question asking whether they had ever attempted suicide during the previous year. Gender. Participants were categorized as female or male, according to their responses to the question on sex. Age. The biological age of each participant was categorized into one of two groups: “aged 12–15” (the usual age for a middle school student in Korea) or “aged 16–19” (the usual age for a high school student in Korea). Socioeconomic Status. Participants’ responses regarding SES were categorized into three levels: high, middle, or low. Academic Achievement. Academic achievement was categorized into three levels, high, middle, or low, according to participants’ responses to the question about their school records over the past year. High Perceived Sadness/Hopelessness. “High perceived sadness/hopelessness” was defined by a positive answer to the yes-orno question asking about the experience of feeling sadness or hopelessness to the extent that respondents had abandoned their daily activities for at least 2 weeks during the past year. High Perceived Stress. “High perceived stress” was defined as a participant’s response of 4 or 5 (on a 5-point scale from 1 = not at all to 5 = extremely much) to the
479 question regarding the participant’s level of perceived stress. Poor Perceived Health Status. Individuals were categorized as having “poor perceived health status” if their responses were 4 or 5 (on a 5-point scale from 1 = very good to 5 = very poor) to the question asking how they perceived their states of health. Heavy Substance Use. “Heavy substance use” was defined as a participant’s meeting one of the following characteristics: current cigarette smoking at least 20 days per month, experience of a blackout due to heavy drinking during the prior month, or substance use history other than nicotine and alcohol for the purpose of mood elevation, hallucinatory experience, or aggressive weight management. The presence of a substance use history other than nicotine and alcohol was assessed by a question asking whether they had ever sniffed glue or butane gas or used illegal drugs such as amphetamines, methamphetamines, or opioids. We conducted analyses using “heavy alcohol use,” “heavy smoking,” and “drug use” separately. Living with a Single Parent or Others. Family structure was categorized as “living with both parents” if a participant reported currently living with both parents, and “living with a single parent or others” if a participant reported currently living in a family with a single parent or with neither mother nor father. Unhealthy Coping Strategies. Coping behaviors were assessed using a question inquiring what the respondent did most often to relax when feeling stressed, based on the instruction to choose one of thirteen specific coping behaviors. In this study, responses were classified into two categories, healthy or unhealthy coping strategies, to distinguish substance use from other coping behaviors. “Unhealthy coping strategies” included using smoking cigarettes or drinking alcoholic beverages to cope. Statistical Analysis Due to the stratified cluster sampling design used to collect the data, we
480
RISK FACTORS
FOR
ADOLESCENT SUICIDE ATTEMPTS
TABLE 1
Psychosocial and Environmental Variables Used in the Analyses; Questions from the 2010 Korea Youth Risk Behavior Web-Based Survey Variable and question Suicidal ideation During the past 12 months, did you ever seriously consider attempting suicide? Suicide attempt During the past 12 months, did you ever actually attempt suicide? Socioeconomic status What is your family’s socioeconomic status? Academic achievement During the past 12 months, how was your school record? Perceived level of sadness/hopelessness During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you stopped doing some usual activities? Perceived level of stress Usually, how often do you feel stressed? Perceived level of health status Usually, how do you feel about your health status? Substance use During the past 30 days, how many cigarettes did you smoke per day? During the past 30 days, on how many days did you have at least one drink of alcohol? Did you ever use butane gas, adhesives, stimulant, philopon, amphetamines, cough medicine, or sedatives with the purpose of mood change, hallucination, or excessive weight loss? Family structure Who are you living with? Coping strategies What do you do most often to relax when you feel stressed?
implemented the Taylor series linearization method and a design-based method using the Complex Samples module of the PASW statistics software package, version 18 (SPSS
Response category Yes or no
Yes or no
Upper, upper middle, middle, middle lower, lower class Upper, upper middle, middle, middle lower, lower class Yes or no
5-point Likert scale, 1 (not at all) to 5 (very often) 5-point Likert scale, 1 (very healthy) to 5 (very unhealthy) 0, 1, 2–5, 6–9, 10–19, more than 20 (cigarettes per day) 0, 1 or 2, 3–5, 6–9, 10–19, 20–29, all 30 (days per month) Never, sometimes in the past (not currently), currently
Mother, father, both, others Watching TV, listening to music, reading a book, engaging in sports, venting emotion by talking to others, online chatting, playing online or mobile games, eating, singing, drinking alcoholic beverages, smoking cigarettes, sleeping, absence of specific coping behavior, using coping behavior that is not listed above
Inc., Chicago, IL, USA). There were no missing values in the portion of the data that we analyzed for this study. Because of the large number of observations, statistical
KIM
ET AL.
