OLDER ADULTS AND THE DESIRE FOR SUICIDE: THE ROLE OF
PERCEIVED BURDENSOMENESS AND THWARTED BELONGINGNESS
by
MEGHAN A. MARTY
B.A., Rice University, 1999
M.A., Ed.M., Teachers College, Columbia University, 2004
A dissertation submitted to the Graduate Faculty of the
University of Colorado at Colorado Springs
in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
Department of Psychology
2011
This dissertation for Doctor of Philosophy degree by
Meghan A. Marty
has been approved for the
Department of Psychology
by
______! Daniel L. Segal, Chair
______! Frederick L. Coolidge
______! Molly Maxfield
______! Aditi Mitra
______! Sara H. Qualls
______Date Marty, Meghan A. (Ph.D., Psychology)
Older Adults and the Desire for Suicide: The Role of Perceived Burdensomeness and
Thwarted Belongingness
Dissertation directed by Associate Professor Daniel L. Segal
The Interpersonal-Psychological Theory of Suicide (IPTS) predicts that in order for individuals to complete suicide, they must have both the desire for suicide, comprised of two components named thwarted belongingness and perceived burdensomeness, and the ability to carry out the act, labeled acquired capability. Mounting evidence supports the tenets of the IPTS; however, little research has tested this theory among older adults, a population at the highest risk for suicide. The purposes of this study were to evaluate the structure and validity of the Interpersonal Needs Questionnaire (INQ), a measure of burdensomeness and belongingness, and to investigate the suicidal desire component of the IPTS among older adults. Acquired capability was not examined in the present study.
Older adult participants (N = 284, M age = 73.3 years, SD = 7.1, age range = 64 – 96,
55.6% female) were recruited through a mailing. They completed anonymously the INQ,
Geriatric Suicide Ideation Scale, Patient Health Questionnaire, Sense of Belonging
Instrument, Beck Hopelessness Scale, Single-Item Self-Esteem Scale, Three-Item
Loneliness Scale, RAND 36-Item Health Survey 1.0, and Short Coolidge Axis II
Inventory. Principle components analysis supported a two-component INQ; however, confirmatory factor analysis indicated several models were not a good fit for the data.
Correlations provided evidence for the validity of the INQ, but the distinction between burdensomeness and belongingness warrants further investigation. The interaction ! iv between burdensomeness and belongingness contributed significant variance in most of the regression analyses, such that individuals who reported high perceived burdensomeness and high thwarted belongingness also reported high current suicide ideation and past suicide attempt. The present study provides provisional support for the structure and validity of the INQ and suggests that the joint presence of burdensomeness and belongingness is associated with suicide among older adults. ! v
ACKNOWLEDGEMENTS
I would like to thank my dissertation committee members for their insightful comments and suggestions; my research lab mates for their encouragement throughout the project and assistance with the preparation of the mailings; the study participants for their willingness to share their thoughts and feelings; the UCCS Copy Center and Mailroom for their logistical expertise; and Julio Barros for his support, encouragement, and technical assistance. !
TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION …………………………………………………… 1
Introduction to the Interpersonal-Psychological Theory of Suicide ……………………………………………………… 4
Empirical Support for the Interpersonal-Psychological Theory of Suicide …………………………………………………… 8
The Interpersonal-Psychological Theory of Suicide and High Risk Populations ……………………………………………. 13
The Interpersonal-Psychological Theory of Suicide and Protective Factors …………………………………………... 16
Belongingness and Suicide in Older Adults ……………………. 17
Burdensomeness and Suicide in Older Adults …………………. 23
The Present Study ………………………………………………. 28
II. METHOD …………………………………………………………… 31
Participants ……………………………………………………... 31
Measures ………………………………………………………... 32
Interpersonal Needs Questionnaire (INQ) ………………... 32
Geriatric Suicide Ideation Scale (GSIS) ………………….. 34
Patient Health Questionnaire (PHQ-9) …………………… 35
Sense of Belonging Instrument (SOBI) …………………... 35
Beck Hopelessness Scale (BHS) …………………………. 36
! ! vii
Single-Item Self-Esteem Scale (SISE) …………………… 37
Three-Item Loneliness Scale (3LS) ………………………. 37
RAND 36-item Health Survey 1.0 (SF-36) ………………. 38
Short Coolidge Axis II Inventory (SCATI) ………………. 38
Procedure ……………………………………………………….. 39
III. RESULTS ………………………………………………………….. 42
Exploratory Principle Components Analysis …………………... 42
Confirmatory Factor Analysis ………………………………….. 49
Construct Validity Analysis ……………………………………. 50
Hierarchical Multiple Regression Analysis for Current Suicide Ideation ……………………………………………………... 58
Model 2 Predicting Current Suicide Ideation …………….. 60
Model 3 Predicting Current Suicide Ideation …………….. 61
Model 4 Predicting Current Suicide Ideation …………….. 63
Validity Measures Predicting Current Suicide Ideation ….. 64
Exploratory Regression Analyses ………………………... 66
Hierarchical Multiple Regression Analysis for Past Suicide Attempt ……………………………………………………... 66
Model 2 Predicting Past Suicide Attempt ………………... 68
Model 3 Predicting Past Suicide Attempt ………………... 69
Model 4 Predicting Past Suicide Attempt ………………... 70
Validity Measures Predicting Past Suicide Attempt ……... 72
IV. DISCUSSION …………………………………………………….... 74
Structure and Validity of the Interpersonal Needs Questionnaire ………………………………………………. 74
! ! viii
Suicidal Desire Component of the Interpersonal-Psychological Theory of Suicide …………………………………………... 78
Limitations and Strengths ………………………………………. 80
Research Implications and Future Research …………………… 82
Clinical Implications and Future Research …………………….. 85
Conclusion ……………………………………………………… 87
REFERENCES ………………………………………………………………….... 89
APPENDICES …………………………………………………………………..... 98
A. INSTITUTIONAL REVIEW BOARD APPROVAL …………….....9 8
B. INFORMED CONSENT FORM ………………………………….... 99
! TABLES
Table
1. Hypotheses of the Interpersonal-Psychological Theory of Suicide ……. 7
2. Measurement of Study Variables ……………………………………..... 29
3. Participant Characteristics ……………………………………………… 31
4. Psychometric Properties of the Major Study Variables ………………... 40
5. Pearson Correlations between Suicide Ideation, Past Suicide Attempt, and Demographic, Clinical, and Personality Variables …………….. 43
6. Inter-item Correlation Matrix for the Interpersonal Needs Questionnaire ……………………………………………………….. 44
7. Component Correlations for Four-Component Interpersonal Needs Questionnaire ……………………………………………………….. 45
8. Component Loadings for Exploratory Principle Components Analysis With Varimax Rotation …………………………………………….. 47
9. Item Loadings for Different Versions of the INQ ……………………… 48
10. Psychometric Properties of Different Versions of the INQ ……………. 49
11. Confirmatory Factor Analysis for the Interpersonal Needs Questionnaire ……………………………………………………….. 50
12. Pearson Correlations Between Belonging, Burden, Validity Measures, and Predicted Variables …………………………………………….. 53
13. Pearson Correlations Between Belonging, Burden, and Demographic and Clinical Variables ……………………………………………… 55
14. Pearson Correlations Between Belonging, Burden, and Personality Disorder Features …………………………………………………... 56
! "! ! 15. Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 2 Belonging and Burden Measures ………… 61
16. Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 3 Belonging and Burden Measures ………… 63
17. Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 4 Belonging and Burden Measures ………… 64
18. Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using SOBI-P and GSIS LPSW Measures ……………….. 65
19. Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 2 Belonging and Burden Measures ………… 69
20. Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 3 Belonging and Burden Measures ………… 71
21. Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 4 Belonging and Burden Measures ………… 72
22. Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using SOBI-P and GSIS LPSW Measures ……………….. 73
! xi
FIGURES
Figure
1. Assumptions of the Interpersonal-Psychological Theory of Suicide …... 5
2. Hypothesized Causal Pathway to Desire for Suicide …………………... 6
3. Scree Plot for Interpersonal Needs Questionnaire ……………………... 46
4. Causes of Burden to Others …………………………………………….. 57
5. Histogram of Transformed Current Suicide Ideation Variable ………… 59
6. Interaction between Model 2 Burden and Belonging for Current Suicide Ideation ……………………………………………………………... 62
7. Histogram of Transformed Past Suicide Attempt Variable …………….. 68
8. Interaction between Model 2 Burden and Belonging for Past Suicide Attempt ……………………………………………………………... 70
CHAPTER 1
INTRODUCTION
Old age is often characterized as a time of decreased physical and cognitive functioning as well as diminished contributions to one’s family and society. Later life is also associated with a number of accumulated losses, including relationships, roles, and status. With such a bleak outlook, it is easy to imagine why some older adults might consider suicide as a solution to life’s problems. In fact, suicide is a significant public health problem in the older adult population. According to the Centers for Disease
Control and Prevention (CDC), between 2000 and 2006 adults aged 65 years and older in the United Sates had a substantially higher death rate by suicide (14.8 per 100,000) than the general population (10.9 per 100,000; CDC, 2010). This difference was especially pronounced between the suicide rate among men aged 65 and older (30.1 per 100,000) as compared to adolescent males aged 15 to 19 (12.3 per 100,000). In addition, men aged 65 and older had a suicide rate over seven times greater than women aged 65 and older
(CDC, 2010). Studies have found that suicidal older adults typically do not utilize mental health services, which would presumably help to reduce suicide ideation and attempts.
Unfortunately, as compared to individuals aged 35 and younger, older adults tend to have had less contact with mental health services within the year before suicide (Luoma,
Martin, & Pearson, 2002). Given these troubling statistics and the increasing average age of individuals in the United States, the problem of suicide among older adults is likely to ! 2 grow in the future. A better understanding of this phenomenon will help to inform the assessment and treatment of older adults who might be at risk for suicide.
The distinction between thinking about suicide, attempting suicide, engaging in non-suicidal self-injury (NSSI), and completing suicide is important. Completed suicides involve those who have actually died by their own hand (Maris, Berman, &
Silverman, 2000), presumably with the intent to die. A suicide attempt refers to
“engagement in potentially self-injurious behavior in which there is at least some intent to die” (Nock & Favazza, 2009, p. 12). Although it is believed that nonfatal suicide attempts outnumber completed suicides, the exact ratio is speculative (Maris et al., 2000). In contrast, NSSI is “the direct, deliberate destruction of one’s own body tissue in the absence of suicidal intent” and is also referred to as self-mutilation, deliberate self-harm, self-cutting, and parasuicide in the literature (Nock & Favazza,
2009, p. 9). Finally, suicide ideators are “individuals who think about or form an intent to suicide of varying degrees of seriousness but do not make an explicit suicide attempt or complete suicide” (Maris et al., 2000, p. 20).
Twelve-month prevalence estimates of suicide ideation, plans, and attempts are 2.6%, 0.7%, and 0.4% respectively among adults aged 18 and older (Borges et al.,
2006). Suicide ideators with a plan are more likely to make an attempt (31.9%) than those without a plan (9.6%; Borges et al., 2006). The prevalence rates of suicide- related behaviors in the general population have remained relatively constant over the past two decades (Kessler et al., 2005). Although the exact figures are unknown, some estimate that approximately 31% of the clinical population and 24% of the general population have considered suicide at some point in their lives (Linehan & ! 3
Laffaw, 1982). Whether or not individuals who fall into the different subcategories of suicidality exist along a continuum or are categorically different from one another is an unanswered empirical question and, unfortunately, is often confounded in the literature (Maris et al., 2000). Research regarding the entire span of suicidality, from those who never contemplate suicide to those who complete suicide, can provide a more thorough understanding of the problem of suicide.
The majority of psychological research on suicide in older adults has focused on risk factors for suicide. In a recent literature review, Heisel (2006) reported that elevated risk for suicide in older adults is associated with communication of suicide ideation and intent, presence of a mood or other mental disorder, presence of a personality disorder or rigid personality style, presence and number of physical illnesses, psychosocial factors such as financial and legal difficulties or social isolation, and functional impairment in activities of daily living. Among the old-old
(i.e., age 75 years and older), family conflict, serious physical illness, loneliness, and both major and minor depression are significant risk factors for suicide (Waern,
Rubenowitz, & Wilhelmson, 2003). Common characteristics of late-life suicide include less expression of suicide ideation, use of more lethal means, higher prevalence of hopelessness and depression, and higher prevalence of physical illness as compared to suicide earlier in the life span (Fiske & Arbore, 2000 – 2001). In addition, older adults appear to have higher suicidal intent than younger adults, with previous suicide attempts and mood disorders being the main factors associated with intent among older adults (Miret et al., 2010). ! 4
Far fewer studies have examined associations between suicide and DSM-IV-
TR Axis II personality disorders (PDs) in older adults, even though empirical evidence suggests approximately 30-40% of all suicides are completed by individuals with a PD (Duberstein & Conwell, 1997). Among individuals of all ages, increased suicide risk appears to be associated with borderline PD, antisocial PD, avoidant PD, and possibly schizoid PD (Duberstein & Conwell, 1997). In a review of controlled studies examining risk factors for late life suicide, Conwell, Duberstein, and Caine
(2002) concluded that personality traits might play an important role; however, further research is needed in order to make more definitive conclusions.
In spite of the high rate of suicide in older adults, most do not contemplate, attempt, or complete suicide. What stops some older adults with significant risk factors from following through on their suicidal thoughts? This question remains largely unanswered.
Introduction to the Interpersonal-Psychological Theory of Suicide
In spite of decades of psychological theorizing and empirical research on the risk factors for suicide, researchers and clinicians today are usually not capable of predicting who, specifically, will attempt or complete suicide. The only major psychological theory of suicide that has been put forth in the past 20 years, the
Interpersonal-Psychological Theory of Suicide (IPTS), was developed by Thomas
Joiner and colleagues in an effort to determine more sensitive and specific predictors of completed suicide. This theory proposes that in order for individuals to complete suicide, they must have both the desire for suicide and the ability to carry out the act
(Joiner, 2005; Van Orden, Merrill, & Joiner, 2005; see Figure 1). Desire for suicide ! 5 contains two components: a thwarted sense of belongingness and a feeling of perceived burdensomeness on others.
