Psychometric Characteristics of Duke Social Support Index Among Elderly Suicide in Rural China

Abstract

We recruited 242 elderly suicides and 242 controls above 60 years to conduct face-to-face interviews by psychological autopsy to examine the psychometric characteristics of the Duke Social Support Index (DSSI) in rural China. DSSI had high internal consistency, with Cronbach’s αs of .89 and .90 in suicides and controls, respectively. DSSI was significantly and negatively correlated to loneliness in both samples. Confirmatory factor analysis basically supported the original structure of DSSI, but Item 4 had low factor loading in controls. In conclusion, DSSI has satisfactory reliability and acceptable validity in evaluating social support in the elderly suicide study in China.

Keywords

Duke Social Support Index reliability validity suicide the elderly

Introduction

Suicide, an important public health and social problem, has aroused extensive concern in the world (Khazaei, Armanmehr, Nematollahi, & Rezaeian, 2017; Turecki & Brent, 2016). According to the report by World Health Organization in 2014, over 800,000 people, namely, one person every 40 seconds, die by suicide. China, where suicide was ranked as the fifth cause of death among the whole population (Phillips, Li, & Zhang, 2002), has made an overwhelming contribution to global suicide rate, accounting for approximately 22% of global suicides, or roughly 200,000 suicides every year (Hvistendahl, 2012). In China, suicide rate in rural area was 3 times higher than that in urban area. Although the difference in suicide rate between rural and urban area is increasingly smaller, the suicide rate of rural population is still higher than that of urban rate (Sun, Guo, Zhang, Jia, & Xu, 2013; Wang, Chan, & Yip, 2014; Zhang, Sun, Liu, & Zhang, 2014). The overall suicide rate has decreased apparently in recent years, while the rate of the elderly suicide increased (Wang et al., 2014). During 2013 and 2014, the elderly suicides accounted for 38.2% of overall suicides, while older adults shared only 8.9% of the total population (Zhong, Chiu, & Conwell, 2016). The average annual elderly suicide rate in China was 34.5 per 100,000 (Zhong et al., 2016). So, it is crucial to pay more attention to the elderly suicides in rural China.

Suicide among older adults is associated with physical health (Conwell, Duberstein, & Caine, 2002; Harwood, Hawton, Hope, Harriss, & Jacoby, 2006; Karbeyaz, Celikel, Emiral, & Emiral, 2017), mental disorder (Chiu et al., 2004; Karbeyaz et al., 2017; Wærn et al., 2002; Yeh, Ng, & Wu, 2017), and some socioeconomic factors (Crnek-Georgeson, Wilson, & Page, 2017; Karbeyaz et al., 2017; Liu, Qin, & Jia, 2018; Torresani, Toffol, Scocco, & Fanolla, 2014). Social support, as an important protective factor for various socioeconomic factors, has a significantly negative influence on suicide (Hollingsworth et al., 2018; Kleiman & Liu, 2013; Kleiman, Riskind, & Schaefer, 2014; Poudel-Tandukar et al., 2011; Zadravec Sedivy, Podlogar, Kerr, & De Leo, 2017). Zadravec Sedivy et al. (2017) conducted an ecological study in the relationship between social support and suicide rates in 75 regions of 23 European countries and found social support had a protective effect against suicide on a regional level for both genders, especially for males. Kleiman and Liu (2013) found social support was negatively correlated to the likelihood of a lifetime suicide attempt in two nationally representative samples: U.S. sample (odds ratio = 0.68, p < .001) and English sample (odds ratio = 0.93, p < .01), after controlling for a variety of related psychopathological and demographic variables. Hollingsworth et al. (2018) discovered that perceived burdensomeness mediated the relationship between perceived social support and suicide ideation and the relationship between social connectedness and suicide ideation in college students, which indicated a lack of social support could result in the feeling of being a burden to others, leading to suicide ideation further. In China, there were studies demonstrating that social support was negatively associated with suicide among the whole population (Zhang, Conwell, Zhou, & Jiang, 2004) and the older adults (Liu et al., 2018). In addition, enhancing the social support networks is one of the four steps in suicide prevention guidelines formulated by the United Nations (1996). All these elucidate that social support acts as an independent factor for suicide and an ameliorable factor to decrease suicide rate. So, it is important to measure the level of social support in suicide research studies.