significance was set a priori at a = .01 (twosided) to limit type-I error. The aforementioned psychosocial and environmental characteristics related to suicide were compared between the groups (i.e., SA group, SI-only group, no suicidality group) using chi-square tests. Due to the cross-sectional nature of the study, we used both Wald chi-square and odds ratios (ORs) as measures of association without implying any causal association (Le Strat & Hoertel, 2011). Weighted percentages with standard errors are presented. Effect size was estimated by the ORs: about 1.5 to 1 = small, about 2.5 to 1 = moderate, about 4 to 1 = strong, and about 10 to 1 = very strong effect (Rosenthal, 1996). To determine the predictors of recent suicide status, we performed a multinomial logistic regression analysis using group membership of recent suicide status (i.e., SA group, SI-only group, no suicidality group) as a dependent variable. We entered all the dependent variables that showed significant differences individually in the chi-square tests as independent variables in a single model. To explore gender-specific predictors, we conducted the analyses separately for males and females.
RESULTS
Group Comparisons on Demographic and Psychosocial–Environmental Variables Among 73,238 individuals, 10,395 (14.3%) reported having experienced SI but not having attempted suicide, and 3,616 (5.8%) reported having attempted suicide at least once within the 12 months prior to the survey. The rates of all psychosocial and environmental characteristics included in this study showed statistically significant differences among groups (Table 2). The SI-only group had higher rates for all factors except younger age and drug use compared with the no suicidality group. The SA group had higher rates for all factors compared with the no suicidality group and the SI-only group.
481 Risk Factors for the SI-Only Group and the SA Group Compared with the No Suicidality Group As all psychosocial and environmental characteristics explored in this study showed significant differences among groups, all were entered as independent variables in the subsequent multinomial logistic regression analyses. When all individuals were included in the model, female, younger age, low SES, living with a single parent or others, heavy alcohol use, poor perceived health status, high perceived stress level, unhealthy coping strategy, and high perceived sadness/hopelessness showed significant effects on the presence of SI only versus no suicidality (Model 1 of Table 3). High perceived sadness/hopelessness and high perceived stress level were the strongest risk factors after adjusting for all other risk factors (high perceived stress level: OR 2.73, 95% CI 2.55, 2.92; high perceived sadness/hopelessness: OR 6.69, 95% CI 6.26, 7.16). All factors except heavy smoking showed significant effects on the presence of SA versus no suicidality. We additionally ran separate analyses for males and females (Models 2 and 3 of Table 3). In males, heavy alcohol use, poor perceived health status, high perceived stress level, unhealthy coping strategy, and high perceived sadness/hopelessness showed significant effects on the presence of SIonly versus no suicidality. These variables as well as drug use and low academic achievement were significantly related to the presence of SA versus no suicidality. In females, all variables except heavy smoking, drug use, and unhealthy coping strategy showed significant effects on the presence of SI only versus no suicidality. These variables as well as drug use and unhealthy coping strategy were significantly related to the presence of SA versus no suicidality. High perceived stress level and high perceived sadness/hopelessness consistently showed the highest ORs, indicating that they contributed the strongest effects on suicide outcomes after adjusting for all other factors.
