Gender-specific factors associated with the use of mental health services for suicidal ideation: Results from the 2013 Korean Community Health Survey

This study examined gender-specific factors associated with the use of mental health services (MHS) for suicidal ideation (SI). We included data on 6,768 males and 12,475 females who had experienced SI over the past year from the nationwide 2013 Korean Community Health Survey. These individuals were grouped as MHS users for SI if they had received professional counseling at medical institutions, professional counseling agencies, or community health centers for SI-related problems. Their information on sociodemographic factors, socio-familial relationships, health behaviors, and health status were included as exposures in a logistic regression analysis. Of the 19,243 individuals, 7.0% of the males and 10.5% of the females used MHS for SI treatment. For males with SI, living in an urban area, being a widower, and having unhealthy behaviors (frequent alcohol consumption and infrequent walking) were associated with underuse of MHS. For females with SI, frequent contact with friends, low level of religious activity, and good self-rated health were associated with underuse of MHS. For both males and females, those who were younger, completed higher education, and experienced depression/suicide attempts in the past year were more likely to use MHS for SI. These findings suggest that gender-specific factors should be used to inform suicide prevention strategies.


Introduction
Suicide is one of the most important public health issues in the world. Suicide rates vary widely across countries, and Korea has twice as high of a suicide rate compared with other Organization for Economic Cooperation and Development countries [1]. Korea's suicide rate was 26.5 deaths per 100,000 people and was the fifth leading cause of death in 2015 [1]. Although most people with suicidal ideation (SI) do not die by suicide, SI is closely related to suicide attempts, PLOS  and these suicidal behaviors are positively related to death by suicide [2]. Therefore, linking people with SI to appropriate mental health services (MHS) is an important strategy for preventing suicide. Studies have found that the majority of people at high risk for suicide do not use any form of MHS [3]. Suicide-related characteristics differ depending on gender. In almost all countries, males have a higher suicide rate than that of females [4], but females attempt suicide more often than do males [5]. Gender-specific associated factors should be considered in the prevention and management of suicide attempts, and the gender-specific factors underlying the use of MHS for SI should also be investigated. Previous studies have examined the associations between socioeconomic factors and use of MHS for SI [3,[6][7][8]. However, there is a lack of research on the differences in factors related to the use of MHS by gender [9]. A Korean study reported that only 8.2% of adults with SI have used MHS for their mental health problems, and revealed the socioeconomic factors related to the use of MHS in the population with SI from the National Health Survey [10]. However, the small sample, limited independent variables, and the lack of an analysis by gender prevented a comprehensive understanding of the use of MHS among the general population with SI.
Therefore, the purpose of this study was to provide information for establishing improved and customized suicide prevention policies by identifying gender-specific factors associated with the use of MHS for SI and by examining gender-specific barriers for those seeking help from MHS.

Study population
The Korean Community Health Survey (KCHS) conducted by the Korea Centers for Disease Control and Prevention is a nationwide survey that has been carried out annually since 2008 by trained surveyors using a computer-assisted personal interviewing method. Multistage, stratified, and random sampling was used to select representative households in 253 local Korean communities based on resident registration information resulting from surveying an average of 900 individuals from each local community. This study used data from the 2013 KCHS collected from August 16, 2013 to October 31, 2013. A total of 228,781 individuals (102,722 males and 106,059 females) aged ! 19 years participated in the 2013 survey. The 2013 KCHS provides population-based estimates of health indicators, including health status, morbidity, health service use, and health behaviors using a standardized questionnaire consisting of 258 questions. After excluding participants with missing data regarding sociodemographic variables, socio-familial relationships, health behaviors, and health status, 19,243 subjects (6,768 males and 12,475 females) who had experienced SI over the past year were included in the final analysis. This study was conducted in accordance with the Declaration of Helsinki guidelines. Written informed consent was obtained from all participants in the KCHS. The study protocol was approved by the Institutional Review Board of Wonkwang University Hospital (WKUH 2017-05-018).

Outcome and variables measurements
Suicide-related behaviors, including SI, suicide attempts, and use of MHS for SI, were evaluated using a questionnaire. SI was defined as having had thoughts of wanting to die in the past year. The use of MHS for SI included subjects who had received professional counseling at a medical institution, professional counseling agency, or community health center for SI-related problems in the past year.
Information on each subject's sociodemographic factors, socio-familial relationships, health behaviors, and health status was collected using a questionnaire. A detailed description of the variables used in this study is provided in Table 1.

Statistical analysis
The participants' characteristics were compared according to gender using the chi-square test. After adjusting for all of the evaluated covariates, the adjusted odds ratio (aOR) with 95% confidence interval (CI) of using MHS for SI were subjected to multivariate logistic regression analysis. Logistic regression analyses were performed separately for males and females. All statistical analyses were performed using SPSS Statistics for Windows ver. 22.0 (IBM Co., Armonk, NY, USA). A P-value < 0.05 was considered significant.