Figure 1. Assumptions of the interpersonal-psychological theory of suicide. Adapted from “The Interpersonal Theory of Suicide,” by K. A. Van Orden, T. K. Witte, K. C. Cukrowicz, S. R. Braithwaite, E. A. Selby, & T. E. Joiner, 2010, Psychological Review, 117(2), p. 576.
Thwarted belongingness refers to a sense of profound alienation, including the feeling that one is not an integral part of any valued group, such as a family, a circle of friends, or society in general. The two main hypothesized components of thwarted belongingness are loneliness and an absence of reciprocal care, defined as relationships “… in which individuals both feel cared about and demonstrate care of another” (Van Orden et al., 2010, p. 582; see Figure 2). Perceived burdensomeness refers to a self-view that one is defective and flawed, to the point of being a liability to others. The two main hypothesized components of perceived burdensomeness are liability and self-hate (Van Orden et al., 2010; see Figure 2). The IPTS does not propose that thwarted belongingness and perceived burdensomeness are the only ! 6 factors that lead to suicidal desire, simply that their joint presence is highly likely to result in an exceptionally dangerous form of suicidal desire (Van Orden, Witte,
Gordon, Bender, & Joiner, 2008). Further, thwarted belongingness and perceived burdensomeness are presumed to be distinct, but related concepts (Van Orden et al.,
2010). The third variable, the capacity to carry out the act of suicide, refers to the acquired capability for self-harm, which includes habituation to pain and sense of fearlessness about death that is learned over time.
Figure 2. Hypothesized causal pathway to desire for suicide. For text of hypotheses, see Table 1. Adapted from “The Interpersonal Theory of Suicide,” by K. A. Van Orden, T. K. Witte, K. C. Cukrowicz, S. R. Braithwaite, E. A. Selby, & T. E. Joiner, 2010, Psychological Review, 117(2), p. 588.
The IPTS posits that the presence of all three components are necessary for completed suicide to occur and, while many individuals may desire suicide, evidence suggests that fewer have the capability for lethal self-harm (Van Orden et al., 2005).
This is consistent with the relatively low rates of attempted and completed suicide ! 7 versus the rate of suicide ideation in the general population (Borges et al., 2006;
Kessler et al., 2005). Thus, an individual may perceive low belongingness and high burdensomeness, but without the capability for self-harm would not be likely to complete suicide. Similarly, an individual may have acquired the capability for self- harm, but without perceived low belongingness and high burdensomeness would not be likely to complete suicide. The theory also proposes that active suicidal desire will develop only if individuals believe their levels of thwarted belongingness and perceived burdensomeness to be unchangeable; that is, they must be hopeless about their perceived interpersonal status (Van Orden et al., 2010; see Table 1).
Table 1
Hypotheses of the Interpersonal-Psychological Theory of Suicide
Number Hypothesis 1 Thwarted belongingness and perceived burdensomeness are proximal and sufficient causes of passive suicidal ideation.
2 The simultaneous presence of thwarted belongingness and perceived burdensomeness, when perceived as stable and unchanging (i.e., hopelessness regarding these states), is a proximal and sufficient cause of active suicidal desire.
3 The simultaneous presence of suicidal desire and lowered fear of death serves as the condition under which suicidal desire will transform into suicidal intent.
4 The outcome of serious suicidal behavior (i.e., lethal or near lethal suicide attempts) is most likely to occur in the context of thwarted belongingness, perceived burdensomeness (and hopelessness about both), reduced fear of suicide, and elevated physical pain tolerance. Note. Adapted from “The Interpersonal Theory of Suicide,” by K. A. Van Orden, T. K. Witte, K. C. Cukrowicz, S. R. Braithwaite, E. A. Selby, & T. E. Joiner, 2010, Psychological Review, 117(2), p. 581.
The absence of either of these variables, perceived as stable and permanent, is likely to be a protective factor against the development of suicidal desire (Van Orden et al., ! 8
2010). The IPTS attempts to distinguish between individuals who are likely to engage in lethal or near-lethal suicide attempts from those who desire suicide or make low- lethality attempts, with the ultimate goal of preventing death by suicide by determining more sensitive and specific predictors of it (Joiner, Van Orden, Witte, &
Rudd, 2009).
Empirical Support for the Interpersonal-Psychological Theory of Suicide
Although the development of this theory is relatively recent, mounting empirical evidence offers support for its primary tenets. Three studies provide support for the hypothesized relationship between perceived burdensomeness and suicide completion, suicide attempt, and current suicide ideation. The first compared the content of suicide notes of 20 individuals who attempted but did not complete suicide
(M age = 35.8 years, SD = 16.73, no age range reported) to the suicide notes of 20 individuals who completed suicide (M age = 37.4 years, SD = 14.33; Joiner et al.,
2002). Burdensomeness was positively correlated with completer status and, among those who completed suicide, with more lethal methods (Joiner et al., 2002). These findings held true even when other relevant variables, such as hopelessness and emotional pain, were statistically controlled. The findings suggest that the feeling of being a burden on loved ones is a significant risk factor for suicide completion.
In the second study regarding the hypothesized relationship between perceived burdensomeness and suicide, Van Orden and colleagues (2006) examined the relationships between perceived burdensomeness and key suicide-related variables, including past number of suicide attempts and current suicidal symptoms, in 343 adult psychotherapy outpatients (M age = 26.5 years, SD = 9.38, age range: 18 ! 9
– 62). Perceived burdensomeness was a significant predictor of both attempt status and current symptoms, even after controlling for age, gender, presence of personality disorder, depression, and hopelessness (Van Orden, Lynam, Hollar, & Joiner, 2006).
These researchers noted that perceived burdensomeness was positively correlated with age (r = .12), such that older participants reported higher levels of feeling like a burden.
A third direct test of the hypothesized relationship between perceived burdensomeness and suicide was conducted with two samples of older adults. Among older adults recruited from a university-affiliated research database (N = 57, M age =
74.14 years, SD = 7.51, no age range reported), perceived burdensomeness accounted for a significant amount of variance in current suicide ideation, after accounting for age, gender, depression, and loneliness (Cukrowicz, Cheavens, Van Orden, Ragain, &
Cook, 2011). In the second sample, which involved 105 older adults recruited from a university-affiliated primary care clinic (M age = 70.89 years, SD = 7.63, no age range reported), perceived burdensomeness accounted for a significant amount of variance in current suicide ideation, after accounting for age, gender, depression, hopelessness, loneliness, and physical functioning (Cukrowicz et al., 2011). Gender was not a moderator between perceived burdensomeness and suicide ideation in either of the models (Cukrowicz et al., 2011). These findings provide support for the importance of perceptions of burden to suicide ideation among older adults. While this study represented an important first step towards applying the IPTS to older adults, it did not examine the combination of perceived burdensomeness and thwarted ! 10 belongingness, which, according to the IPTS, is the key determinant of suicidal desire
(see Table 1).
A naturalistic study of the relationship between belongingness and suicide rates also provided preliminary support for the theory. Joiner, Hollar, and Van Orden
(2006) hypothesized that a “pulling together” effect may occur when individuals perceive themselves to be part of a valued group (e.g., fans of a sports team) and that interactions with other group members (e.g., other fans of the same team) may engender feelings of belongingness. According to the IPTS, increased feelings of belongingness ultimately should result in lower suicide rates. For example, the researchers found that suicide rates in two communities in which there was wide interest in the local college football team (Columbus, Ohio, home of the Ohio State
Buckeyes and Gainesville, Florida, home of the University of Florida Gators) correlated with the teams’ final national rankings, even after adjusting for economic factors. That is, when the teams performed poorly, there were higher suicide rates in the respective communities (Joiner et al., 2006). For comparison, they also examined the suicide rates in a community with relatively little interest in the local college football team (Miami-Dade County, Florida, home of the University of Miami
Hurricanes) and no such relationship between suicide rates and national rankings was found. Joiner et al. also found that significantly fewer suicides occurred in the U.S. on
February 22, 1980 (the day of the “Miracle on Ice”) than any other February 22 during the 1970s and 1980s, which they attributed to a “pulling together” effect of the
U.S. Olympic hockey team’s surprising defeat over the U.S.S.R. Finally, Joiner et al. discovered that fewer suicides occurred in the U.S. on Super Bowl Sundays, as ! 11 compared to the Sundays before and after the games, since the mid-1980s when the
Super Bowl became an important social event. Although not entirely conclusive, the findings do provide converging evidence that sports-related “pulling together” increases belongingness, which, in turn, reduces suicide rates.
An exploratory study examined the mediating role of belongingness with regard to changes in suicide ideation over time. Suicide ideation was highest among undergraduates in a large state university (N = 309, M age = 19 years, age range: 17 –
51) in the summer semester, when most students were not enrolled in classes, as compared to fall and spring semesters (Van Orden, Witte, James et al., 2008). In addition, level of belongingness was the lowest during the summer and significantly predicted suicide ideation. Belongingness mediated the relationship between semester and suicide ideation, such that variation in suicide ideation was accounted for by decreased belongingness (Van Orden, Witte, James et al., 2008).
More recently, researchers have begun investigating the interaction between the three variables in the IPTS. In a series of three studies, Van Orden and colleagues
(2008) tested the main hypotheses of the IPTS (see Table 1). The first study, conducted with a sample of undergraduates (N = 309, M age = 19 years, age range: 17
– 51), found that, after controlling for age, gender, and depression, perceived burdensomeness significantly predicted current suicide ideation, whereas thwarted belongingness did not. However, the joint presence of thwarted belongingness and perceived burdensomeness significantly predicted current suicide ideation, above and beyond age, gender, and depression (Van Orden, Witte, Gordon, Bender, & Joiner,
2008). The form of the interaction was such that individuals with both the highest ! 12 ratings of perceived burdensomeness and thwarted belongingness reported the highest level of suicide ideation. The second study found that past experience with painful and “provocative” events, including past suicide attempts, increased an individual’s capability for self-harm among adult outpatients of a community mental health clinic
(N = 228, M age = 26.21 years, age range: 18 – 54), even after controlling for age, gender, depressive symptoms, and suicide ideation (Van Orden, Witte, Gordon et al.,
2008). The third study found that, among another sample of adult outpatients from the same community mental health clinic (N = 153, ages were similar to the sample in study 2), perceived burdensomeness was significantly correlated with risk for suicide whereas acquired capability was not significantly related to risk for suicide. However, the interaction between perceived burdensomeness and acquired capability did significantly predict clinician-rated risk for suicide, after controlling for age, gender, and depression (Van Orden, Witte, Gordon et al., 2008). The form of the interaction was such that individuals with both the highest ratings of perceived burdensomeness and acquired capability were at higher risk for suicide. Together, these findings provide support for the hypothesized interaction between the variables described in the IPTS. Unfortunately, older adults were not included in any of these studies, limiting the application of the theory to the older adult population.
Another set of studies found support for the main predictions of the IPTS in two samples of young adults (Joiner, Van Orden, Witte, Selby, et al., 2009). The first study examined mattering (i.e., perceived burdensomeness) and family social support
(i.e., social alienation or low belonging) among a large, ethnically-diverse sample of young adults with depressive symptoms (N = 815, M age = 20.02 years, age range: 19 ! 13
– 26). The two variables significantly predicted suicide ideation after controlling for current and lifetime depression, both individually and as an interaction, such that those who were low in mattering and low in family social support reported the highest levels of suicide ideation (Joiner, Van Orden, Witte, Selby, et al., 2009). The second study examined the simultaneous presence of all three variables of the IPTS in predicting serious suicide attempt, versus suicide ideation, in a clinical inpatient sample of young adults (N = 313, M age = 22.17 years, SD = 2.76, no age range reported). After accounting for a number of relevant variables, including age, gender, depression, mania, hopelessness, borderline personality disorder, and family psychiatric background, the three-way interaction between perceived burdensomeness, thwarted belongingness, and acquired capability predicted whether or not participants’ recent suicidal crises involved suicide attempts versus only suicide ideation (Joiner, Van Orden, Witte, Selby, et al., 2009). The second study represents the most rigorous test of the predictions of the IPTS to date.
The Interpersonal-Psychological Theory of Suicide and High-Risk Populations
Researchers are also beginning to apply the IPTS to specific high-risk populations, such as individuals in the military. One qualitative study explored the
IPTS concepts among combat veterans (N = 16, no mean age reported, age range: 18
– 55, 93.75% male) returning from areas such as Iraq and Afghanistan (Brenner et al.,
2008). The veterans described variability in individuals’ ability to tolerate pain (i.e., acquired capability) and an emotional numbing that accompanied habituation to pain and developed in response to fear. Perceived burdensomeness was associated with a loss of sense of self post-discharge, often related to a decrease in status or purpose ! 14 upon return to civilian life. Finally, failed belongingness was linked to a sense of disconnectedness from civilian life and people not in the military (Brenner et al.,
2008). Another study investigated whether the IPTS could discriminate between a group of Air Force personnel who had died by suicide (n = 60, M age = 30 years, no
SD or age range reported) and a group of Air Force personnel who were living (n =
122, M age = 32 years; Nademin et al., 2008). All three variables taken together differentiated between the two samples, but when controlling for each variable independently, only acquired capability was a significant predictor of suicide risk.
Bryan and colleagues (2010) found partial support for the IPTS in an active duty Air Force sample (N = 88, no mean age reported, 93.2% of the sample was 18 –
24 years old). Specifically, they discovered a significant interaction between burdensomeness and acquired capability that predicted suicidal history. In contrast, the three-way interaction between belongingness, burdensomeness, and acquired capability did not significantly predict suicidal history (Bryan, Morrow, Anestis, &
Joiner, 2010). Given that belongingness did not emerge as a significant factor, the researchers suggested that it may be less robustly related to suicidal desire than burdensomeness. Alternatively, the unexpected findings may be due to the fact that they were measuring past suicidality, but not current suicidality (Bryan et al., 2010).