Social support can be measured by various instruments, such as the Social Support Inventory with 27 items in two-part question (Sarason, Levine, Basham, & Sarason, 1983), the Multidimensional Scale of Perceived Social Support with 12 items in 7-point Likert-type format (Zimet, Dahlem, Zimet, & Farley, 1991), and the Social Support Network Inventory with 10 items in 6-point Likert-type response (Flaherty, Gaviria, & Pathak, 1983). But the Duke Social Support Index (DSSI; Koenig et al., 1993) and the Social Support Rating Scale (SSRS; Xiao, 1994) were often used to evaluate the social support in China. SSRS is a self-report questionnaire with 10 items in three dimensions: subjective support, objective support, and utility of support. It evaluated the social support from family, friends, neighbors, colleagues, and various organizations. It has satisfactory reliability and validity in various Chinese population (Li, 2017; Lu, Jia, Zhang, et al., 2011; Su, Wei, Ling, & Tang, 2009; Yang, Feng, & Xia, 2006). Due to its few items and being made on the basis of Chinese culture, it has been widely applied to explore the relationship between social support and suicide in many Chinese research studies (Chen, Ma, & Fan, 2008; Lu, Jia, & Ma, 2011; Sun & Jia, 2014). To our knowledge, this scale was few used in psychological autopsy (PA) research studies. And maybe it is inappropriate to make comparison with results from other countries by SSRS because it is more apt to Chinese environment. In addition, there have been few references about the comparison between the level of social support by SSRS and by other scales in other countries.

DSSI has been extensively used to study the relationship between social support and suicide (Hu, Jiang, & Liu, 2005; Zhang et al., 2003; Zhang & Zhou, 2009). Mao et al. (2015) found DSSI had satisfactory reliability and acceptable validity, and it was better to adjust several items in the older adults by stratified cluster random sampling strategy in a rural elderly population. Jia and Zhang (2012) examined the psychometric characteristics of DSSI in young rural Chinese suicides and controls both aged 15 and 34 years and got similar conclusion. However, these two research studies are either young suicides or general elder population, and there are not much related studies on the examination of DSSI among completed suicides in rural and elderly Chinese. So, this study recruited 242 completed suicides and 242 matched controls both aged 60 years and older to measure the psychometric characteristics of DSSI among elderly suicides and controls in rural China, further to pave a way for the research studies about the relationship between suicide and social support or other-related factors.

Method

Subjects

Based on 1:1 matched case-control design, 242 elderly cases and 242 controls were recruited. Cases were the older adults who were aged 60 years or above and conducted completed suicide during October 8, 2011 to September 5, 2015. Controls were matched on gender, age (±3 years), same or neighboring community, and without suicide history. All suicides and controls were chosen from 12 rural counties from disease surveillance points of Shandong, Hunan, and Guangxi province, which were located in Eastern China, South Central China, and Southern China, respectively. According to death certification system, the project cooperators from the Center for Disease Control and Prevention of each county helped identify and contact cases, with the assistance of village doctors or leaders. They were also responsible for finding suitable controls from the same or near village to match suicides.

Procedures

We conducted face-to-face interview with control himself and two informants of each control and suicide by PA, which was a scientific and systematic method to reconstruct suicidal deaths through informants (Beskow, Runeson, & Asgard, 1991; Isometsa, 2001). The first informant was usually one next of kin, for example, a spouse, a parent, a son, or a daughter. The second informant was always a friend, a neighbor, a colleague, or a remote relative. PA, widely used in suicide studies in China, has been proved to have good validity and reliability (Phillips, Yang, et al., 2002; Yang et al., 2005; Zhang et al., 2002, 2003; Zhou, Jiang, & Zhang, 2006). Interviews were conducted between 2 and 6 months after suicide in order to try to avoid bereavement reaction and minimize recall bias (Conner et al., 2012). After the written informed consent, trained interviewers began interview in a separated room of the village clinic or the interviewee’s home. The average interview time was about 3 hours.

Instruments

Demographic characteristics

Demographic characteristics included gender, age, education (illiteracy, primary school, middle school, high school and above), marital status (stable vs. unstable; stable marital status refers to being married and living with a spouse and unstable marital status refers to being never married, divorced, widowed, or not living with a spouse), employment (employed, unemployed, and retired), family annual income (≤3,600 RMB, 3,600∼10,000 RMB, and ≥10,000 RMB, according to the tertiles), living alone (yes vs. no), religion belief (yes vs. no), superstition (yes vs. no), and family suicide history (yes vs. no).

Duke Social Support Index

The DSSI (Koenig et al., 1993) is a self-report questionnaire to assess social support, including 23 items in three dimensions. There are 4 items expressing social interaction (SIS), 7 items demonstrating subjective social support (SSS), and 12 items stating instrumental social support (ISS). SIS and SSS are designed on a 3-point Likert-type format, while ISS is dichotomous. The total score varies from 11 to 45, with higher scores reflecting more social support. The corrected Cronbach’s α of the University of California Los Angeles Loneliness Scale-6 (ULS-6) was .89 in completed suicides and .90 in controls.