Sample characteristics by gender
Characteristics according to gender are presented in Table 2. Of the 19,243 subjects who had experienced SI, 1,780 (9.3%) received professional counseling for SI, which was a significantly higher proportion in females (10.5%) than in males (7.0%). There were significant differences in age group, residence type, marital status, household composition, education level, monthly household income, and employment status between the genders; no significant difference in National Basic Livelihood Security (NBLS) status according to gender was observed. Compared with males, females contacted their family and neighbors more frequently and participated in religious activities more often. In contrast, males contacted their friends more frequently and participated more often in activities with friends and leisure than females. A greater proportion of males than females were current or past smokers and males tended to drink more frequently, but walk less frequently. Sleep duration and self-rated health status differed significantly according to gender. Compared with males, the proportions with high perceived daily stress, experienced depressive mood, having more than three chronic diseases, and diagnosis of depression were higher in females. The proportion of suicide attempts was higher in males (4.6%) than in females (3.9%).

Gender-specific factors predicting the use of MHS for SI
Univariate analyses showed that among males, sociodemographic factors (age group, residence type, marital status, household composition, education level, employment status, and NBLS), socio-familial relationships (contact with family, contact with neighbors, religious activity, and activities with friends), and health behaviors and health status (smoking status, frequency of alcohol use, walking activity, self-rated health, perceived daily stress, experience of depressed mood, number of chronic diseases, diagnosis of depression, and suicide attempts) were significantly associated with the use of MHS for SI. Among females, sociodemographic factors (age group, residence type, marital status, education level, employment status, and NBLS), sociofamilial relationships (contact with family, contact with neighbors, religious activity, activities with friends, leisure activities, and charitable activities), and health behaviors and health status (smoking status, walking activity, sleep duration, self-rated health, perceived daily stress, experience of depressed mood, diagnosis of depression, and suicide attempts) were significantly associated with the use of MHS for SI (data not shown).
Fully adjusted gender-specific relationships between the use of MHS for SI and sociodemographic factors, socio-familial relationships, and health behaviors and health status, as determined by logistic regression analysis, are presented in Tables 3 (males) and 4 (females). After  After adjusting for related variables, more religious activity (aOR = 1.28, 95% CI = 1.09-1.49) was positively associated, and more contact with friends (aOR = 0.80, 95% CI = 0.68-0.95) was negatively associated with the use of MHS in females. However, none of the socio-familial factors showed an association in males after full adjustment. Compared with non-drinkers, the OR of using MHS was lower in males who drank ! 4 times/week in the fully adjusted model (aOR = 0.58, 95% CI = 0.39-0.86). Although walking activity was negatively associated with the use of MHS in both genders in the unadjusted model, only males who walked less had a lower likelihood of using MHS after full adjustment (aOR = 0.70, 95% CI = 0.55-0.90). After adjusting for related factors, the OR for using MHS was greater in females with good health (aOR = 0.78, 95% CI = 0.61-0.99) than in those with poor health. After adjusting for related factors, the OR for using MHS among those who experienced depressed mood was significantly higher (2.18-fold, 95% CI = 1.70-2.79) in males and (2.15-fold, CI = 1.85-2.50) higher in females compared with those who did not experience depressed mood. The OR for using MHS was significantly higher among males (aOR = 29.95, 95% CI = 23.38-38.36) and females (aOR = 27.43, 95% CI = 23.55-31.95) who were diagnosed with depression compared with those who were not diagnosed with depression. Males (aOR = 2.31, 95% CI = 1.58-3.39) and females (aOR = 3.12, 95% CI = 2.40-4.06) who had attempted suicide had a higher OR for using MHS compared with those who had not attempted suicide.