Case studies of active duty post-deployers have also been used to illustrate the application of the IPTS to a high-risk population (Anestis, Bryan, Cornette, & Joiner,
2009).
Another population at elevated risk for suicide is individuals with substance abuse disorders. The relationship between burdensomeness and lifetime presence of a ! 15 suicide attempt was examined among individuals with opiate dependence (N = 139,
M age = 41.8 years, age range: 21 – 59); however, the results were somewhat mixed
(Conner, Britton, Sworts, & Joiner, 2007). Lower burdensomeness was associated with lower probability of an attempt after accounting for demographic (e.g., age) and clinical covariates (e.g., depression), but this significant association was not observed after accounting for other interpersonal variables including belongingness and loneliness. In contrast, higher belongingness was associated with lower probability of a history of attempted suicide and this significant association remained after accounting for demographic, clinical, and other interpersonal variables. These associations were unique to suicide attempt, as burdensomeness and belongingness were both unrelated to unintentional drug overdoses. Interestingly, loneliness was not significantly associated with probability of suicide attempt, suggesting that loneliness is a slightly different construct from belongingness and burdensomeness (Conner et al., 2007).
More recently, another study investigated various dimensions of social connectedness, and their associations with past suicide ideation and past suicide attempt, among individuals with substance abuse disorders, including alcohol, cocaine, and other drugs (N = 814, M age = 39.0 years, SD = 11.3, no age range reported; You, Van Orden, & Conner, 2010). All indices of social connectedness, including interpersonal conflict, low perceived social support, low belongingness, and living alone, were associated with history of suicide attempt and, with the exception of living alone, history of suicide ideation (You et al., 2010). The findings provide ! 16 additional empirical support for the IPTS and its applicability to high-risk populations.
The Interpersonal-Psychological Theory of Suicide and Protective Factors
Another emerging area of research is relating the three IPTS variables to protective factors against suicide. For example, dispositional hope was negatively correlated with thwarted belongingness and perceived burdensomeness in a sample of undergraduates (N = 129, M age = 20.19 years, age range: 18 – 38; Davidson,
Wingate, Rasmussen, & Slish, 2009). However, dispositional hope unexpectedly had a positive relationship with acquired capability. Burdensomeness, belongingness, and acquired capability together significantly predicted suicide ideation, accounting for
36%, 6%, and 4% of the variance, respectively (Davidson et al., 2009).
Cultural beliefs that are viewed as protective factors against suicide among certain groups may also be relevant. A recent study of Mexican and Mexican
American women (N = 61, M age = 34.55, SD = 12.05, no age range reported) investigated the relationships between burdensomeness, familism, and suicide ideation (Garza & Pettit, 2010). Familism was defined as, “a strong sense of identification with and attachment to one’s family, as evidenced by strong feelings of loyalty, reciprocity, and solidarity among family members” (Garza & Pettit, 2010, p.
562). The researchers hypothesized that burdensomeness would significantly predict suicide ideation, particularly among women who endorsed high levels of familism, such that those with high perceived burdensomeness and a high level of familism would have the highest level of suicide ideation. Results showed that burdensomeness did significantly predict suicide ideation; however, neither familism by itself nor the ! 17 interaction between burdensomeness and familism significantly predicted suicide ideation (Garza & Pettit, 2010). The findings suggest perceived burdensomeness is a risk factor for suicide ideation in this population, regardless of one’s feelings towards family.
In summary, existing empirical findings support the hypothesized relationships between burdensomeness, belongingness, acquired capability, and various suicidal behaviors among younger adults and high-risk groups, such as individuals in the armed forces and individuals with substance abuse disorders. There is also mounting support for the hypothesized interactions between the IPTS variables. That is, each variable alone does not necessarily indicate increased risk for suicide, but the interaction between the variables does. The IPTS seems particularly well suited to explain late life suicide because of the increased likelihood of experiencing a shrinking social network and dependence on others due to functional decline in late life, but very little research has directly tested the theory among older adults. The findings of studies examining similar constructs as those posited by the
IPTS suggest the theory may aid in gaining a better understanding late life suicide.
Belongingness and Suicide in Older Adults
Over a century of theoretical development and decades of empirical research have established social isolation as a significant contributing factor to suicidal behavior (Trout, 1980). Social isolation may be defined as “a state in which interpersonal contacts and relationships are disrupted or nonexistent” (Trout, 1980, p.
10). As mentioned previously, poor social integration, family conflict, and loneliness are robust risk factors for suicide among older adults (Heisel, 2006; Heisel & ! 18
Duberstein, 2005; Waern, Rubenowitz, & Wilhelmson, 2003). However, in a review of case control studies using the psychological autopsy method, researchers concluded that, while life events (e.g., interpersonal losses) and lack of social support appear to represent risk factors for late life suicide, it is unclear to what extent their effects may be mediated or moderated by their relationships with other variables, such as depression, personality, and culture (Conwell, Duberstein, & Caine, 2002).
More recent evidence suggests that lack of social interaction and religious involvement do increase the risk for completed suicide among older adults, over and above the effects of a mood disorder and occupational status (Duberstein et al., 2004).
Nevertheless, the associations between social support variables and suicidality among older adults may not be as straightforward as they appear, especially when they are examined in the context of other important variables.
It is also evident that it is not just the quantity of social interaction that is relevant to late life suicide, but also the quality. The definition of social isolation, as described by Trout (1980), encompasses both the physical and emotional elements of social isolation. That is, both objective and subjective measures of social isolation appear to be useful in understanding the relationship between social isolation and suicide (Trout, 1980). For example, one survey study of 104 older adults living in their own homes or in nursing homes (M age = 77.9 years, age range: 65 – 100) investigated the relationships between suicide ideation and several personal and social support variables, including marital status, living arrangements, social isolation, and support satisfaction (Mireault & de Man, 1996). Among the social variables, only social isolation (r = .22) and support satisfaction (r = .42) were significantly ! 19 correlated with suicide ideation. In a multiple regression analysis, social support satisfaction accounted for the largest proportion of variance in suicide ideation
(Mireault & de Man, 1996). Together, these findings suggest that perceptions of social support may be more important than the quantity of support or social contact, at least with regard to suicide ideation among older adults.
Hagerty and colleagues (1992) posited that sense of belonging is a concept that is similar to but distinct from others such as loneliness, alienation, and social support, as well as distinct from quantitative reports of social interaction. They defined sense of belonging as “the experience of personal involvement in a system or environment so that persons feel themselves to be an integral part of that system or environment” (Hagerty et al., 1992, p. 173). They proposed two defining psychological attributes of sense of belonging: (1) valued involvement, in which individuals feel valued, needed, and accepted; and (2) fit, in which individuals perceive themselves to complement their environment (Hagerty et al., 1992). They also proposed that individuals must possess the energy and potential desire for social involvement, which they termed antecedents to sense of belonging. They later named these separate components belonging-psychological and belonging-antecedent
(Hagerty & Patusky, 1995). Their view of sense of belonging is very much in line with the concept of belongingness proposed by the IPTS. According to the IPTS, thwarted belongingness refers to the feeling that one is not an integral part of any valued group, such as a family, a circle of friends, or society in general. The two main hypothesized components of thwarted belongingness are loneliness and an absence of reciprocal care (Van Orden et al., 2010). ! 20
McLaren and colleagues in Australia have conducted most of the existing research on belongingness and suicide in older adults to date, using a measure of sense of belonging developed by Hagerty and Patusky (1995). One of their earliest studies examined the associations between physical activity, performed alone or with others, sense of belonging, depression, and suicide ideation (Bailey & McLaren,
2005). They found that suicide ideation was moderately correlated with sense of belonging-antecedent (r = -.35) and strongly correlated with sense of belonging- psychological (r = -.59) in the community-dwelling sample of 194 older adults (for men M age = 69.59 years, SD = 6.64 and for women M age = 68.21 years, SD = 7.78, no age ranges reported; Bailey & McLaren, 2005). Interestingly, simply performing activities with others was not significantly associated with sense of belonging or mental health, again suggesting that it is the quality, not necessarily the quantity, of interaction that is important (Bailey & McLaren, 2005).
Another study investigated various human relatedness variables, including marital status, social support resources, and belongingness, as predictors of depression and suicide ideation in a community sample of 110 older adults (M age =
76.67 years, SD = 8.11, no age range reported; Van Der Horst & McLaren, 2005).
Although suicide ideation was moderately associated with social support (r = -.44), sense of belonging-antecedent (r = -.35), and sense of belonging-psychological (r =
-.30), only social support was a significant negative predictor of suicide ideation in a subsequent hierarchical multiple regression analysis (Van Der Horst & McLaren,
2005). That is, sense of belonging did not predict suicide ideation beyond that which ! 21 was attributable to marital status and social support (Van Der Horst & McLaren,
2005).
Sense of belonging is also associated with protective factors against suicide among older adults. Kissane and McLaren (2006) found that a higher sense of belonging predicted more reasons to live overall and, specifically, was related to child-related concerns, responsibility to family, and survival and coping beliefs in a sample of 104 older adults (M age = 77.5 years, age range: 61 – 95). McLaren and colleagues (2007) investigated the applicability of various resiliency models to describe the relationship between sense of belonging and suicide ideation among older adults (N = 352, M age = 71.31 years, SD = 7.99, no age range reported).
Suicide ideation was moderately correlated with sense of belonging-antecedent (r = -
.30 for women and r = -.38 for men) and strongly correlated with sense of belonging- psychological (r = -.49 for women and r = -.51 for men; McLaren, Gomez, Bailey, &
Van Der Horst, 2007). The findings provided support for the notion that both components of belongingness were protective factors against suicide ideation. For example, a high level of belongingness reduced the effects of depression on suicide ideation for both older men and older women (McLaren et al., 2007). These findings were replicated with a sample of 99 male farmers (M age = 48.14 years, age range: 19
– 77), such that when participants reported high levels of both social support and sense of belonging, there was no longer a relationship between suicide ideation and depression (McLaren & Challis, 2009). Notably, the correlation between sense of belongingness-psychological and suicide ideation was stronger in this sample than in previous ones (r = -.66; McLaren & Challis, 2009). ! 22
More recently, another group of researchers investigated the relationships between suicidal thoughts and attempts, perceived social support, and chronic interpersonal difficulties among older adults (Harrison et al., 2010). They considered chronic interpersonal difficulties to be an indicator of personality factors that were likely to have an impact on relationships. By comparing suicidal depressed (n= 29, M age = 69.8 years, SD = 7.8), non-suicidal depressed (n = 58, M age = 72.1 years, SD =
8.4), and non-depressed (n = 19, M age = 69.2 years, SD = 8.6) individuals, they discovered that those with suicide attempts and severe suicide ideation reported lower perceived social support and more chronic interpersonal difficulties (Harrison et al.,
2010). In addition, low perceived social support (which included a measure of belongingness) partially explained the relationship between interpersonal difficulties and suicidal behavior (Harrison et al., 2010). The findings point to the importance of subjective feelings of social support and suggest that personality factors may play a role in late-life suicide, although this was not assessed directly.
In summary, for the most part, only one research group has investigated belongingness and late-life suicide. The findings from their studies provide evidence for feelings of belongingness as a protective factor against suicide in older adults, but the relationship between the absence of belongingness and suicidality is somewhat less clear. Belongingness is strongly associated with other social support variables, and it appears to be a distinct concept from other, more widely researched, concepts such as social isolation, loneliness, and alienation. As of yet, belongingness has not been examined in combination with burdensomeness in older adults, which is one of the main hypotheses of the IPTS (see Table 1). ! 23
Burdensomeness and Suicide in Older Adults
According to the IPTS, perceived burdensomeness refers to a self-view that one is defective and flawed, to the point of being a liability to others. The two main hypothesized components of perceived burdensomeness are liability and self-hate
(Van Orden et al., 2010). To date, there are few published studies examining perceived burdensomeness and its relation to suicidality in older adults (e.g.,
Cukrowicz et al., 2011) and almost no theoretical development in this area. Although not the main focus of investigation, perceived burdensomeness emerged as a theme among the suicide notes of older adults who completed suicide in one psychological autopsy study (Foster, 2003). Older adults’ (age 65 and older, no mean age or age ranges reported) notes were significantly more likely to contain the topic of burden to others than the notes of younger and middle aged adults (40% vs. 3%, p = .03; Foster,
2003). Although this study had a small sample size (N = 42), restricting the generalizability of the results, it provides support for the importance of perceived burdensomeness in late life suicide.
In a systematic literature review on self-perceived burden among people who are nearing the end of their lives, McPherson and colleagues (2007b) found that, although knowledge about the prevalence of self-perceived burden is limited, it is a common concern among those with advanced disease. A number of studies they reviewed pointed to self-perceived burden as an important value underlying decisions regarding life-sustaining treatment and physician-assisted suicide. For instance, one qualitative study regarding the attitudes of terminally ill cancer patients toward euthanasia and physician-assisted suicide found 46% of the participants (N = 70, ! 24
M age = 64.5 years, SD = 12.1, no age range reported) could imagine asking for euthanasia or assisted suicide in the future (Wilson et al., 2000). Among the reasons participants gave for wanting to hasten death in the future were if they become a burden to others and/or develop a feeling of hopelessness about their overall situations (Wilson et al., 2000). In another qualitative study involving terminally ill cancer patients, those with a high wish to hasten death (n = 16, M age = 61 years, SD
= 16.85, no age range reported) perceived themselves to be more of a burden on others than those with moderate or no wish to hasten death (n = 26, M age = 71, SD =
10.99 and n = 30, M age = 63, SD = 13.22 respectively; Kelly et al., 2002). Consistent with the hypotheses of the IPTS, the participants with high wish to hasten death also more frequently reported hopelessness and lack of social support than the moderate or no wish to hasten death groups (Kelly et al., 2002).