Simplified ULS

The initial ULS-20 (Stephan, Fath, & Lamm, 1988) was deleted 12 items and became ULS-8 (Russell, Peplau, & Cutrona, 1980). The psychometric properties of the ULS had been widely evaluated (Neto, 2014; Russell et al., 1980; Swami, 2009; Wilson, Cutts, Lees, Mapungwana, & Maunganidze, 1992; Wu & Yao, 2008). Zhou, Li, Hu, and Xiao (2012) evaluated it among the older adults in rural China and deleted two items further. The six items are all positive loneliness, on a 4-point Likert-type, from 1 (never) to 4 (always). The range of total score is from 6 to 24, and higher score means more serious loneliness. The ULS-6 has been assessed to have good reliability and validity in Chinese culture (Zhou et al., 2012). In this study, the corrected Cronbach’s α of the ULS-6 was .94 in completed suicides and .91 in controls.

Integration of information

The responses from two informants were largely similar. However, there were also inevitable differences, so it was necessary to integrate different information. For the demographic characteristics and suicide behavior, we relied on the information provided by the first informant. For the diagnosis of mental disease, we tended to adopt the positive response. For the DSSI and ULS-6, we were inclined to use lower score of DSSI and higher score of ULS-6.

Data analysis

IBM SPSS Statistics 24.0 (online edition) and AMOS 24.0 were used for statistical analyses. t Tests or Kolmogorov–Smirnov Z tests were used to compare continual variables. χ² tests were used to compare categorical variables. The Cronbach’s α was conducted to test internal consistency of DSSI. Spearman correlations between ULS-6 and three dimensions of DSSI and the whole DSSI were examined for criterion-related validity. Conﬁrmatory factor analysis was used to assess construct validity. Models were evaluated by χ², Goodness-of-Fit Index, Adjusted Goodness-of-Fit Index, Comparative Fit Index, and root mean square error of approximation (Carleton et al., 2013).

Ethical issues

This study was approved by the institutional review board of Shandong University, Central South University, and Guangxi Medical University. All experiments involving human participants were performed in accordance with relevant guidelines and regulations. Trained interviewers explained the aim and procedure of the research and other information that interviewees want to know to all participants. Written informed consent was obtained from all participants before the interviews.

Results

Demographic Characteristics

Among 242 suicides and 242 controls, 135 (55.8%) of them were male for each group. The mean age was 74.4 (standard deviation [SD] = 8.2) years in suicides and 74.1 (SD = 8.2) years in controls.

As shown in Table 1, compared with controls, suicides were more likely to be unemployed, living alone, superstitious, and to have unstable marital status, less social support, and more feeling of loneliness. However, there were no statistically significant differences in education, family annual income, religion, and family suicide history between suicides and controls.

Table 1.

Demographic Characteristics in Suicide and Control Samples.

Variables	Suicide sample	Control sample	t/z/χ²	p
Variables	n(%)/M(SD)	n(%)/M(SD)	t/z/χ²	p
Education
Illiteracy	111 (45.9)	96 (39.7)	2.015	.569
Primary school	105 (43.3)	116 (47.9)
Middle school	20 (8.3)	22 (9.1)
High school and above	6 (2.5)	8 (3.3)
Marital status
Stable marital status	122 (50.4)	170 (70.2)	19.890	<.001
Unstable marital status	120 (49.6)	72 (29.8)
Employment
Employed	41 (16.9)	61 (25.2)	8.274	.016
Unemployed	194 (80.2)	167 (69.0)
Retired	7 (2.9)	14 (5.8)
Family annual income^a
≤3,600	88 (36.4)	74 (30.6)	1.865	.394
3,600∼10,000	88 (36.4)	98 (40.5)
≥10,000	66 (27.3)	70 (28.9)
Living alone
Yes	64 (26.4)	35 (14.5)	10.679	.001
No	178 (73.6)	207 (85.5)
Religion
Yes	12 (5.0)	10 (4.1)	0.190	.663
No	230 (95.0)	232 (95.9)
Superstition
Yes	36 (14.9)	15 (6.2)	9.666	.002
No	206 (85.1)	227 (93.8)
Family suicide history
Yes	37 (15.3)	21 (8.7)	5.015	.025
No	205 (84.7)	221 (91.3)
Social support	22.9 (6.0)	27.5 (6.8)	3.500	<.001
Loneliness	15.6 (5.0)	11.0 (4.0)	4.545	<.001

Note. SD = standard deviation; M = mean.

^aFamily annual income: 3,600 was the 33.3 percentile; 10,000 was the 66.67 percentile.