0.3 0.3 0.1 0.2 0.00 0.3 0.1 0.3 0.2 0.3
21.5
15.2
2.8 7.4 0.5 35.8
6.2
36.5
4.1
27.2
77.6
0.5
0.4
0.6
0.5
0.3 0.4 0.1 0.7
0.5
0.7
1.7 1.1
SE
85.5
14.1
77.9
19.5
12.7 18.0 4.3 51.8
24.8
35.2
59.5 65.0
Wt%
0.8
0.8
0.9
0.8
0.7 0.9 0.4 1.1
1.0
1.1
1.8 1.5
SE
SA (N = 3,616)
5842.34
396.26
3072.59
952.28
568.33 300.49 305.19 270.08
215.98
424.76
384.95 12.43
OR
Adj Wald v2 p value
2.55 2.36, 2.76 4.81 4.53, 5.11 1.73 1.55, 1.92
9.27 8.70, 9.87 15.78 13.95, 17.85 1.70 1.50, 1.93
477
Psychosocial–Environmental Risk Factors for Suicide Attempts in Adolescents with Suicidal Ideation: Findings from a Sample of 73,238 Adolescents SUN MI KIM, MD, JI HYUN BAEK, MD, DOUG HYUN HAN, MD, PHD, YOUNG SIK LEE, MD, PHD, AND DEBORAH A. YURGELUN-TODD, PHD
We determined risk factors that discriminate between suicide attempt (SA) adolescents and suicidal ideation only (SI only) adolescents using data from the 2010 Korea Youth Risk Behavior Web-based Survey (12–19 years; N = 73,238). In males, heavy alcohol use, drug use, and high perceived sadness/ hopelessness showed significant effects on the presence of SA versus the presence of SI only. In females, along with these variables, low academic achievement, poor perceived health status, high perceived stress, and unhealthy coping strategy were also significantly related to the presence of SA versus SI only. Therefore, clinical interventions targeting adolescents’ psychological distress are warranted to prevent suicide.
Suicide is a major public health problem that merits increased research attention. In 2010, the suicide rate among Korean adolescents was 5.2 per 100,000 people, and suicide was the leading cause of death for Korean adolescents 10 to 19 years of age (Korean Association for Suicide Prevention & Korea Centers for Disease Control & Prevention [KCDC], 2011). The suicide rate of Korean adolesSUN MI KIM, Department of Psychiatry, College of Medicine, Chung-Ang University, Seoul, Korea; JI HYUN BAEK, School of Medicine, Sungkyunkwan University, Seoul, Korea and Bipolar Clinic and Research Program, Massachusetts General Hospital, Boston, MA, USA; DOUG HYUN HAN and YOUNG SIK LEE , Department of Psychiatry, College of Medicine, Chung-Ang University, Seoul, Korea; DEBORAH A. YURGELUNTODD, Brain Institute, University of Utah, Salt Lake City, UT, USA. Addresss correspondence to Young Sik Lee, MD, PhD, Department of Psychiatry, Chung-Ang University Medical Center, 102 Heukseok-ro, Dongjak-gu, Seoul 156-755, Korea; E-mail: [email protected]
cents has been increasing in recent decades, in contrast to other Organization for Economic Cooperation and Development (OECD) countries that have had a decreasing suicide rate during the same period (OECD Social Policy Division, 2011; Statistics Korea, 2013). The number of suicides per 100,000 people aged 10 to 19 years rose 57.2% to 5.58 in 2011 from 3.19 in 2001, according to data from the state-run Korea Health Promotion Foundation. In 2009, the 1-year prevalence of suicidal ideation (SI) was 15.2% in males and 23.5% in females, and the rates of suicide attempts (SAs) were 3.3% and 6.0%, respectively, for Korean youth 12 to 18 years of age (KCDC, 2010). Numerous psychosocial–environmental risk factors associated with adolescent suicidal behaviors have been extensively researched. A number of studies have reported that psychosocial distress, such as feelings of sadness, hopelessness, and loneliness, are related to SI (Peltzer & Pengpid, 2012) and SA (Nock et al., 2008). Suicidal