Perception of being a burden on self or others in the present or the future was conceptualized as a chronic antecedent or trigger that preceded expression of desire for hastened death in participants with advanced cancer (N = 7, no mean age reported, age range: 40 – 79) in one qualitative study (Coyle & Sculco, 2004). Some of these participants believed their hastened death would be a gesture of altruism in that it would spare the family of the burden of care and of witnessing the participants’ decline (Coyle & Sculco, 2004). In another qualitative study of patients with advanced cancer (N = 69, M age = 64.5 years, SD = 12.2, no age range reported),
37.6% of participants reported moderate to extreme levels of distress around being a burden to others (Wilson, Curran, & McPherson, 2005). Self-perceived burden had small to moderate correlations with desire for death (r = .35) and suicide ideation ! 25
(r = .27). However, further analysis revealed no statistically significant differences in desire for death or suicide ideation between the low burden and high burden groups
(Wilson, Curran, & McPherson, 2005).
Although the relationship between the wish to hasten one’s death and perceptions of burdensomeness on others may provide insight regarding the relationship between self-perceived burden and suicide ideation, the wish to hasten death may not necessarily be equivalent to suicide ideation. One group of researchers investigated the factors associated with current suicidality among terminally ill cancer patients (N = 140, M age = 61 years, age range: 38 – 89) using semi-structured interviews (Akechi et al., 2004). The four major reasons for suicide ideation, as described by participants with current suicide ideation, included physical distress, anxiety/fear regarding the future course of illness, hopelessness, and burden on family. Perceived burdensomeness, as well as anxiety and depression, were the only statistically significant differences between the group of patients with suicide ideation and those without. However, logistic regression analysis revealed anxiety and depression, but not perceived burdensomeness, to be significant positive predictors of suicide ideation (Akechi et al., 2004).
Another study examined the characteristics of terminal cancer patients who completed suicide without the assistance of a physician during a home palliative care program (Filiberti et al., 2001). The researchers conducted a psychological autopsy by reviewing the medical records and patient reports of symptoms during their care and by conducting interviews of caregivers and medical professionals involved in the care of the five individuals who completed suicide (M age = 65 years; age range: 50 – 76). ! 26
Consistent with the hypotheses of the IPTS, although multiple vulnerability factors were present, a key characteristic shared among the deceased patients was the fear of being a burden on others, accompanied by a pervasive sense of hopelessness about their clinical conditions (Filiberti et al., 2001).
Whereas the issue of family caregiver burden is well documented (e.g.,
Pinquart & Sorensen, 2003), less attention has been directed at the corresponding issue of care recipients’ self-perceived burden. There is evidence that receiving help from others can lead to feelings of burdensomeness. For example, in one qualitative study of 50 older adult primary care patients (M age = 77.92, age range: 69 – 90) and their family caregivers, three main themes emerged in relation to burden (Cahill,
Lewis, Barg, & Bogner, 2009). First, the older adult care recipients discussed burden in relation to not wanting to complicate the busy lives of their adult children. The largest number of care recipients discussed this theme, which included dimensions of physical and social burden. Second, care recipients expressed guilt about their health problems and the limitations caused by their illness. Third, care recipients mentioned concern that sharing information about their health would cause unnecessary worry for their family caregivers (Cahill et al., 2009). Similarly, another qualitative study of
15 individuals receiving palliative care (no mean age reported, age range: 42 – 78) found three main categories of self-perceived burden: concern for others, implications for self, and minimizing burden (McPherson, Wilson, & Murray, 2007a).
Concern for others included the physical, social, and emotional hardships the participants believed they were creating for others, in addition to the likely effect of their death on those around them. Implications for self included feelings of ! 27 responsibility for causing difficulties for others, which resulted in distress and a diminished sense of self. Minimizing burden included coping strategies that participants used to lighten the burden on others and to reduce the negative impact on themselves. Two of the participants mentioned a desire for hastened death, connected to their wish to alleviate burden on their families (McPherson et al., 2007a). Although these studies do not include information on suicide specifically, their findings highlight important dimensions of self-perceived burden.
In summary, the extant literature regarding self-perceived burden and suicide in older adults has largely focused on those with advanced medical illness and physician-assisted suicide. There is limited information regarding perceptions of burdensomeness in older adults who are not dealing with a terminal illness. One restriction of this body of literature, at least with regard to the current study, is that the wish to hasten death is not necessarily equivalent to suicide ideation, plan, or intent. Further, most of this research is qualitative and/or conducted retrospectively.
Nevertheless, the findings suggest self-perceived burden is a significant factor in older adults with advanced disease who wish to hasten death. There is also cross- cultural support for the findings, including data from the United States (Coyle &
Sculco, 2004), Canada (Wilson, Curran, & McPherson, 2005; Wilson et al., 2000),
Australia (Kelly et al., 2002), and Japan (Akechi et al., 2004). In their review,
McPherson and colleagues (2007b) noted that researchers have used various ways of assessing self-perceived burden, making comparisons across studies difficult. In addition, they suggested that work regarding self-perceived burden could benefit from ! 28 a hypothesis-driven approach in order to fully understand the relationships between self-perceived burden and other variables.
The Present Study
The purpose of the present study was two-fold: (1) to examine the structure and validity of a previously-developed measure of burdensomeness and belongingness (i.e., Interpersonal Needs Questionnaire; Joiner, Van Orden, Witte, &
Rudd, 2009) in the older adult population; and (2) to evaluate the applicability of the suicidal desire component of the IPTS to community-dwelling older adults. The third
IPTS variable, acquired capability for suicide, was not investigated in the current study, as the rate was expected to be too low to detect in a community-dwelling older adult population. Thus, the focus of the current study was on the first two hypotheses of the IPTS (see Table 1), concerning perceived burdensomeness and thwarted belongingness.
With regard to the first aim, it was hypothesized that the Interpersonal Needs
Questionnaire (INQ) would be a valid and reliable instrument for use with older adults, as evidenced by: an exploratory principle components analysis and confirmatory factor analysis, used to determine the number of factors in the latent structure of the INQ in the general population of older adults (hypothesis 1a); a significant positive correlation between the INQ Burdensomeness subscale and a measure of low self-worth and a significant negative correlation between the INQ
Burdensomeness subscale and a measure of self-esteem (hypothesis 1b); and a significant negative correlation between the INQ Belongingness subscale and a ! 29 measure of high belonging and a significant positive correlation between the INQ
Belongingness subscale and a measure of loneliness (hypothesis 1c; see Table 2).
Table 2
Measurement of Study Variables
Variable Measure Current suicide ideation GSIS Suicide Ideation subscale
Past suicide attempt GSIS “I have tried ending my life in the past” item
Perceived Burdensomeness INQ Burdensomeness subscale
Validity measures for GSIS Loss of Personal and Social Worth subscale burdensomeness Rosenberg Self-Esteem Scale (1-item version)
Thwarted Belongingness INQ Belongingness subscale
Validity measures for Sense of Belonging Instrument–Psychological belongingness Three-Item Loneliness Scale
Depression Patient Health Questionnaire
Hopelessness Beck Hopelessness Scale
Dysfunctional personality traits Short Coolidge Axis II Inventory
Physical health RAND 36-Item Health Survey Physical Health subscale Note. GSIS = Geriatric Suicide Ideation Scale; INQ = Interpersonal Needs Questionnaire.
In addition, the correlations between INQ Burdensomeness and the self-worth measure and INQ Burdensomeness and the self-esteem measure were expected to be stronger than the correlations between INQ Belongingness and the self-worth measure and INQ Belongingness and the self-esteem measure (hypothesis 1d). The correlations between INQ Belongingness and the belongingness measure and INQ
Belongingness and the loneliness measure were expected to be stronger than the ! 30 correlations between INQ Burdensomeness and the belonging measure and INQ
Burdensomeness and the loneliness measure (hypothesis 1e).
With regard to the second aim, hypotheses included the following: perceived burdensomeness would be positively correlated with current suicide ideation
(hypothesis 2a); thwarted belongingness would be positively correlated with current suicide ideation (hypothesis 2b); the interaction between burdensomeness and belongingness would significantly predict current suicide ideation in older adults after accounting for other robust risk factors (hypothesis 2c); perceived burdensomeness would be positively correlated with previous suicide attempt (hypothesis 2d); thwarted belongingness would be positively correlated with previous suicide attempt
(hypothesis 2e); and the interaction between burdensomeness and belongingness would significantly predict previous suicide attempt in older adults after accounting for other robust risk factors (hypothesis 2f). These hypotheses were consistent with the first two hypotheses of the IPTS (see Table 1).
! CHAPTER 2
METHOD
Participants
Participants in the final sample were 284 community-dwelling older adults (M age
= 73.3 years, SD = 7.1 years, age range = 64 – 96 years; 55.6% female; see Table 3),
Table 3
Participant Characteristics (N = 284)
M SD Range Age 73.3 7.1 64 – 96 Education 15.0 3.0 8 – 25 Religion 7.21 2.6 1 – 10
n % Gender Female 158 55.6 Male 126 44.4
Racial/Ethnic European American 245 86.3 Native American/Alaskan Native 10 3.5 African American 9 3.2 Latino(a) 8 2.8 Multiracial 8 2.8 Asian American 2 0.7
Relationship Married 193 68.0 Widowed 58 20.4 Divorced 24 8.5 Single/Never Married 6 2.1 Domestic partnership 3 1.1
! 32 recruited through a mailing to 789 older adults whose names were randomly drawn from an El Paso county voter registration list, representing a 36.0% return rate. In the original sample, there were 433 women (54.9%) and 356 men (45.1%), with a mean age of 73.9 years (age range = 64 – 99 years). A chi-square goodness-of-fit test indicated there was no significant difference in the proportion of men and women between the original sample and the respondent sample, !2 (1, n = 284) = .06, p = .81. No other demographic data were available for the original sample. Three additional packets were returned, but excluded from all analyses because the participant chose to discontinue filling out the questionnaires before completion and one further packet was excluded because the participant’s spouse completed the packet for the participant.
In the final respondent sample, mean years of participants’ education was 15.0 years (SD = 3.0 years, range = 8 – 25 years). Ethnicities were reported as follows:
European American (86.3%), Native American/Alaskan Native (3.5%), African
American (3.2%), Latino(a) (2.8%), Multiracial (2.8%), and Asian American (0.7%).
Relationship statuses were reported as follows: married (68.0%), widowed (20.4%), divorced (8.5%), single/never married (2.1%), and domestic partnership (1.1%). Mean rating for degree of religiousness or spirituality was 7.21 (SD = 2.6, range = 1 – 10). A chi-square goodness-of fit test indicated there was no significant difference in the proportion of men (44.4%) and women (55.6%) in the respondent sample, !2 (1, n = 284)
= 3.61, p = .06.
Measures
Interpersonal Needs Questionnaire (INQ). The INQ is an 18-item self-report measure of beliefs about the extent to which individuals feel that they are connected to ! 33 others (i.e., Belongingness) and the extent to which they feel like a burden on the people in their lives (i.e., Burdensomeness; Joiner, Van Orden, Witte, & Rudd, 2009). Ten items measure Belongingness (e.g., “These days, other people care about me.”). Eight items measure Burdensomeness (e.g., “These days, I feel like a burden on the people in my life.”). Respondents are instructed to indicate the extent to which each item is true for them recently. Answers are based on a 7-point scale ranging from 1 (not at all true for me) to 7 (very true for me). Eight of the items are reverse coded so that higher scores reflect higher levels of thwarted belongingness and perceived burdensomeness.
In a sample of undergraduate students who took a shorter version of the current
INQ, internal consistency was high, with a Cronbach alpha of .85 for the Belongingness subscale and .89 for the Burdensomeness subscale (Van Orden, Witte, Gordon et al.,
2008). A confirmatory factor analysis conducted with a sample of older adults recruited from primary care clinics and an older adult research registry (N = 156, M age = 71.89 years, age range: 57 – 90) suggested the 18-item INQ was an adequate measure of belongingness and burdensomeness for that population (Van Orden, 2009). In the current sample, Cronbach alpha coefficients were .87 for the original Belongingness subscale and
.67 for the original Burdensomeness subscale. Refinement of the INQ is underway and a manuscript describing a 15-item version, containing six burdensomeness items and nine belongingness items, is currently under review (K. A. Van Orden, personal communication, March 3, 2011).
In the current study, exploratory principle components analysis (PCA) was used to determine on which subscale each item loaded. The PCA resulted in 8 items loading on the Belongingness subscale and 10 items loading on the Burdensomeness subscale, ! 34 somewhat inconsistent with previous analyses with other samples. Basing the subscales on the results of the current PCA, internal consistency was high for the Belongingness subscale (! = .91) and acceptable for the Burdensomeness subscale (! = .75). Internal consistency for the entire scale was high for the current sample (! = .87).
Geriatric Suicide Ideation Scale (GSIS). The GSIS is a 31-item self-report measure of suicide ideation and related factors in older adults consisting of four subscales: Suicide Ideation (10 items), Death Ideation (5 items), Loss of Personal and
Social Worth (7 items), and Perceived Meaning in Life (8 items; Heisel & Flett, 2006).
One auxiliary item is not included in any of the subscales: “I have tried ending my life in the past.” Respondents are instructed to decide to what extent they agree or disagree with each item. Answers are based on a 5-point scale ranging from 1 (strongly disagree) to 5
(strongly agree). Items on the Perceived Meaning in Life subscale are reverse-scored so that higher GSIS total scores indicate higher overall suicide ideation.