Discussion
Among the community-dwelling general population with SI, this study examined genderspecific associations between sociodemographic factors, socio-familial relationships, health behaviors, and health status with use of MHS. Significant relationships were observed between using MHS and residence type, marital status, frequency of alcohol use, and walking activity in males, whereas contact with friends, religious activity, and self-rated health were significantly associated with use of MHS in females. In contrast, our findings showed that age, education level, experience of depressed mood, depression diagnosis, and suicide attempts were associated with the use of MHS for SI in males and females. In a review of 12 studies, the utilization rate of professional mental health providers by people with SI, suicide plans, and/or suicide attempts during the past year was approximately 29.5% [3]. We defined the use of MHS as professional counseling at a medical institution, professional counseling agency, or community health center for SI-related problems. Especially, in this study, the use of mental health counseling only included counseling received from visiting a mental health professional and did not include telephone or internet counseling. In addition, only SI-related issues, and no other mental health problems, were included in the definition of the use of MHS for SI. In this study, 7.0% of males and 10.5% of females (9.3% in total) used MHS for SI, which is slightly higher than the rate reported in a previous Korean study [10]. Previous studies have consistently reported that age is an important predictor of using MHS for SI [8,10]. Older people are less sensitive to psychiatric symptoms, whereas younger people are more aware of the need for MHS [11,12]. Similar to previous studies, the present study identified that age and use of MHS for SI were inversely associated in males and females, although the magnitude of the association was greater in males. Education level, which is an important indicator of socioeconomic status, was positively associated with the use of MHS for SI in both genders in the present study. A previous study reported that education level was not a significant determinant of utilizing MHS in males or females [9], but other studies have shown a significant association between education level and the use of MHS for SI [8,10,13]. Highly educated people generally use MHS because they are less stigmatized about mental illness and have a positive attitude toward the effectiveness of treatment for mental illness [14,15], whereas those with lower levels of education are more economically burdened by MHS and are less aware of mental illness problems and treatment, resulting in limited use of MHS [15,16]. In addition, older age and lower education levels are related to less knowledge about suicide, which may affect seeking help for SI [17]. Marital status, especially widowed or divorced, has a greater influence on suicide mortality in males than females [18,19]. In addition, our previous study identified that widowed, divorced, or separated males attempted suicide significantly more frequently than did married males, but widowed, divorced, or separated females attempted suicide significantly less frequently than did married females [20]. In the present study, the frequency of MHS use was significantly lower in widowers compared with married persons, indicating that widowers are vulnerable to suicide in Korea. Death or divorce of a spouse is a significant risk factor for suicide in both genders, but the impact on females is somewhat weaker, because females continue to receive support through social and family connections even after losing their spouse [21]. In addition, traditional male gender roles, including greater levels of strength and independence, often prevent them from seeking help for suicidal feelings and depression [22]. A Korean study reported that the risk of not using MHS was 2.75-fold higher for widows than for married people, but the results were not evaluated by gender [10]. Two Canadian studies in adults with SI reported an association between marital status and use of MHS [8,9]. One study found a significantly higher use of MHS in unmarried and divorced people than in married people [8], while the other study reported that marital status and use of MHS were not related in males or females [9].
Social interaction and religious involvement are independently related to suicide [23]. One study showed that the incidence of suicide decreases with increasing social integration, indicating that higher levels of social integration are associated with protection against suicide in females [24]. In our study, less contact with friends and greater religious activities were significantly related to the use of MHS for SI in females only. These results are difficult to explain. However, determining the extent of involvement in a range of social relationships may provide useful information for assessing suicidal risk and establishing a tailored strategy [24]. It is possible that females with a better social support system may seek help from people around them rather than seeking help through MHS, while males do not seek such help. Further research is needed on the gender-specific associations between socio-familial relationships and the use of MHS.
In this study, unhealthy behaviors, such as frequent drinking and lack of exercise, were significantly associated with not using MHS for SI in males, but not in females. Although not significant on multivariate analysis, the univariate analysis showed a significant association between current smoking and MHS use in males. To the best of our knowledge, no previous studies have examined gender-specific associations between health-related behaviors and MHS utilization by adults with SI. Unlike females, males who engage in unhealthy behaviors are more likely not to use MHS; thus, preventing suicide among males may require healthcare policies and societal concern to encourage the use of MHS for SI in males who are behaving in an unhealthy manner. Depression is an important risk factor for suicide [25]. Previous studies have shown that depression, psychiatric disorders, and psychiatric distress are important underlying factors for MHS use [8][9][10]26]. Our study found that a diagnosis of depression, experience of depressed mood, and suicide attempts were independently associated with the use of MHS in both genders, of which depression was the most potent factor in the use of MHS.
Some limitations should be considered when interpreting the results of this study. First, due to the cross-sectional design, this study could not derive causal relationships. Second, information on SI, suicide attempts, and the use of MHS was collected retrospectively, so recall bias may have occurred. Third, although socio-familial relationships were included in our analysis, the distribution of community resources such as medical institutions and community health centers, access to healthcare facilities, and regional cultural differences that affect the use of MHS were not included [13]. Fourth, attitudes and stigma about mental illness or MHS use were not evaluated. Promoting a positive attitude and reducing the stigma associated with the use of MHS in the general public may facilitate seeking the help of mental health professionals. Lastly, although this study included a large number of samples, statistical significance can be influenced by sample size and the variance of variables between the genders. Despite these limitations, in this study, we analyzed data from a national health survey and assessed a representative large-scale general population. In addition, multiple covariates, such as socioeconomic information, socio-familial relationships, health behaviors, and health status, were investigated simultaneously according to gender.

Conclusions
This study identified gender-specific factors associated with the use of MHS in the general Korean population with SI. These findings suggest that gender-specific factors should be used to inform suicide prevention strategies. Further studies are required to demonstrate genderspecific causal relationships between MHS utilization and related factors in individuals with SI.