Reliability

Table 2 showed mean, SD, corrected item-total correlation, and Cronbach’s α coefficients if item deleted of DSSI in both suicide and control samples. The corrected Cronbach’s α of the DSSI was .89 in completed suicides and .90 in controls. In the suicide sample, the Cronbach’s α was .54 for SIS, .89 for SSS, and .92 for ISS. In the control sample, the Cronbach’s α was .54 for SIS, .92 for SSS, and .91 for ISS. Cronbach’s α coefficients if item deleted ranged from .88 to .89 in suicides and varied from .88 to .90 in controls.

Table 2.

Mean, SD, Corrected Item-Total Correlation, and Cronbach’s α If Item Deleted of Duke Social Support Index in Suicide and Control Samples.

Item	Suicide sample			Control sample
Item	M(SD)	r	α	M(SD)	r	α
Social interaction
Item 1	1.62 (0.63)	.225	.892	1.88 (0.66)	.277	.897
Item 2	1.48 (0.59)	.195	.892	1.84 (0.64)	.376	.894
Item 3	1.17 (0.41)	.224	.888	1.47 (0.61)	.405	.893
Item 4	1.12 (0.33)	−.056	.893	1.21 (0.49)	.054	.900
Subjected support
Item 5	1.26 (0.53)	.489	.882	1.59 (0.68)	.711	.883
Item 6	1.20 (0.48)	.424	.884	1.56 (0.68)	.681	.884
Item 7	1.28 (0.54)	.531	.881	1.64 (0.71)	.733	.883
Item 8	1.20 (0.50)	.554	.880	1.47 (0.68)	.685	.884
Item 9	1.23 (0.51)	.535	.881	1.49 (0.69)	.709	.883
Item 10	1.23 (0.48)	.530	.881	1.65 (0.70)	.663	.885
Item 11	1.69 (0.65)	.465	.884	2.23 (0.69)	.437	.892
Instrumental support
Item 12	0.83 (0.38)	.547	.881	0.93 (0.26)	.409	.893
Item 13	0.81 (0.40)	.623	.880	0.87 (0.34)	.462	.891
Item 14	0.69 (0.47)	.579	.880	0.76 (0.43)	.431	.892
Item 15	0.80 (0.40)	.579	.880	0.86 (0.35)	.507	.891
Item 16	0.68 (0.47)	.567	.880	0.74 (0.44)	.429	.892
Item 17	0.69 (0.46)	.579	.880	0.69 (0.46)	.447	.891
Item 18	0.54 (0.50)	.553	.880	0.60 (0.49)	.489	.890
Item 19	0.68 (0.47)	.578	.880	0.83 (0.37)	.466	.891
Item 20	0.56 (0.50)	.601	.879	0.76 (0.43)	.557	.889
Item 21	0.60 (0.49)	.642	.878	0.77 (0.42)	.578	.889
Item 22	0.75 (0.43)	.576	.880	0.80 (0.40)	.526	.890
Item 23	0.79 (0.41)	.550	.881	0.85 (0.36)	.397	.892

Note. SD = standard deviation; M = mean.

Validity

Table 3 presented that the total DSSI and three subscales were all negatively related to loneliness in male, female, and total suicide and control samples, which verified satisfactory criterion validity.

Table 3.

Correlation Between Scores of Loneliness and Subscales of DSSI and Total DSSI Among Male and Female in Suicide and Control Samples.

	Suicide sample			Control sample
	Total	Male	Female	Total	Male	Female
Loneliness and SIS	−0.214**	−0.283**	−0.110	−0.255**	−0.285**	−0.217*
Loneliness and SSS	−0.372**	−0.491**	−0.209**	−0.421**	−0.415**	−0.437**
Loneliness and ISS	−0.352**	−0.337**	−0.377**	−0.360**	−0.326**	−0.409**
Loneliness and DSSI	−0.443**	−0.495**	−0.367**	−0.466**	−0.457**	−0.489**

Note. DSSI = Duke Social Support Index; SIS = social interaction; ISS = instrumental social support; SSS = subjective social support.

*p < .05. **p < .01.

As shown in Table 4, conﬁrmatory factor analysis was conducted to confirm the three-factor model. Goodness-of-Fit Index, Adjusted Goodness-of-Fit Index, and Comparative Fit Index were approximately or above 0.80 for both suicide and control samples. The values of root mean square error of approximation were about 0.09 among suicides and controls. Factor loadings in every factor were 0.31–0.80, 0.48–0.86, and 0.60–0.79 in suicides and 0.28–0.73, 0.49–0.88, and 0.55–0.79 in controls, respectively. Apparently, Item 4 had lower factor loadings in both suicides and controls. These results basically fitted the original structure of DSSI.