The initial reliability and validity information for the GSIS was gathered among heterogeneous samples of institutionalized and community-dwelling adults aged 65 years old and older. Cronbach alpha coefficients for the subscales and total scale ranged from
.82 to .93 and test-retest reliability coefficients ranged from .61 to .86 (Heisel & Flett,
2006). Construct validity of the GSIS was supported by a positive association with the
Scale for Suicide Ideation (SSI; Heisel & Flett, 2006). Convergent and discriminant validity were indicated by significant positive associations with measures of depression, hopelessness, and poor perceived physical health, and by negative associations with psychological well-being and life satisfaction (Heisel & Flett, 2006). In addition, GSIS scores differentiated between psychiatric and non-psychiatric groups and there is support ! 35 for using the Suicide Ideation subscale as a brief measure of late-life suicide ideation
(Heisel & Flett, 2006). A recent study of dispositional coping strategies, suicide ideation, and protective factors against suicide in older adults also provides support for the validity of the GSIS (Marty, Segal, & Coolidge, 2010). In the current study, Cronbach alpha coefficients were .92 for the total GSIS, .88 for the Suicide Ideation scale, .82 for the
Loss of Personal and Social Worth scale, .68 for the Death Ideation scale, and .83 for the
Meaning in Life scale.
Patient Health Questionnaire (PHQ-9). The PHQ-9 is a 9-item self-report measure of depressive symptoms, based on DSM-IV diagnostic criteria for Major
Depressive Disorder (Kroenke, Spitzer, & Williams, 2001). Respondents rate how often they experienced each symptom over the previous two weeks, on a 4-point scale ranging from 0 (not at all) to 3 (nearly every day). Scores for each item are added together so that higher overall scores indicate greater severity of depression, with total scores ranging from 0 to 27. Scores of 5, 10, 15, and 20 correspond to mild, moderate, moderately severe, and severe depression, respectively (Kroenke et al., 2001). The PHQ-9 has demonstrated good reliability and validity among the general population (M age = 48.8 years, SD = 18.1 years, range: 14 – 93 years; Martin, Rief, Klaiberg, & Braehler, 2006).
In the current study, the Cronbach alpha coefficient was .83.
Sense of Belonging Instrument (SOBI). The SOBI is a 27-item self-report measure of sense of belonging in adults, consisting of two separately scored subscales:
Psychological State (SOBI-P) and Antecedents (SOBI-A; Hagerty & Patusky, 1995).
Only the Psychological State subscale (SOBI-P) was used in the current study. The
SOBI-P contains 18 items (e.g., “If I died tomorrow, very few people would come to my ! 36 funeral.”). Respondents rate how much they agree with each statement, on a 4-point scale ranging from 1 (strongly disagree) to 4 (strongly agree). A high score on the SOBI-P indicates that the individual feels valued, needed, and accepted. Among older adults (N =
37, M age = 73.8 years, age range: 43 – 84), Cronbach alpha coefficients were .91 for the
SOBI-P and .76 for the SOBI-A (Hagerty & Patusky, 1995). Evidence of validity was demonstrated by significant positive associations with measures of social support and reciprocity and a significant negative association with a measure of loneliness (Hagerty &
Patusky, 1995). The SOBI has been consistently used in studies examining belongingness in older adults (e.g., Kissane & McLaren, 2006; McLaren, Gomez, Bailey, & Van Der
Horst, 2007). In the current study, the Cronbach alpha coefficient for the SOBI-P was high (! =.94).
Beck Hopelessness Scale (BHS). The BHS is a 20-item self-report measure that assesses pessimism and hopelessness (Beck, Weissman, Lester, & Trexler, 1974). The response choices were modified from a true/false format to a 5-point Likert scale, ranging from 1 (rarely or none of the time) to 4 (most or all of the time). This type of response format has been used successfully with research samples of older adults (Heisel & Flett,
2005; Neufeld, O’Rourke, & Donnelly, 2010) and was used to increase the variability of the scores and avoid a possible floor effect. Eleven items are negatively-worded (e.g., “I might as well give up because I can’t make things better for myself.”) and nine items are positively-worded (e.g., “I look forward to the future with hope and enthusiasm.”). The positively-worded items are reverse-scored so that higher overall scores indicate greater hopelessness or frequency of negative expectancies for the self or for the future, with total scores ranging from 20 to 80. Although the initial development and validation of the ! 37 scale was conducted with younger adult psychiatric patients, subsequent studies have established reliability and validity among depressed older adult outpatients (e.g., Hill,
Gallagher, Thompson, & Ishida, 1988) and older adults in the general population (e.g.
Greene, 1981). In the group of depressed older adult outpatients, the BHS had a coefficient alpha of .84 and had a moderate positive correlation with a measure of depression (r = .41; Hill et al., 1988). In the current study, internal consistency was high
(! = .88).
Single-Item Self-Esteem Scale (SISE). The SISE is a modified version of the 10- item self-report measure of global self-esteem developed by Rosenberg in 1965 (Robins,
Hendin, & Trzesniewski, 2001). Respondents rate how much they agree with the statement, “I see myself as someone who has high self-esteem,” on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The SISE has good reliability and validity among diverse samples of children and adults and has been shown to be an acceptable alternative to the Rosenberg scale with the context of research (Robins et al.,
2001). The SISE has also been successfully used in a large-scale cross-sectional study of self-esteem over the lifespan (M age = 24 years, SD = 9.7, age range = 9 – 90; Robins,
Trzesniewski, Tracy, Gosling, & Potter, 2002). No internal consistency was calculated for the current sample.
Three-Item Loneliness Scale (3LS). The 3LS is a 3-item self-report measure of subjective loneliness, originally designed for use on a telephone survey (Hughes, Waite,
Hawkley, & Cacioppo, 2004). Respondents rate how often they feel a certain way, ranging from 1 (hardly ever) to 3 (often). For example, one of the items is, “How often do you feel that you lack companionship?” Answers from the three items are added together ! 38 so that overall scores indicate more loneliness, with total scores ranging from 3 to 9. In two large-scale studies of older adults (M ages 66.5 years and 57.5 years, respectively), using both in-person self-administration and telephone administration, the 3LS had good internal consistency, with a Cronbach alpha coefficient reported of .72 in both studies
(Hughes et al., 2004). The 3LS also demonstrated convergent and construct validity
(Hughes et al., 2004). In the current study, internal consistency was .82.
RAND 36-Item Health Survey 1.0 (SF-36). The SF-36 is a 36-item self-report measure of subjective health status, consisting of eight subscales: physical functioning
(10 items), social functioning (2 items), role limitations due to physical problems (4 items), role limitations due to emotional problems (3 items), general mental health (5 items), energy/vitality (4 items), bodily pain (2 items), and general health perception (5 items; Ware & Sherbourne, 1992). Only the physical functioning scale was used in the current study. The original version of the SF-36 was adapted from longer instruments completed by patients in the Medical Outcomes Study (MOS); the RAND version is identical to the MOS SF-36 in content, but the scoring procedures differ slightly. Each item is scored so that a high score defines a more favorable health state, with transformed item scores ranging from 0 to 100. Within each subscale, item scores are averaged together to create a total subscale score. Reliability and validity of the SF-36 have been established among older adults (Bowling, 1997; McHorney, 1996). In the current study, internal consistency for the physical functioning subscale was high (! = .94).
Short Coolidge Axis II Inventory (SCATI). The SCATI is a 70-item self-report inventory designed to assess the 10 standard personality disorders listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR, American Psychiatric ! 39
Association, 2000): Antisocial, Avoidant, Borderline, Dependent, Histrionic, Narcissistic,
Obsessive-Compulsive, Paranoid, Schizotypal, and Schizoid (Coolidge, Segal, Cahill, &
Simenson, 2010). The SCATI can be used with both clinical and non-clinical populations, as it views personality disorders as dimensional, rather than categorical, in nature. Items are answered on a 4-point scale ranging from 1 (strongly false) to 4
(strongly true), with higher scores indicating the presence of traits associated with a particular personality disorder. In a community sample of adults age 16 to 88 years (M age = 33.9 years, SD = 14.9), internal scale reliabilities ranged from .58 to .74 and test- retest reliabilities over a 1-week interval for a subset of participants ranged from .72 to
.89 (Coolidge et al., 2010). The SCATI also showed evidence of construct validity
(Coolidge et al., 2010). Cronbach alpha coefficients for the current study are provided in
Table 4 showing a range from .40 to .79, with a median alpha of .62 for the 10 scales.
Procedure
Participants received a brief pre-notice letter delivered by the post office introducing the study. Approximately two weeks later, they received an envelope containing the following items:
• A cover letter addressed directly to the potential participant with a general
description of the study;
• An IRB-approved informed consent form, including contact information for the
primary investigator should the individual have a question about the study;
• A questionnaire packet, containing the measures described above; ! 40
• A debrief form, including the rationale for study and crisis resources should
participation temporarily cause distress or raise issues requiring clinical attention;
and,
• A self-addressed stamped envelope for returning the questionnaire packet.
Table 4
Psychometric Properties of the Major Study Variables
Range Variable M SD ! Potential Actual Hopelessness (20) 30.74 8.79 .88 20–80 20–70 Loneliness (3) 3.91 1.41 .82 3–9 3–9 Physical Health (10) 73.32 23.96 .92 0–100 0–100 Depression (9) 2.73 3.57 .83 0–27 0–22 SOBI-P Belonging (18) 65.09 7.77 .94 18–72 28–72 Self-Esteem (1) 4.28 .87 N/A 1–5 1–5
GSIS Total GSIS (31) 43.18 13.25 .92 31–155 31–128 Suicide Ideation (10) 11.83 4.26 .88 10–50 10–46 Loss of Worth (7) 10.91 4.51 .82 7–35 7–29 Death Ideation (5) 7.66 3.27 .68 5–25 5–25 Meaning in Life (8) 11.90 4.31 .83 8–40 8–32 Past suicide attempt (1) 1.12 .58 N/A 1–5 1–5
SCATI Antisocial (5) 6.27 1.46 .55 5–20 5–13 Avoidant (5) 7.34 2.59 .79 5–20 5–17 Borderline (5) 5.75 1.31 .55 5–20 5–13 Dependent (5) 6.77 1.77 .63 5–20 5–16 Histrionic (5) 8.00 2.32 .65 5–20 5–18 Narcissistic (5) 8.44 2.32 .62 5–20 5–18 Obsessive-Compulsive (5) 9.28 2.55 .64 5–20 5–18 Paranoid (5) 6.95 1.96 .62 5–20 5–14 Schizotypal (5) 6.49 1.72 .40 5–20 5–15 Schizoid (5) 7.78 2.20 .49 5–20 5–14 Note. Number of items in each scale listed in parentheses. SOBI-P = Sense of Belonging Inventory-Psychological; GSIS = Geriatric Suicide Ideation Scale; SCATI = Short Coolidge Axis II Inventory. GSIS Meaning in Life is reverse-scored.
! 41
All potential participants received a thank you postcard approximately one week after the questionnaire packet. Completion of the packet was estimated to take 30 minutes. All questionnaire packets were kept anonymous and stored in a locked, secure location. No deception was used in this study.
! CHAPTER 3
RESULTS
Means and standard deviations were calculated for all measures included in the study (see Table 4). The GSIS Suicide Ideation subscale score suggested a low level of current suicide ideation in the whole sample, with some variability (M = 11.83, SD =
4.26, range = 10 – 46), and was consistent with scores from other community-based older adult samples (e.g., Cukrowicz et al., 2011). With regard to past suicide attempt, 2.5% of the sample answered agree or strongly agree to the GSIS question, “I have tried ending my life in the past” (M = 1.12, SD = .58, range = 1 – 5). See Table 5 for Pearson correlation coefficients between suicide ideation, past suicide attempt, and the demographic, clinical, and personality variables used in the study.