478 behaviors have also been reported to be associated with low socioeconomic status (SES; Kokkevi, Rotsika, Arapaki, & Richardson, 2012) and various family-related adverse factors, such as family discord and poor parent–child relationships (Borowsky, Resnick, Ireland, & Blum, 1999; Bridge, Goldstein, & Brent, 2006; Canetto, 1997). Several studies of Korean adolescents have found that depressive mood, hostility, low self-esteem, discordance in parent–child relationships, parental divorce, low SES, and body dissatisfaction have positive associations with SI in both males and females (An, Ahn, & Bhang, 2010; Kim & Kim, 2009; Park, Schepp, Jang, & Koo, 2006). Similarly, other studies have found that being involved in school bullying, academic stress, insufficient coping strategies, poor parent–child relationships, depression, substance abuse, and psychosomatic symptoms are risk factors for adolescent SA in both sex groups (Juon, Nam, & Ensminger, 1994; Kim, 2008). Relatively few studies have specifically assessed whether any risk factors show different patterns between individuals with SI only and those who have made attempts. Attempters have been reported to be significantly more likely to report histories of stressful life events, being sexually active, smoking more than one cigarette per day, and having smoked marijuana, compared with participants with SI only (King, Strunk, & Sorter, 2011; Smith, Cukrowicz, Poindexter, Hobson, & Cohen, 2010). In a study on bulimic female adolescents, it was found that, compared with those who had SI only, adolescents with histories of SA were more likely to have been exposed to more frequent violence during their childhoods, to live alone, and to be dissatisfied with their relationships with their partners (Nickel et al., 2006). Therapeutic interventions that decrease rates of SA have failed to achieve similar decreases in SI (Brown et al., 2005; Linehan et al., 2006). Therefore, research efforts identifying the similarities and differences between SI and SA are important to develop interventions to prevent transitioning from ideation to actual attempts. In this
RISK FACTORS
FOR
ADOLESCENT SUICIDE ATTEMPTS
study, we aimed to identify psychosocial– environmental risk factors that can distinguish adolescents who will attempt suicide from those with SI only in a nationally representative sample of Korean adolescents. We also examined the gender differences in those risk factors for SA.
MATERIALS AND METHODS
Participants and Data Collection Data from the 2010 Korea Youth Risk Behavior Web-based Survey (K-YRBWS) were used for the analyses. This survey was conducted by the KCDC and administered to 74,980 participants in a nationally representative sample of Korean adolescents aged 12 to 19 years. The self-report questionnaires were written in Korean, and participants entered their answers to the questions online in the computer classrooms of individual schools. The K-YRBWS questionnaires consisted of 128 questions regarding participants’ demographic characteristics and their status related to their physical and mental health. Every survey questionnaire was required to be completed during the online survey session. Logical fallacies and extreme values were treated as missing data. Consequently, the response rate to the 2010 K-YRBWS was 97.7%. In this study, data from 38,391 male and 34,847 female adolescents (total N = 73,238) were included in the analyses. These data included responses regarding participants’ suicidal behaviors and other family, social, and individual factors known to be related to suicidal behaviors. The institutional review board of Chung-Ang University Medical Center permitted the data analysis without obtaining informed consent because the 2010 K-YRBWS data were collected and analyzed anonymously. Study Procedures Subjects were categorized into one of three mutually exclusive groups depending
KIM
ET AL.