Table 4.

Factor Loadings of Each Item in SIS, SSS, and ISS of Duke Social Support Index From Confirmatory Factor Analysis in Suicide and Control Samples.

Item	Suicide sample			Control sample
Item	SIS	SSS	ISS	SIS	SSS	ISS
Item 1	0.50			0.57
Item 2	0.80			0.73
Item 3	0.32			0.37
Item 4	0.31			0.28
Item 5		0.71			0.82
Item 6		0.68			0.82
Item 7		0.80			0.84
Item 8		0.86			0.85
Item 9		0.86			0.88
Item 10		0.80			0.80
Item 11		0.48			0.49
Item 12			0.72			0.58
Item 13			0.79			0.66
Item 14			0.63			0.55
Item 15			0.72			0.68
Item 16			0.77			0.73
Item 17			0.79			0.72
Item 18			0.60			0.67
Item 19			0.73			0.67
Item 20			0.67			0.69
Item 21			0.72			0.79
Item 22			0.70			0.67
Item 23			0.70			0.64
GFI	0.795			0.816
AGFI	0.751			0.776
CFI	0.835			0.854
χ²	708.459			655.302
df	227			227
p	<.001			<.001
RMSEA	0.094			0.088

Note. SIS = social interaction; ISS = instrumental social support; SSS = subjective social support; GFI = Goodness-of-Fit Index; AGFI = Adjusted Goodness-of-Fit Index; CFI = Comparative Fit Index; RMSEA = root mean square error of approximation.

Discussion

This study aimed to assess the reliability and validity of DSSI in the elderly suicides and controls in rural China. The findings indicated that DSSI had satisfactory reliability and acceptable validity in Chinese culture, which was consistent with the previous study (Jia & Zhang, 2012). In terms of demographic characteristics, some differences between suicides and controls were statistically significant, which was also similar to the previous findings (Liu et al., 2018; Zhang et al., 2004).

Many studies have proved that DSSI has good internal consistency in various Chinese populations. For instance, the Cronbach’s α coefficient was .84 for suicide group and .79 for control group in young rural Chinese population (Jia & Zhang, 2012). Another study showed high Cronbach’s α of .88 in a rural elderly population (Mao et al., 2015). In this study, the Cronbach’s α coefficients were both above .89 in suicides and controls, indicating good internal consistency and satisfactory reliability of DSSI in the elderly suicide research in rural China.

The relationship between social support and loneliness has been tested (O’Donovan & Hughes, 2007; Wright, 2005; Yildirim & Kocabiyik, 2010; Zhang, Gao, Fokkema, Alterman, & Liu, 2015). So, it is suitable to use loneliness to assess the concurrent validity of DSSI. In this study, the ULS-6 was adopted, which had high degree of internal consistency and validity in Chinese culture (Zhou et al., 2012). In the total samples, both in suicide group and control group, loneliness was statistically correlated to SIS, SSS, ISS, and DSSI, indicating DSSI had good concurrent validity. Most of results in Table 3 were statistically significant (p < .01), whereas there was no statistical significance in the relationship between loneliness and SI among female suicide sample (p > .05) and no strong statistical significance among female control sample (p < .05). Maybe it is because majority of Chinese females pay most attention to taking care of family and have few friends and few interactions with the world outside family after their marriage. But they are pleased to do so and do not feel lonely.

The three-factor model of DSSI had an acceptable goodness of fit in the suicide and control samples in rural China. In addition, the factor loadings of most items were high, except Item 4 (0.31 in suicide sample, 0.28 in control sample). Item 4 was how often did you go to meetings of clubs, religious meetings, or other groups that you belonged to in the past week or in the past week before suicide. It could be due to cultural diversity. Different from Western countries, the rural Chinese might not understand the meaning of club and group and seldom participated in collective activities.

Despite the reliability and validity of DSSI had been reported in China, this was the first study to assess DSSI in suicide research by PA among the older adults in rural China. It paved the way for the further study of the relationship between the elderly suicide and social support and of other risk factors in the elderly suicides in rural China. What’s more, samples in this study were recruited from three provinces of China, making the research more representative.

There are also some limitations in this study. First, in the light of PA, some data were collected from informants. Although the reliability and validity of PA has been proved in previous studies (Phillips et al., 2002; Zhang et al., 2002, 2003; Zhou et al., 2006), the information bias is inevitable. Second, in consideration of interview time and expenditure, we did not add SSRS in the questionnaire and did not conduct test–retest reliability. ULS-6 as a criterion for concurrent validity might not be appreciated and it might be better to use SSRS.