Exploratory Principle Components Analysis
The first aim of the current study was to investigate the structure and validity of the Interpersonal Needs Questionnaire (INQ) in the older adult population. To determine the number of factors in the latent structure of the INQ in the present sample of older adults (hypothesis 1a), the 18 items of the original version of the INQ were subjected to a principal components analysis (PCA) using SPSS Version 19. The suitability of the data for PCA was assessed. Inspection of the correlation matrix revealed the presence of several coefficients of .3 and above and only one coefficient greater than .9 (r = .91 for items 1 and 2), suggesting that the data were suitable for factor analysis and ! 43
Table 5
Pearson Correlations between Suicide Ideation, Past Suicide Attempt, and Demographic Clinical, and Personality Variables
Suicide Past Suicide Ideation Attempt Demographic Variables Age .03 .10 Gender -.02 -.04 Race < .01 -.11 Education -.01 -.11 Relationship -.11 -.08 Religion -.05 -.01
Clinical Variables Hopelessness .65*** .21** Loneliness .59*** .31*** Physical Health -.09 -.02 Depression .73*** .20** SOBI-P Belonging -.63*** -.39*** Self-Esteem -.54*** -.22***
SCATI Antisocial .32*** .27*** Avoidant .45*** .37*** Borderline .60*** .30*** Dependent .16* .08 Histrionic .18** .13* Narcissistic .14* .08 Obsessive-Compulsive .16* .18** Paranoid .41*** .25*** Schizotypal .29*** .27*** Schizoid .40*** .23*** Note. SOBI-P = Sense of Belonging Inventory-Psychological; SCATI = Short Coolidge Axis II Inventory. *p < .05, ** p < .01, *** p < .001 multicollinearity was not a problem (Field, 2005; see Table 6). In addition, the Kaiser-
Meyer-Oklin value (a measure of sampling adequacy) was .86, exceeding the recommended value of .5, and Bartlett’s Test of Sphericity reached statistical significance
(p < .001), also providing evidence that factor analysis was appropriate (Field, 2005). ! 44
18 1.00
17 .72 1.00
16 .81 .67 1.00
15 .06 .12 .09 1.00
14 .45 .27 .33 .26 1.00
13 .37 .09 .63 .71 .62 1.00
12 .26 .35 .22 .04 .16 .21 1.00
11 .21 .70 .35 .20 .62 .65 .57 1.00
10 .81 .09 .60 .17 .10 .59 .58 .46 1.00
9 .78 .73 .11 .58 .13 .08 .48 .52 .47 1.00
8 .05 .14 .13 .17 .26 .13 .23 .24 .04 .03 - 1.00
7 .57 .09 .14 .09 .19 .11 .11 .11 .12 .06 .03 1.00
6 .12 .16 .15 .11 .21 .10 .21 .21 .10 .14 .13 .18 1.00
5 .17 .06 .05 .44 .42 .41 .16 .41 .16 .04 .36 .36 .37 1.00
4 .14 .28 .24 .36 .15 .11 .21 .23 .18 .39 .20 .15 .14 .11 1.00
3 .45 .24 .26 .30 .32 .16 .17 .22 .18 .19 .42 .25 .18 .21 .12 1.00
2 .41 .45 .17 .42 .28 .47 .20 .15 .25 .26 .34 .37 .22 .22 .19 .17 1.00
1
.91 .41 .45 .14 .43 .30 .42 .15 .13 .27 .24 .35 .36 .20 .25 .21 .19 1.00 item Correlation Matrix for the Interpersonal Needs Questionnaire -
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Table 6 Table Inter
! 45
PCA with Varimax rotation revealed the presence of four components with eigenvalues exceeding 1, explaining 28.0%, 15.3%, 11.0%, and 9.6% of the variance respectively (referred to as Model 1). Cronbach alpha coefficients for the components were .91, .73, .60, and .62 respectively. Because PCA with Varimax rotation assumes that the components are not correlated and there were small-to-moderate correlations between the components (see Table 7), a PCA with Oblimin rotation was also performed, with results that were nearly identical to the PCA with Varimax rotation. An inspection of the
Table 7
Component Correlations for Four-Component Interpersonal Needs Questionnaire Belonging Burden-1 Burden-2 Burden-3 Belonging 1.00 .31 .27 .10 Burden-1 1.00 .42 .45 Burden-2 1.00 .28 Burden-3 1.00 Note. Burden-1 = Social Hindrance; Burden-2 = Social Disconnection; Burden-3 = Social Liability. scree plot revealed a clear break after the second component and a slight break after the fourth component, suggesting that either a four-component or a two-component solution may be appropriate (see Figure 3). Items that loaded on to Component 1 (see Table 8) appeared to represent a sense of give-and-take in interpersonal relationships, one important hypothesized component of belonging (see Figure 2), and was named Social
Reciprocity. Note that all items on this subscale were reverse-scored so that high scores indicate low reciprocity. Items that loaded on to Component 2 appeared to represent a sense of holding others back or hindering others in some way and was named Social
Hindrance. Items that loaded on to Component 3 appeared to represent a sense of ! 46
Figure 3. Scree plot for Interpersonal Needs Questionnaire disconnection and isolation from others and was named Social Disconnection. Items that loaded on to Component 4 appeared to represent a sense of being a liability or detriment to others and was named Social Liability.
Because the INQ was originally designed to measure two constructs (i.e., burdensomeness and belongingness) and the scree plot revealed a clear break after the second component, a PCA with Varimax rotation was performed with two components extracted (referred to as Model 2). The two-component solution explained 50.3% of the variance, with Component 1 contributing 28.3% and Component 2 contributing 22.0%.
Cronbach alpha coefficients were .91 and .75 respectively. Component 1 contained ! 47 identical items as those in Component 1 of the first PCA, reflecting social reciprocity, which could be construed as belongingness (see Table 8). Items from Components 2, 3,
Table 8
Component Loadings for Exploratory Principle Components Analysis With Varimax Rotation
Model 1 Model 2 4 component 2 component Item solution solution 1 2 3 4 1 2 5. contribute to well-being .552 .112 .023 .072 .546 .126 of others 9. matter to others .802 .005 -.040 .233 .776 .090 10. others care about me .849 -.055 -.004 .248 .822 .069 11. I belong .850 .089 .183 .126 .848 .211 13. caring and supportive .792 .228 .167 -.008 .802 .245 friends 16. others I can turn to .803 .170 .054 -.093 .812 .108 17. close to others .829 .116 .183 -.136 .848 .110 18. satisfying interaction .749 .134 .157 -.229 .776 .068 1. others would be better .140 .858 .149 .203 .143 .802 off 2. others would be happier .135 .837 .168 .242 .136 .814 3. failed others .127 .451 .372 .264 .135 .637 4. burden on society .069 .524 .341 .237 .078 .664 6. burden on others .130 .699 -.040 -.063 .141 .470 12. rarely interact with .116 .082 .587 .170 .140 .424 others 14. disconnected from .207 .307 .759 -.062 .264 .567 others 15. feel like an outsider .024 .027 .758 .084 .065 .427 7. others wish they could .058 .159 .073 .794 -.001 .520 be rid of me 8. make things worse for .008 .285 .229 .782 -.039 .684 others Note. Component loadings > .40 are in boldface. and 4 from the previous PCA combined into one component, which reflected the concept of burdensomeness. Although each item had a high loading on only one component, some of the items loaded differently than in the original 18-item INQ or the new 15-item ! 48
INQ (see Table 9). For example, in the 18-item version, item 5 was on the
Burdensomeness subscale, but loaded onto the Belongingness scale in the current
Table 9
Item Loadings for Different Versions of the INQ
Model 1 Model 2 Model 3 Model 4 4-component 2-component 18-item 15-item Item solution solution INQ INQ2
1. others would be better off B1 B B B 2. others would be happier B1 B B B 3. failed others B1 B B 4. burden on society B1 B B B 5. contribute to well-being of A A B others1 6. burden on others B1 B B 7. others wish they could be rid of B3 B B B me 8. make things worse for others B3 B B B 9. matter to others1 A A A 10. others care about me1 A A A A 11. I belong1 A A A A 12. rarely interact with others B2 B A A 13. caring and supportive friends1 A A A A 14. disconnected from others B2 B A A 15. feel like an outsider B2 B A A 16. others I can turn to1 A A A A 17. close to others1 A A A A 18. satisfying interaction1 A A A A Note. A = Belonging; B = Burden; B1 = Social Hindrance; B2 = Social Disconnection; B3 = Social Liability; INQ = Interpersonal Needs Questionnaire. 1Reverse-scored item. 2One item from this scale is missing in the current study. analysis. Item 5 was removed from the 15-item INQ. In addition, items 12, 14, and 15 originally loaded onto the Belongingness subscale, but loaded onto the Burdensomeness scale in the current analysis. Means and standard deviations were calculated for all versions of the INQ (see Table 10).
! 49
Table 10
Psychometric Properties of Different Versions of the INQ
Range Variable M SD ! Potential Actual
4-component solution (Model 1) Belonging (8) 17.83 11.59 .91 8–56 8–56 Burden-1 (5) 6.43 2.96 .73 5–35 5–24 Burden-2 (3) 6.14 3.59 .60 3–21 3–21 Burden-3 (2) 2.51 1.60 .62 2–14 2–14
2-component solution (Model 2) Belonging (8) 17.83 11.59 .91 8–56 8–56 Burden (10) 15.05 6.37 .75 10–70 10–52
18-item INQ (Model 3) Belonging (10) 21.27 11.85 .87 10–70 10–63 Burden (8) 11.47 4.83 .67 8–56 8–38
15-item INQ (Model 4) Belonging (9) 19.40 10.69 .85 9–63 9–58 Burden (6) 1 5.98 2.75 .76 6–42 5–24 Note. Number of items in each scale listed in parentheses. INQ = Interpersonal Needs Questionnaire; Burden-1 = Social Hindrance; Burden-2 = Social Disconnection; Burden- 3 = Social Liability. 1One item is missing from this scale in the current study.
Confirmatory Factor Analysis
To determine which of the models best fit the data, a confirmatory factor analysis
(CFA) was performed using SPSS AMOS Version 19 (continuation of hypothesis 1a).
Model 1 was the four-component solution derived from the first PCA. Model 2 was the two-component solution derived from the second PCA. Model 3 was the original 18-item version of the INQ. Model 4 was the recently-developed 15-item version of the INQ. The chi-square statistic, comparative fit index (CFI), and root mean square residual (RMSEA) were used to determine goodness-of-fit (see Table 11). The chi-square statistic should be low and non-significant if the model is a good fit to the data (Roberts, 1999). Values of ! 50
.90 and greater for the CFI and values of .05 and less for the RMSEA indicate a better fit of the model (Roberts, 1999). In addition, higher values for the PCFI, a parsimony- adjusted measure of fit, are more desirable (Roberts, 1999).
Table 11
Confirmatory Factor Analysis for the Interpersonal Needs Questionnaire !2 CFI RMSEA PCFI Model 1 539.56* .83 .11 .63 Model 2 660.71* .78 .12 .61 Model 3 765.23* .74 .13 .58 Model 4 492.25* .78 .14 .57 * p < .001
CFA for Model 1 revealed a significant chi-square value, !2 (129) = 539.56, p <
.001, .83 for CFI, .11 for RMSEA, and .63 for PCFI. CFA for Model 2 revealed a significant chi-square value, !2 (134) = 660.71, p < .001, .78 for CFI, .12 for RMSEA, and .61 for PCFI. CFA for Model 3 revealed a significant chi-square value, !2 (134) =
765.23, p < .001, .74 for CFI, .13 for RMSEA, and .58 for PCFI. CFA for Model 4 revealed a significant chi-square value, !2 (76) = 492.25, p < .001, .78 for CFI, .14 for
RMSEA, and .57 for PCFI. Although none of the models met suggested criteria for adequate goodness-of-fit, out of the two-component models, Model 2 appeared to be the best fit for these data, with Model 4 as a close second best fit.
Construct Validity Analysis
To examine the validity of the INQ, each two-component model (described previously) was assessed by correlating Burden and Belonging with measures that were hypothesized to be related to each. Burden was expected to be positively correlated with the Geriatric Suicide Ideation Loss of Personal and Social Worth subscale (GSIS LPSW) and negatively correlated with the Single-Item Self-Esteem Scale (SISE; hypothesis 1b). ! 51
Belonging was expected to be negatively correlated with the Sense of Belonging
Inventory-Psychological subscale (SOBI-P) and positively correlated with the Three-item
Loneliness Scale (3LS; hypothesis 1c). In addition, the correlations between Burden and
GSIS LPSW and Burden and SISE were expected to be stronger than the correlations between Belonging and GSIS LPSW and Belonging and SISE (hypothesis 1d). The correlations between Belonging and SOBI-P and Belonging and 3LS were expected to be stronger than the correlations between Burden and SOBI-P and Burden and 3LS
(hypothesis 1e).
Construct validity was first assessed with Pearson correlations using the components derived from the two-component PCA (Model 2; see Table 12). Burden and
Belonging were moderately correlated with each other (r = .33). As hypothesized, the
Burden subscale was positively and strongly correlated with GSIS LPSW (representing liability in the theoretical model, r = .66; see Figure 2) and moderately negatively correlated with SISE (representing the opposite of self-hate in the theoretical model, r =
-.38; hypothesis 1b). Also as hypothesized, the Belonging scale was negatively correlated with SOBI-P (representing reciprocal care in the theoretical model, r = -.38) and positively correlated with 3LS (representing loneliness in the theoretical model, r = .37; hypothesis 1c). Unexpectedly, Burden had stronger correlations with SOBI-P (r = -.65) and 3LS (r = .62) than the Belonging subscale (hypothesis 1e). Also unexpectedly, the correlation between SISE and Belonging was only slightly lower than the correlation between SISE and Burden (-.33 and -.38 respectively; hypothesis 1d). Thus, the Pearson correlations provide mixed evidence for the construct validity of Model 2. ! 52
Because the analysis did not provide clear support for the validity of the two components derived from the PCA (Model 2), the validity of the original subscales from the 18-item INQ was assessed with Pearson correlations (Model 3 described previously; see Table 12). These results were also mixed. In this model, Burden and Belonging were strongly and positively correlated with each other (r = .51). The Burden subscale was positively correlated with GSIS LPSW (r = .62) and negatively correlated with SISE (r =
-.31; hypothesis 1b). The Belonging scale was negatively correlated with SOBI-P (r =
-.50) and positively correlated with 3LS (r = .49; hypothesis 1c). Unexpectedly, Burden had nearly the same strength of correlation with SOBI-P (r = -.49) and a slightly stronger correlation with 3LS (r = .52) than the Belonging subscale (hypothesis 1e). Also unexpectedly, the correlation between Belonging and SISE was slightly stronger than the correlation between Burden and SISE (-.39 and -.31 respectively; hypothesis 1d).
The validity of the subscales from the 15-item INQ (Model 4 described previously) was assessed with Pearson correlations as well, with mixed results (see Table
12). In this model, Burden and Belonging were moderately correlated with each other
(r = .33). The Burden subscale was positively correlated with GSIS LPSW (r = .52) and negatively correlated with SISE (r = -.24; hypothesis 1b). The Belonging scale was negatively correlated with SOBI-P (r = -.52) and positively correlated with 3LS (r = .49; hypothesis 1c). In this model, the relative strength of the correlations between the INQ subscales and the validity measures was as expected, except for the correlation between
Belonging and SISE, which was slightly stronger than the correlation between Burden and SISE (-.39 and -.24 respectively; hypotheses 1d and 1e). ! 53
< ** *** ***
*** *** *** p
.20 .24 .41
.52 .43 .49 Burden - - P = Sense
-
= .33 r Model 4 Model < .05, ** p *** ***
*** *** *** *** SOBI .39 .52 .45 .49 .44 .28 - - Belonging
*** ***
*** *** *** *** Esteem Scale; Scale; Esteem - 4
.31 .49
.62 .52 .53 .2 Burden - -
= .51 r Item Loneliness Scale. * Scale. Loneliness Item Item Self Item Model 3 Model - - *** ***
*** *** *** *** .39 .50 .45 .49 .45 .25 - - Belonging
; 3LS = Three ; SISE = Single *** ***
*** *** *** ***
.38 .65
.66 .62 .54 .34 Burden - -
= .33
r Model 2 Model ** *** ***
*** *** *** .18 .33 .38 .35 .37 .39 - - Belonging Psychological Subscale Psychological
-
P)
-
(GSIS) (SISE) (3LS) (SOBI < .001
Correlations BetweenVariables Burden, Belonging, Validity Correlations and Measures, Predicted p esteem of Worth - . GSIS = Geriatric Suicide Ideation Scale; Scale; Ideation Suicide . GSIS = Geriatric elonging Inventory elonging Loss Belonging Table 12 Table Pearson Measures Validity Self Loneliness Variables Predicted Ideation Suicide Attempt Suicide Note of B .01. ***
! 54
In an exploratory analysis, Pearson correlations were used to establish the relationships between Burden and Belonging (for Models 2, 3, and 4) and various demographic and clinical variables (see Table 13). Gender was coded into 1 (male) and 2
(female). Race was coded into 0 (not European American) and 1 (European American).