on their recent suicide status: (1) adolescents who had attempted suicide during the previous year (“suicide attempt: SA” group); (2) adolescents who had had suicide ideas, but had not attempted suicide in the previous year (“suicidal ideation only: SI only” group); and (3) those who had never considered or attempted suicide during the previous year (“no suicidality” group). The following psychosocial and environmental characteristics, which have been reported as risk factors for suicidal behaviors in previous studies, were used for the analyses (Table 1). Suicidal Ideation and Attempt. To assess SI, participants were asked whether they had ever seriously wanted to commit suicide during the past year (yes or no). The presence of an SA was assessed by a yes-orno question asking whether they had ever attempted suicide during the previous year. Gender. Participants were categorized as female or male, according to their responses to the question on sex. Age. The biological age of each participant was categorized into one of two groups: “aged 12–15” (the usual age for a middle school student in Korea) or “aged 16–19” (the usual age for a high school student in Korea). Socioeconomic Status. Participants’ responses regarding SES were categorized into three levels: high, middle, or low. Academic Achievement. Academic achievement was categorized into three levels, high, middle, or low, according to participants’ responses to the question about their school records over the past year. High Perceived Sadness/Hopelessness. “High perceived sadness/hopelessness” was defined by a positive answer to the yes-orno question asking about the experience of feeling sadness or hopelessness to the extent that respondents had abandoned their daily activities for at least 2 weeks during the past year. High Perceived Stress. “High perceived stress” was defined as a participant’s response of 4 or 5 (on a 5-point scale from 1 = not at all to 5 = extremely much) to the
479 question regarding the participant’s level of perceived stress. Poor Perceived Health Status. Individuals were categorized as having “poor perceived health status” if their responses were 4 or 5 (on a 5-point scale from 1 = very good to 5 = very poor) to the question asking how they perceived their states of health. Heavy Substance Use. “Heavy substance use” was defined as a participant’s meeting one of the following characteristics: current cigarette smoking at least 20 days per month, experience of a blackout due to heavy drinking during the prior month, or substance use history other than nicotine and alcohol for the purpose of mood elevation, hallucinatory experience, or aggressive weight management. The presence of a substance use history other than nicotine and alcohol was assessed by a question asking whether they had ever sniffed glue or butane gas or used illegal drugs such as amphetamines, methamphetamines, or opioids. We conducted analyses using “heavy alcohol use,” “heavy smoking,” and “drug use” separately. Living with a Single Parent or Others. Family structure was categorized as “living with both parents” if a participant reported currently living with both parents, and “living with a single parent or others” if a participant reported currently living in a family with a single parent or with neither mother nor father. Unhealthy Coping Strategies. Coping behaviors were assessed using a question inquiring what the respondent did most often to relax when feeling stressed, based on the instruction to choose one of thirteen specific coping behaviors. In this study, responses were classified into two categories, healthy or unhealthy coping strategies, to distinguish substance use from other coping behaviors. “Unhealthy coping strategies” included using smoking cigarettes or drinking alcoholic beverages to cope. Statistical Analysis Due to the stratified cluster sampling design used to collect the data, we
480
RISK FACTORS
FOR
ADOLESCENT SUICIDE ATTEMPTS
TABLE 1
Psychosocial and Environmental Variables Used in the Analyses; Questions from the 2010 Korea Youth Risk Behavior Web-Based Survey Variable and question Suicidal ideation During the past 12 months, did you ever seriously consider attempting suicide? Suicide attempt During the past 12 months, did you ever actually attempt suicide? Socioeconomic status What is your family’s socioeconomic status? Academic achievement During the past 12 months, how was your school record? Perceived level of sadness/hopelessness During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you stopped doing some usual activities? Perceived level of stress Usually, how often do you feel stressed? Perceived level of health status Usually, how do you feel about your health status? Substance use During the past 30 days, how many cigarettes did you smoke per day? During the past 30 days, on how many days did you have at least one drink of alcohol? Did you ever use butane gas, adhesives, stimulant, philopon, amphetamines, cough medicine, or sedatives with the purpose of mood change, hallucination, or excessive weight loss? Family structure Who are you living with? Coping strategies What do you do most often to relax when you feel stressed?
implemented the Taylor series linearization method and a design-based method using the Complex Samples module of the PASW statistics software package, version 18 (SPSS
Response category Yes or no
Yes or no
Upper, upper middle, middle, middle lower, lower class Upper, upper middle, middle, middle lower, lower class Yes or no
5-point Likert scale, 1 (not at all) to 5 (very often) 5-point Likert scale, 1 (very healthy) to 5 (very unhealthy) 0, 1, 2–5, 6–9, 10–19, more than 20 (cigarettes per day) 0, 1 or 2, 3–5, 6–9, 10–19, 20–29, all 30 (days per month) Never, sometimes in the past (not currently), currently
Mother, father, both, others Watching TV, listening to music, reading a book, engaging in sports, venting emotion by talking to others, online chatting, playing online or mobile games, eating, singing, drinking alcoholic beverages, smoking cigarettes, sleeping, absence of specific coping behavior, using coping behavior that is not listed above
Inc., Chicago, IL, USA). There were no missing values in the portion of the data that we analyzed for this study. Because of the large number of observations, statistical
KIM
ET AL.