Footnotes

Acknowledgments

The authors are very grateful for the support of Shandong University, Central South University, and Guangxi Medical University. The authors also appreciate all the Center for Disease Control and Prevention cooperators, village doctors or directors, interviewers, and interviewees from Shandong, Hunan, and Guangxi.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the American Foundation of Suicide Prevention to Dr. Liang Zhou under Grant No. SRG-0–169-12.

Author Biographies

Yan-Fei Pan, is a graduate student in Shandong University. She is mainly engaged in suicide research.

Zhen-Yu Ma, PhD, is an associate professor in Guangxi Medical University. Her research interests are league studies and injury research.

Liang Zhou, is a professor in the Affiliated Brain Hospital of Guangzhou Medical University (Guangzhou Huiai Hospital). He mainly focuses on research of mental illness and suicide.

Cun-Xian Jia, PhD, is a professor of epidemiology at Shandong University School of Public Health. He obtained his Master degree in 1998 in Shandong Medical University and his PhD degree in 2004 in Shandong University. His research areas are injury (suicide) epidemiology and adolescent health.

References

Beskow

Runeson

Asgard

(1991). Ethical aspects of psychological autopsy. Acta Psychiatrica Scandinavica, 84(5), 482.

Carleton

R. N.

Thibodeau

M. A.

Teale

M. J. N.

Welch

P. G.

Abrams

M. P.

Robinson

Asmundson

G. J. G.

(2013). The Center for Epidemiologic Studies Depression Scale: A review with a theoretical and empirical examination of item content and factor structure. PLoS One, 8(3), e58067.

Chen

L. F.

S. B.

Fan

C. X.

(2008). Social support and it’s effects on suicidal ideation among undergraduates. Chinese Journal of Public Health, 24(8), 925–926.

Chiu, H. F., Yip, P. S., Chi, I., Chan, S., Tsoh, J., Kwan, C. W., Li, S. F., Conwell, Y., Caine, E. (2004). Elderly suicide in Hong Kong—A case-controlled psychological autopsy study. Acta Psychiatrica Scandinavica, 109(4), 299–305.

Conner

K. R.

Beautrais

A. L.

Brent

D. A.

Conwell

Phillips

M. R.

Schneider

(2012). The next generation of psychological autopsy studies: Part 2. Interview procedures. Suicide and Life-Threatening Behavior, 42(1), 86–103.

Conwell

Duberstein

P. R.

Caine

E. D.

(2002). Risk factors for suicide in later life. Biological Psychiatry, 52(3), 193–204.

Crnek-Georgeson

K. T.

Wilson

L. A.

Page

(2017). Factors influencing suicide in older rural males: A review of Australian studies. Rural Remote Health, 17(4), 4020.

Flaherty

J. A.

Gaviria

F. M.

Pathak

D. S.

(1983). The measurement of social support: The Social Support Network Inventory. Comprehensive Psychiatry, 24(6), 521–529.

Harwood

D. M.

Hawton

Hope

Harriss

Jacoby

(2006). Life problems and physical illness as risk factors for suicide in older people: A descriptive and case-control study. Psychological Medicine, 36(9), 1265–1274.

10.

Hollingsworth

D. W.

Slish

M. L.

Wingate

L. R.

Davidson

C. L.

Rasmussen

K. A.

O’Keefe

V. M.

, . . . Grant

D. M.

(2018). The indirect effect of perceived burdensomeness on the relationship between indices of social support and suicide ideation in college students. Journal of American College Health, 66(1), 9–16.

11.

D. M.

Jiang

Liu

Q. G.

(2005). A psychological autopsy study on rural suicides using conditional logistic regression model. Chinese Mental Health Journal, 19(2), 131–133.

12.

Hvistendahl

(2012). Making sense of a senseless act. Science, 338, 1025–1027.

13.

Isometsa

E. T.

(2001). Psychological autopsy studies—A review. European Psychiatry the Journal of the Association of European Psychiatrists, 16(7), 379.

14.

Jia

C. X.

Zhang

(2012). Psychometric characteristics of the Duke Social Support Index in a young rural Chinese population. Death Studies, 36(9), 858–869.

15.

Karbeyaz

Celikel

Emiral

G. O.

(2017). Elderly suicide in Eskisehir, Turkey. Journal of Forensic and Legal Medicine, 52, 12–15.

16.

Khazaei

Armanmehr

Nematollahi

Rezaeian

(2017). Suicide rate in relation to the Human Development Index and other health related factors: A global ecological study from 91 countries. Journal of Epidemiology and Global Health, 7(2), 131–134.

17.