Relationship status was coded into 0 (not in a relationship) and 1 (in a relationship; i.e., domestic partnership and married). Results showed that Burden and Belonging were not significantly correlated with age, gender, race, education, or religion/spirituality. Burden and Belonging had small significant negative correlations with relationship status, meaning that not being in a relationship was slightly more associated with Burden and
Belonging than being in a relationship. Burden and Belonging had medium-to-large positive correlations with hopelessness, depression, and meaning in life (note that GSIS
Perceived Meaning in Life was reverse-scored, so that these relationships were actually negative). Burden had small significant negative relationships with physical health, meaning that poorer physical health was associated with higher Burden. Belonging was not correlated with physical health. Burden and Belonging had small-to-medium significant positive correlations with several different types of personality disorder features (based on the SCATI), including Antisocial, Avoidant, Borderline, Dependent,
Obsessive-Compulsive, Paranoid, Schizotypal, and Schizoid (see Table 14). Burden and
Belonging were mostly unrelated to Histrionic and Narcissistic personality disorder features.
To further assess the construct validity of Burden among older adults, a single open-item was asked: “Please list the top three life events or circumstances that might make you feel like you are a burden to others (e.g., family members, friends), even if just ! 55
.04 .03 .11 .01 .05 .02
- - - -
.23*** .47*** .45*** .31*** Burden -
Model 4 Model
.04 .10 .09 .06 .07 .01 .15* - - - - - .50*** .42*** .54*** Belonging
.06 .10 .06 .08 .02 - -
.18** .25*** - .58*** .53*** .42*** Burden - < .001
p
Model 3 Model
.03 .09 .08 .05 .06 .03 .15* - - - - - .50*** .42*** .54*** Belonging < .01. *** p
6** .07 .08 .04 < .05, ** .06 .03 - - -
.1 p .24*** - .59*** .57*** .46*** Burden -
Model 2 Model
.02 .10 .07 .02 .06 .04 .15* - - - - - .33** .43*** .49*** Belonging
rrelations Between Burden, Belonging, andVariables andrrelations Demographic Clinical
. Meaning in Life is reverse scored. * scored. is reverse Life in . Meaning
Table 13 Table Co Pearson Age Gender Race Education Relationship Religion Hopelessness Health Physical Depression Life in Meaning Note
! 56
.08 .12
.18** .17** .16** .26*** .29*** .30*** .31*** .35*** Burden
Model 4 Model
.12 .14* .28*** .50*** .41*** .30*** .22*** .42*** .22*** .43*** Belonging
.01 .04 .15* .13*
.20** .27*** .39*** .34*** .37*** .41*** Burden
Model 3 Model
.12 .14* .21** .28*** .48*** .40*** .28*** .22*** .41*** .41*** Belonging
.07 .12*
.35*** .50*** .45*** .29*** .23*** .41*** .25*** .49*** Burden
Model 2 Model
.11 .06 .16* .14* .20** .22*** .36*** .30*** .35*** .32*** een Belonging, Burden, and Personality Disorder Features Burden,Disorder Belonging, andeen Personality Belonging < .001 p
< .01. *** p
Compulsive
-
< .05, ** p Table 14 Table Betw Correlations Pearson Antisocial Avoidant Borderline Dependent Histrionic Narcissistic Obsessive Paranoid Schizotypal Schizoid *
! 57 a little bit.” Space was provided for up to three answers. Many participants left the item blank or indicated they weren’t a burden to others (n = 116, 40.8%). Those who responded (n = 168, 59.2%) supplied one, two, or three answers. Out of a total of 380 answers, physical illness or disability was the most frequent answer (n = 117, 30.7%), followed by loss or lack of support (n = 56, 14.7%), finances (n = 42, 11.1%), cognitive or mental health issues (n = 38, 10.0%), inability to care for oneself or others (n = 38,
10.0%), miscellaneous (n = 31, 8.2%), reliance on others (n = 28, 7.4%), transportation issues (n = 25, 6.6%), and age (n = 5, 1.3%; see Figure 4). Examples of items from each
!
Figure 4. Causes of burden to others. A total of 380 answers were supplied by 168 participants (59.2% of the sample). Percentages in figure represent proportion of total answers for each category. category are as follows: “Stroke resulting in full or partial paralysis” (physical illness or disability); “If I lost my spouse, my family could feel burdened” (loss or lack of support);
“My finances are bad at this time” (finances); “Onset of Alzheimer’s disease” (cognitive or mental health issues); “If I was too sick or injured to care for myself” (inability to care ! 58 for self or others); “Having to move to unknown territory” (miscellaneous); “Asking family, friends, to do things for me” (reliance on others); “Unable to continue driving”
(transportation issues); and “Old age” (age).
Hierarchical Multiple Regression Analysis for Current Suicide Ideation
Hierarchical regression was used to determine if burdensomeness and belongingness predicted current suicide ideation above and beyond other suicide risk factors and if the interaction between these two variables contributed a significant amount of variance (hypothesis 2c). Model 1 was excluded from the regression analyses because
Burden in that model was split into three components, which would have made the interpretation of the interaction between Burden and Belonging nearly impossible.
Regardless of how Burden and Belonging were measured, they each had medium-to-large positive correlations with current suicide ideation (see Table 12; hypotheses 2a and 2b).
Both Burden and Belonging were centered before they were entered into the regressions, and the centered variables were used for the interaction. Predictor variables were chosen based on past research on important risk and protective factors for suicide among older adults: age (CDC, 2010), gender (CDC, 2010), religious beliefs/spirituality (June, Segal,
Coolidge, & Klebe, 2009; Marty et al., 2010), hopelessness (Fiske & Arbore, 2000 –
2001), borderline personality traits (Heisel, 2006; Segal, Marty, Meyer, & Coolidge,
2010), loneliness (Heisel, 2006; Waern et al., 2003), depression (Heisel, 2006; Fiske &
Arbore, 2000 – 2001), physical functioning (Heisel, 2006), and self-esteem (Chatard,
Selimbegovic, & N’Dri Konan, 2009).
A preliminary analysis assessed the normality of the predicted variable, current suicide ideation (as measured by the 10-item GSIS Suicide Ideation subscale). The ! 59
Kolmogorov-Smirnov test statistic indicated the distribution deviated from normality,
D(272) = .33, p < .05, and was positively skewed (skewness = 4.24, kurtosis = 23.56). A logarithmic transformation slightly improved the shape of the distribution; however, the
Kolmogorov-Smirnov test statistic indicated the distribution still deviated from normality, D(272) = .34, p < .05, and remained positively skewed (skewness = 2.64, kurtosis = 8.04; see Figure 5). Square root and inverse transformations were computed, but significantly deviated from normality also. The logarithmic transformed suicide ideation variable was used for all analyses.
Figure 5. Histogram of transformed current suicide ideation variable. ! 60
Further preliminary analyses were conducted for all the regressions predicting current suicide ideation to detect the assumptions of normality of the standardized residuals, linearity of the standardized residuals, homoscedasticity of the standardized residuals, and multicollinearity of the predictor variables. For all of the regression models, the assumptions of normality, linearity, and multicollinearity were met.
However, because of the positively skewed distribution, it was questionable whether the assumption of homoscedasticity was met. For all of the regression models, some cases exceeded the critical value for Mahalanobis distance (27.88), indicating the presence of multivariate outliers; however, all of the values for Cook’s distance were below 1.0, suggesting that no cases were having undue influence on the results for the model as a whole. Cases were removed listwise, resulting in slightly different sample sizes for each regression; however, all the regressions had more than the recommended number of cases
(Field, 2005). Criteria used for determining magnitude of effect sizes was based on recommendations from Cohen (1992): .02 (small), .15 (medium), and .35 (large).
Model 2 Predicting Current Suicide Ideation. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 65.1% of the variance in current suicide ideation (see Table 15). Depression, hopelessness, borderline personality traits, and loneliness were all significant positive predictors, whereas self- esteem was the only significant negative predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 65.6%. Burden and
Belonging explained an additional 0.7% of the variance in current suicide ideation, after controlling for other risk and protective variables. Burden was an additional significant ! 61
Table 15
Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 2 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .65 .67*** Depression .01 < .01 .41*** Hopelessness < .01 < .01 .14* Borderline .01 .01 .14* Loneliness .02 < .01 .19** Self-esteem -.02 .01 -.13* Physical Health < .01 < .01 .09 Religion < .01 < .01 .02 Gender -.01 .01 -.06 Age < .01 < .01 .08 Step 2 .66 .01 Belonging < .01 < .01 .02 Burden < .01 < .01 .11* Step 3 .66 .01* Interaction < .01 < .01 .12* Note. N = 215. Step 3 f 2 = .03. *p < .05. **p < .01. ***p < .001. predictor in Step 2 (" = .11, p < .05). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 66.2%, explaining an additional 0.8% of the variance, with a small effect size (f 2 = .03). The interaction was significant (p < .05), such that the simultaneous presence of high perceived burdensomeness and high thwarted belongingness resulted in the highest level of current suicide ideation (see Figure 6). That is, the combination of high perceived burdensomeness and high thwarted belongingness had the strongest correlation with current suicide ideation.
Model 3 Predicting Current Suicide Ideation. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 65.1% of the variance in current suicide ideation (see Table 16). Depression, hopelessness, borderline ! 62
Figure 6. Interaction between Model 2 Burden and Belonging for current suicide ideation. personality traits, and loneliness were all significant positive predictors, whereas self- esteem was the only significant negative predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 66.7%. Burden and
Belonging explained an additional 1.8% of the variance in current suicide ideation, after controlling for other risk and protective variables. Burden was an additional significant predictor in Step 2 (! = .18, p < .01). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 67.5%, explaining an additional 0.9% of the variance, with a small effect size (f 2 = .02). The interaction was significant (p < .05), ! 63 such that the simultaneous presence of high perceived burdensomeness and high thwarted belongingness resulted in the highest level of current suicide ideation.
Table 16
Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 3 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .65 .67*** Depression .01 < .01 .41*** Hopelessness < .01 < .01 .14* Borderline .01 .01 .14* Loneliness .02 < .01 .19** Self-esteem -.02 .01 -.13* Physical Health < .01 < .01 .09 Religion < .01 < .01 .02 Gender -.01 .01 -.06 Age < .01 < .01 .08 Step 2 .67 .02** Belonging < .01 < .01 -.04 Burden < .01 < .01 .18** Step 3 .68 .01* Interaction < .01 < .01 .12* Note. N = 215. Step 3 f 2 = .02. *p < .05. **p < .01. ***p < .001.
Model 4 Predicting Current Suicide Ideation. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 64.8% of the variance in current suicide ideation (see Table 17). Depression, hopelessness, borderline personality traits, and loneliness were all significant positive predictors, whereas self- esteem was the only significant negative predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 66.4%. Burden and
Belonging explained an additional 1.8% of the variance in current suicide ideation, after controlling for other risk and protective variables. Burden was an additional significant predictor in Step 2 (" = .16, p < .01). After the entry of the interaction at Step 3, the total ! 64 variance explained by the model as a whole was 66.7%, explaining an additional 0.4% of the variance, with a small effect size (f 2 < .01). The interaction was not significant (p >
.05).
Table 17
Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using Model 4 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .65 .66*** Depression .01 < .01 .40*** Hopelessness < .01 < .01 .14* Borderline .01 < .01 .14* Loneliness .01 < .01 .18** Self-esteem -.02 < .01 -.13** Physical Health < .01 < .01 .09 Religion < .01 < .01 .02 Gender -.01 .01 -.05 Age < .01 < .01 .09 Step 2 .66 .02** Belonging < .01 < .01 -.02 Burden < .01 < .01 .16** Step 3 .67 < .01 Interaction < .01 < .01 .09 Note. N = 218. Step 3 f 2 < .01. *p < .05. **p < .01. ***p < .001.
Validity Measures Predicting Current Suicide Ideation. A regression using
SOBI-P for Belonging and GSIS LPSW for Burden was also performed because of the somewhat questionable validity of the INQ in the current sample. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 62.7% of the variance in current suicide ideation (see Table 18). Depression, hopelessness, borderline personality traits, loneliness, physical health, and age were all significant positive predictors, whereas self-esteem was the only significant negative predictor. After
! 65
Table 18
Hierarchical Multiple Regression Analysis Predicting Current Suicide Ideation Using SOBI-P and GSIS LPSW Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .63 .64*** Depression .01 < .01 .43*** Hopelessness < .01 < .01 .13* Borderline .01 < .01 .12* Loneliness .01 < .01 .15** Self-esteem -.02 .01 -.16** Physical Health < .01 < .01 .11* Religion < .01 < .01 .01 Gender -.02 .01 -.08 Age < .01 < .01 .10* Step 2 .68 .05*** Belonging < -.01 < .01 -.07 Burden .01 < .01 .40*** Step 3 .69 .01** Interaction < .01 < .01 -.16** Note. SOBI-P = Sense of Belonging Inventory – Psychological subscale; GSIS LPSW = Geriatric Suicide Ideation Scale – Loss of Personal and Social Worth subscale. N = 220. Step 3 f 2 = .04. *p < .05. **p < .01. ***p < .001. entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 68.0%. Burden and Belonging explained an additional 5.3% of the variance in current suicide ideation, after controlling for other risk and protective variables. Burden was an additional significant predictor in Step 2 (" = .40, p < .001). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 69.0%, explaining an additional 1.1% of the variance, with a small effect size (f 2 = .04). The interaction was significant (p < .01), such that the simultaneous presence of higher feelings of low self-worth (i.e., perceived burdensomeness) and lower feelings of high belongingness (i.e., thwarted belongingness) resulted in the highest level of current suicide ideation. ! 66
Exploratory Regression Analyses. Exploratory hierarchical multiple regression analyses were performed with a subsample of participants (n = 71) who scored 11 or higher on the current suicide ideation measure (GSIS Suicide Ideation). All four of the
Models predicting current suicide ideation (Model 2, Model 3, Model 4, and the validity measures) were calculated with this “higher risk” subsample. These regression models accounted for approximately the same amount of total variance as the models calculated on the whole sample (66.4% for Model 2, 68.1% for Model 3, 67.9% for Model 4, and
67.8% for the validity measures). None of the interactions between Burden and
Belonging in Step 3 were significant. Given the small sample size in the exploratory analyses, the lack of significant results may be attributed to reduced statistical power.