significance was set a priori at a = .01 (twosided) to limit type-I error. The aforementioned psychosocial and environmental characteristics related to suicide were compared between the groups (i.e., SA group, SI-only group, no suicidality group) using chi-square tests. Due to the cross-sectional nature of the study, we used both Wald chi-square and odds ratios (ORs) as measures of association without implying any causal association (Le Strat & Hoertel, 2011). Weighted percentages with standard errors are presented. Effect size was estimated by the ORs: about 1.5 to 1 = small, about 2.5 to 1 = moderate, about 4 to 1 = strong, and about 10 to 1 = very strong effect (Rosenthal, 1996). To determine the predictors of recent suicide status, we performed a multinomial logistic regression analysis using group membership of recent suicide status (i.e., SA group, SI-only group, no suicidality group) as a dependent variable. We entered all the dependent variables that showed significant differences individually in the chi-square tests as independent variables in a single model. To explore gender-specific predictors, we conducted the analyses separately for males and females.
RESULTS
Group Comparisons on Demographic and Psychosocial–Environmental Variables Among 73,238 individuals, 10,395 (14.3%) reported having experienced SI but not having attempted suicide, and 3,616 (5.8%) reported having attempted suicide at least once within the 12 months prior to the survey. The rates of all psychosocial and environmental characteristics included in this study showed statistically significant differences among groups (Table 2). The SI-only group had higher rates for all factors except younger age and drug use compared with the no suicidality group. The SA group had higher rates for all factors compared with the no suicidality group and the SI-only group.
481 Risk Factors for the SI-Only Group and the SA Group Compared with the No Suicidality Group As all psychosocial and environmental characteristics explored in this study showed significant differences among groups, all were entered as independent variables in the subsequent multinomial logistic regression analyses. When all individuals were included in the model, female, younger age, low SES, living with a single parent or others, heavy alcohol use, poor perceived health status, high perceived stress level, unhealthy coping strategy, and high perceived sadness/hopelessness showed significant effects on the presence of SI only versus no suicidality (Model 1 of Table 3). High perceived sadness/hopelessness and high perceived stress level were the strongest risk factors after adjusting for all other risk factors (high perceived stress level: OR 2.73, 95% CI 2.55, 2.92; high perceived sadness/hopelessness: OR 6.69, 95% CI 6.26, 7.16). All factors except heavy smoking showed significant effects on the presence of SA versus no suicidality. We additionally ran separate analyses for males and females (Models 2 and 3 of Table 3). In males, heavy alcohol use, poor perceived health status, high perceived stress level, unhealthy coping strategy, and high perceived sadness/hopelessness showed significant effects on the presence of SIonly versus no suicidality. These variables as well as drug use and low academic achievement were significantly related to the presence of SA versus no suicidality. In females, all variables except heavy smoking, drug use, and unhealthy coping strategy showed significant effects on the presence of SI only versus no suicidality. These variables as well as drug use and unhealthy coping strategy were significantly related to the presence of SA versus no suicidality. High perceived stress level and high perceived sadness/hopelessness consistently showed the highest ORs, indicating that they contributed the strongest effects on suicide outcomes after adjusting for all other factors.
0.3 0.3 0.1 0.2 0.00 0.3 0.1 0.3 0.2 0.3
21.5
15.2
2.8 7.4 0.5 35.8
6.2
36.5
4.1
27.2
77.6
0.5
0.4
0.6
0.5
0.3 0.4 0.1 0.7
0.5
0.7
1.7 1.1
SE
85.5
14.1
77.9
19.5
12.7 18.0 4.3 51.8
24.8
35.2
59.5 65.0
Wt%
0.8
0.8
0.9
0.8
0.7 0.9 0.4 1.1
1.0
1.1
1.8 1.5
SE
SA (N = 3,616)
5842.34
396.26
3072.59
952.28
568.33 300.49 305.19 270.08
215.98
424.76
384.95 12.43
OR
Adj Wald v2 p value
2.55 2.36, 2.76 4.81 4.53, 5.11 1.73 1.55, 1.92
9.27 8.70, 9.87 15.78 13.95, 17.85 1.70 1.50, 1.93