Kleiman

E. M.

Liu

R. T.

(2013). Social support as a protective factor in suicide: Findings from two nationally representative samples. Journal of Affective Disorders, 150(2), 540–545.

18.

Kleiman

E. M.

Riskind

J. H.

Schaefer

K. E.

(2014). Social support and positive events as suicide resiliency factors: Examination of synergistic buffering effects. Archives of Suicide Research, 18(2), 144–155.

19.

Koenig

H. G.

Westlund

R. E.

George

L. K.

Hughes

D. C.

Blazer

D. G.

Hybels

(1993). Abbreviating the Duke Social Support Index for use in chronically ill elderly individuals. Psychosomatics, 34(1), 61–69.

20.

Y. Y.

(2017). The reliability and validity of Social Support Rating Scale in firemen. China Journal of Health Psychology, 25(9), 1342–1345.

21.

Liu

B. P.

Qin

Jia

C. X.

(2018). Behavior characteristics and risk factors for suicide among the elderly in rural China. Journal of Nervous and Mental Disease, 206, 195–201.

22.

C. F.

Jia

C. X.

J. X.

(2011). Social support and rural suicides-a matched case-control study. Chinese Journal of Public Health, 27(3), 312–314.

23.

C. F.

Jia

C. X.

Zhang

J. Y.

Wang

X. X.

Wang

D. F.

Liu

L. Y.

(2011). Reliability and validity of Social Support Rating Scale in rural suicides. Chinese Mental Health Journal, 25(3), 218–222.

24.

Mao

Qiu

P. Y.

Wen

Yang

Hao

T. Y.

Gao

X. Q.

Yuan

(2015). Reliability and validity of DSSI-23 Scale in a rural elderly population. Journal of Sichuan University (Medical Sciences), 46(1), 90–93.

25.

Neto

(2014). Psychometric analysis of the short-form UCLA Loneliness Scale (ULS-6) in older adults. European Journal of Ageing, 11(4), 313–319.

26.

O’Donovan

Hughes

(2007). Social support and loneliness in college students: Effects on pulse pressure reactivity to acute stress. International Journal of Adolescent Medical Health, 19(4), 523–528.

27.

Phillips

M. R.

Zhang

(2002). Suicide rates in China, 1995-99. Lancet, 359(9309), 835–840.

28.

Phillips

M. R.

Yang

Zhang

Wang

Zhou

(2002). Risk factors for suicide in China: A national case-control psychological autopsy study. Lancet, 360(9347), 1728–1736.

29.

Poudel-Tandukar, K., Nanri, A., Mizoue, T., Matsushita, Y., Takahashi, Y., Noda, M., Inoue, M., Tsugane, S. (2011). Social support and suicide in Japanese men and women—The Japan Public Health Center (JPHC)-based prospective study. Journal of Psychiatric Research, 45(12), 1545–1550.

30.

Russell

Peplau

L. A.

Cutrona

C. E.

(1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39(3), 472.

31.

Sarason

I. G.

Levine

H. M.

Basham

R. B.

Sarason

B. R.

(1983). Assessing social support: The Social Support Questionnaire. Journal of Personality and Social Psychology, 44(44), 127–139.

32.

Stephan

Fath

Lamm

(1988). Loneliness as related to various personality and environmental measures: Research with the German adaptation of the UCLA Loneliness Scale. Social Behavior and Personality, 16(16), 169–174.

33.

Wei

Ling

X. F.

Tang

J. H.

(2009). Study on the reliability, validity and norm of Social Support Scale in Chuang peasants. Modern Preventive Medicine, 36(23), 4411–4413.

34.

Sun

Guo

Zhang

Jia

(2013). Suicide rates in Shandong, China, 1991-2010: Rapid decrease in rural rates and steady increase in male-female ratio. Journal of Affective Disorders, 146(3), 361–368.

35.

Sun

S. H.

Jia

C. X.

(2014). Completed suicide with violent and non-violent methods in rural Shandong, China: A psychological autopsy study. PLoS One, 9(8), e104333.

36.

Swami

(2009). Psychometric analysis of the Malay version of the UCLA Loneliness Scale (ULS-8) and a comparison of loneliness among sojourning and non-sojourning Malaysian students. International Journal of Culture and Mental Health, 2(1), 38–50.

37.

Torresani

Toffol

Scocco

Fanolla

(2014). Suicide in elderly South Tyroleans in various residential settings at the time of death: A psychological autopsy study. Psychogeriatric Society, 14(2), 101–109.

38.

Turecki

Brent

D. A.

(2016). Suicide and suicidal behaviour. Lancet, 387(10024), 1227–1239.

39.

United Nations. (1996). Prevention of suicide: Guidelines for the formulation and implementation of national strategies. New York, NY: United Nations.

40.

Wærn

Runeson

B. S.

Allebeck

Beskow

Rubenowitz

Skoog

Wilhelmsson

(2002). Mental disorder in elderly suicides: A case-control study. American Journal of Psychiatry, 159(3), 450–455.

41.

Wang

C. W.

Chan

C. L.

Yip

P. S.

(2014). Suicide rates in China from 2002 to 2011: An update. Social Psychiatry and Psychiatric Epidemiology, 49(6), 929–941.

42.

Wilson

Cutts

Lees

Mapungwana

Maunganidze

(1992). Psychometric properties of the revised UCLA Loneliness Scale and two short-form measures of loneliness in Zimbabwe. Journal of Personality Assessment, 59(1), 72.

43.

World Health Organization. (2014). Preventing suicide: A global imperative. Geneva, Switzerland: World Health Organization.

44.

Wright

S. L.

(2005). Organizational climate, social support and loneliness in the workplace. Research on Emotion in Organizations, 1(1), 123–142.

45.

C. H.

Yao

(2008). Psychometric analysis of the short-form UCLA Loneliness Scale (ULS-8) in Taiwanese undergraduate students. Personality and Individual Differences, 44(8), 1762–1771.

46.

Xiao

S. Y.

(1994). The Social Support Rating Scale and its applications. Journal of Clinical Psychological Medicine, 4(2), 98–100.

47.

Yang

G. H.

Phillips

M. R.

Zhou

M. G.

Wang

L. J.

Zhang

Y. P.

(2005). Understanding the unique characteristics of suicide in China: National psychological autopsy study. Biomedical and Environmental Sciences, 18(6), 379–389.

48.

Yang

G. Y.

Feng

Z. Z.

Xia

B. L.

(2006). The reliability, validity and national norm of Social Support Scale in servicemen. Chinese Mental Health Journal, 20(5), 309–312.

49.

Yeh

S. T.

Y. Y.

S. C.

(2017). Risk of suicide according to the level of psychiatric contact in the older people: Analysis of national health insurance databases in Taiwan. Comprehensive Psychiatry, 74, 189–195.

50.

Yildirim

Kocabiyik

(2010). The relationship between social support and loneliness in Turkish patients with cancer. Journal of Clinical Nursing, 19(5-6), 832–839.

51.

Zadravec Sedivy

Podlogar

Kerr

D. C. R.

De Leo

(2017). Community social support as a protective factor against suicide: A gender-specific ecological study of 75 regions of 23 European countries. Health Place, 48, 40–46.

52.

Zhang

Gao

Fokkema

Alterman

Liu

(2015). Adolescent interpersonal relationships, social support and loneliness in high schools: Mediation effect and gender differences. Social Science Research, 53, 104–117.

53.

Zhang

Conwell

Wieczorek

W. F.

Jiang

Jia

Zhou

(2003). Studying Chinese suicide with proxy-based data: Reliability and validity of the methodology and instruments in China. Journal of Nervous and Mental Disease, 191(7), 450–457.

54.

Zhang

Conwell

Zhou

Jiang

(2004). Culture, risk factors and suicide in rural China: A psychological autopsy case control study. Acta Psychiatrica Scandinavica, 110(6), 430.

55.

Zhang

Sun

Liu

Zhang

(2014). The change in suicide rates between 2002 and 2011 in China. Suicide and Life-Threatening Behavior, 44(5), 560–568.

56.

Zhang, J., Wieczorek, W. F., Jiang, C., Zhou, L., Jia, S., Sun, Y., Jin, S., Conwell, Y. (2002). Studying suicide with psychological autopsy: Social and cultural feasibilities of the methodology in China. Suicide and Life-Threatening Behavior, 32(4), 370–379.

57.

Zhang

Zhou

(2009). A case control study of suicides in China with and without mental disorder. Crisis, 30(2), 68–72.

58.

Zhong

B. L.

Chiu

H. F.

Conwell

(2016). Rates and characteristics of elderly suicide in China, 2013-14. Journal of Affective Disorders, 206, 273–279.

59.

Zhou

Jiang

Zhang

(2006). The applicability of psychological autopsy for suicide research in China. Chinese Mental Health Journal, 20(1), 19–22.

60.

Zhou

Xiao

(2012). Reliability and validity of ULS-8 Loneliness Scale in elderly samples in a rural community. Journal of Central South University (Medical Sciences), 37(11), 1124.

61.

Zimet

G. D.

Dahlem

N. W.

Zimet

S. G.

Farley

G. K.

(1991). The Multidimensional Scale of Perceived Social Support. Journal of Clinical Psychology, 47(6), 756–761.