Hierarchical Multiple Regression Analysis for Past Suicide Attempt
A second series of hierarchical regressions was used to determine if burdensomeness and belongingness predicted past suicide attempt above and beyond other suicide risk factors and if the interaction between these two variables contributed a significant amount of variance (hypothesis 2f). Again, Model 1 was excluded from the regression analyses because Burden in that model was split into three components, complicating the interpretation of the interaction between Burden and Belonging.
Regardless of how Burden and Belonging were measured, they each had small-to- medium positive correlations with past suicide attempt (see Table 12; hypotheses 2d and
2e). As with the previous set of regressions, both Burden and Belonging were centered before they were entered into the regressions, and the centered variables were used for the interaction. The same predictor variables used in the previous series of regressions were used for the current series of regressions. ! 67
A preliminary analysis assessed the normality of the predicted variable, past suicide attempt (as measured by the GSIS item, “I have tried ending my life in the past”).
The Kolmogorov-Smirnov test statistic indicated the distribution deviated from normality, D(276) = .53, p < .05, and was positively skewed (skewness = 5.42, kurtosis =
29.54). A logarithmic transformation slightly improved the shape of the distribution; however, the Kolmogorov-Smirnov test statistic indicated the distribution still deviated from normality, D(276) = .54, p < .05, and remained positively skewed (skewness = 4.85, kurtosis = 22.87; see Figure 7). Square root and inverse transformations were computed, but significantly deviated from normality also. The logarithmic transformed past suicide attempt variable was used for all analyses.
Further preliminary analyses were conducted for all the regressions predicting past suicide attempt to detect the assumptions of normality of the standardized residuals, linearity of the standardized residuals, homoscedasticity of the standardized residuals, and multicollinearity of the predictor variables. For all of the regression models, the assumptions of normality, linearity, and multicollinearity were met. However, because of the positively skewed distribution, it was questionable whether the assumption of homoscedasticity was met. For all of the regression models, some cases exceeded the critical value for Mahalanobis distance (27.88), indicating the presence of multivariate outliers; however, all of the values for Cook’s distance were below 1.0, suggesting that no cases were having undue influence on the results for the model as a whole. Cases were removed listwise, resulting in slightly different sample sizes for each regression; however, all the regressions had more than the recommended number of cases (Field, ! 68
2005). Criteria used for determining magnitude of effect sizes was based on recommendations from Cohen (1992): .02 (small), .15 (medium), and .35 (large).
Figure 7. Histogram of transformed past suicide attempt variable.
Model 2 Predicting Past Suicide Attempt. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 9.2% of the variance in past suicide attempt (see Table 19). Borderline personality was the only significant (positive) predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 11.1%. Burden and Belonging explained an additional 2.6% of the variance in past suicide attempt, after controlling for other risk and protective variables. Burden was an additional significant predictor in Step 2 (! = ! 69
Table 19
Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 2 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .09 .13** Depression < .01 < .01 .09 Hopelessness < .01 < .01 -.17 Borderline .02 < .01 .23* Loneliness .01 < .01 .13 Self-esteem -.01 .01 -.10 Physical Health < .01 < .01 .06 Religion < .01 < .01 .02 Gender -.02 .01 -.09 Age < .01 < .01 .09 Step 2 .11 .03* Belonging < .01 < .01 .08 Burden < .01 < .01 .19* Step 3 .17 .06*** Interaction < .01 < .01 .32*** Note. N = 219. Step 3 f 2 = .08. *p < .05. **p < .01. ***p < .001.
.19, p < .05). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 17.4%, explaining an additional 6.3% of the variance, with a small effect size (f 2 = .08). The interaction was significant (p < .001), such that the simultaneous presence of high perceived burdensomeness and high thwarted belongingness resulted in the highest occurrence of past suicide attempt (see Figure 8).
Model 3 Predicting Past Suicide Attempt. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 9.2% of the variance in past suicide attempt (see Table 20). Borderline personality was the only significant (positive) predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 10.1%. Burden and Belonging explained ! 70 an additional 1.7% of the variance in past suicide attempt, after controlling for other risk and protective variables. Neither Burden nor Belonging was a significant predictor in
Figure 8. Interaction between Model 2 Burden and Belonging for Past Suicide Attempt. Step 2, but Belonging approached significance (p = .07). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 11.2%, explaining an additional 1.5% of the variance, with a small effect size (f 2 = .01). The interaction was not significant (p > .05).
Model 4 Predicting Past Suicide Attempt. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 7.2% of the ! 71
Table 20
Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 3 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .09 .13** Depression < .01 < .01 .09 Hopelessness < -.01 < .01 -.17 Borderline .02 .01 .23* Loneliness .01 .01 .13 Self-esteem -.01 .01 -.10 Physical Health < .01 < .01 .06 Religion < .01 < .01 .02 Gender -.02 .01 -.09 Age < .01 < .01 .09 Step 2 .10 .02 Belonging < .01 < .01 .15 Burden < .01 < .01 .03 Step 3 .11 .02 Interaction < .01 < .01 .16 Note. N = 219. Step 3 f 2 = .01. *p < .05. **p < .01. variance in past suicide attempt (see Table 21). Borderline personality was the only significant (positive) predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 7.8%. Burden and Belonging explained an additional 1.4% of the variance in past suicide attempt, after controlling for other risk and protective variables. Neither Burden nor Belonging was a significant predictor in
Step 2, but Belonging approached significance (p = .08). After the entry of the interaction at Step 3, the total variance explained by the model as a whole was 10.0%, explaining an additional 2.5% of the variance, with a small effect size (f 2 = .02). The interaction was significant (p < .05), such that the simultaneous presence of high perceived burdensomeness and high thwarted belongingness resulted in the highest occurrence of past suicide attempt.
! 72
Table 21
Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using Model 4 Belonging and Burden Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .07 .11** Depression < .01 < .01 .05 Hopelessness < -.01 < .01 -.13 Borderline .02 < .01 .23* Loneliness .01 < .01 .09 Self-esteem -.01 .01 -.12 Physical Health < .01 < .01 .07 Religion < .01 < .01 < .01 Gender -.01 .01 -.07 Age < .01 < .01 .12 Step 2 .08 .01 Belonging < .01 < .01 .14 Burden < .01 < .01 .01 Step 3 .10 .03* Interaction < .01 < .01 .22* Note. N = 222. Step 3 f 2 = .02. *p < .05. **p < .01.
Validity Measures Predicting Past Suicide Attempt. A regression using SOBI-
P for Belonging and GSIS LPSW for Burden was also performed because of the somewhat questionable validity of the INQ in the current sample. Depression, hopelessness, borderline personality traits, loneliness, self-esteem, physical functioning, religious beliefs/spirituality, gender, and age were entered at Step 1, explaining 8.4% of the variance in past suicide attempt (see Table 22). Borderline personality was the only significant (positive) predictor. After entry of Burden and Belonging at Step 2, the total variance explained by the model as a whole was 15.0%. Burden and Belonging explained an additional 7.1% of the variance in past suicide attempt, after controlling for other risk and protective variables. Belonging was an additional significant predictor in Step 2 (" =
-.40, p < .001). After the entry of the interaction at Step 3, the total variance explained by ! 73 the model as a whole was 14.7%, explaining an additional 0.01% of the variance, with a small effect size (f 2 < .01). The interaction was not significant (p > .05).
Table 22
Hierarchical Multiple Regression Analysis Predicting Past Suicide Attempt Using SOBI-P and GSIS LPSW Measures
Predictor Adj. R2 !R2 B SE B " Step 1 .08 .12** Depression < .01 < .01 .01 Hopelessness < -.01 < .01 -.11 Borderline .02 < .01 .25** Loneliness < .01 < .01 .11 Self-esteem < -.01 < .01 -.13 Physical Health < .01 < .01 .07 Religion < .01 < .01 .02 Gender -.02 .01 -.08 Age < .01 < .01 .11 Step 2 .15 .07*** Belonging < -.01 < .01 -.40*** Burden < .01 < .01 -.03 Step 3 .15 < .01 Interaction < .01 < .01 -.04 Note. SOBI-P = Sense of Belonging Inventory – Psychological subscale; GSIS LPSW = Geriatric Suicide Ideation Scale – Loss of Personal and Social Worth subscale. N = 223. Step 3 f 2 < .01. **p < .01. ***p < .001.
CHAPTER 4
DISCUSSION
Structure and Validity of the Interpersonal Needs Questionnaire
The first aim of the present study was to examine the structure and validity of a previously-developed measure of burdensomeness and belongingness (INQ) among a large sample of community-dwelling older adults. An exploratory principle components analysis (PCA) indicated that either a four-component or a two-component solution would be appropriate for the current sample of older adults. Unexpectedly, four items loaded on to components differently from what was anticipated based on other versions of the INQ (Joiner et al., 2009; Van Orden, 2009; K. A. Van Orden, personal communication, March 2, 2011; Van Orden, Witte, Gordon et al., 2008). Three of these four items were closely related to social disconnection: “I rarely interact with people who care about me,” “I feel disconnected from other people,” and “I often feel like an outsider in social gatherings.” The medium-to-large positive correlations between Belonging and
Burden in the different models examined in the present study suggests there is a significant amount of overlap between the two variables, perhaps related to shared aspects of social disconnectedness. This raises the question of whether the items belong on the Burden subscale or the Belonging subscale. Interestingly, in a recent article on the
IPTS and late-life suicide, Van Orden and Conwell (2011) conceptualized perceived burdensomeness and thwarted belongingness together as representing social disconnectedness. The definitions used for thwarted belongingness and perceived ! 75 burdensomeness both included the importance of interpersonal connectedness, with belonging signifying the absence of positively-valenced connections and burden signifying the presence of negatively valenced-connections (Van Orden & Conwell,
2011).
Confirmatory factor analysis (CFA) revealed that neither the models derived from the PCA nor the alternative models were a very good fit for the data. These findings may provide evidence against the external validity of the INQ (i.e., the results did not generalize from a younger adult sample to an older adult sample). However, because
CFA is sensitive to non-normally distributed data (Thompson, 2004), it is more likely that the floor effects found for both Burden and Belonging influenced the results. Of the two-component models tested, the model derived from the PCA and the 15-item INQ were roughly equivalent, as far as coming close to meeting suggested criteria for goodness-of fit. Thus, the findings from the PCA and CFA provide partial support for the hypothesis that the INQ would be a structurally sound instrument for use with a community-dwelling older adult population (hypothesis 1a). The findings for each model do show clear evidence of the internal reliability of the INQ, as measured by Cronbach alpha coefficients (hypothesis 1a).
For all of the two-component models, Pearson correlations showed the Burden subscale had a significant positive relationship with a measure of low self-worth and a significant negative relationship a measure of self-esteem, providing evidence of convergent and discriminant validity (hypothesis 1b). This suggests that older adults who feel worthless and/or have low self-esteem are also likely to feel that they are a liability to others. In addition, for each of the two-component models, the Belonging subscale had a ! 76 significant negative relationship with a measure of belonging and a significant positive relationship with a measure of loneliness, providing evidence of convergent and discriminant validity (hypothesis 1c). This suggests that older adults who do not feel close to others and/or feel lonely are also likely feel that they are not an integral part of any valued group. These findings support the hypothesized components of burden and belonging according to the IPTS (Van Orden et al., 2010; see Figure 2). Also providing support for the validity of the INQ were the correlations between Burden and Belonging and the demographic, clinical, and personality disorder feature variables, all of which indicated perceptions of burdensomeness and thwarted belongingness were positively associated with known risk factors for late-life suicide (e.g., hopelessness, depression).
Mitigating the strength of the validity findings were the Pearson correlations between Belonging and the measures hypothesized to correlate more strongly with
Burden (hypothesis 1d). For Model 2, the relative strengths were as hypothesized for the self-worth and self-esteem measures. For Model 3 and Model 4, Belonging had a stronger relationship with self-esteem than Burden. Also mitigating the strength of the validity results were the Pearson correlations between Burden and the measures hypothesized to correlate more strongly with Belonging (hypothesis 1e). For Model 2, Burden had stronger relationships with the belonging and loneliness measures than Belonging. For
Model 3, Burden had a stronger relationship with the loneliness measure than Belonging.
For Model 4, the relative strengths were as hypothesized for the belonging and loneliness measures. These findings provide mixed support for the hypothesized strength of relationships between Burden and Belonging and the validity measures (hypotheses 1d and 1e). This could suggest that the measures chosen to assess validity in the present ! 77 study were not a good representation of the construct in question. For example, a one- item self-esteem question was chosen to represent the opposite of self-hate, which is one of the hypothesized components of Burden according to the IPTS (see Figure 2). Perhaps the assumption that self-esteem and self-hate are on opposite ends of a continuum of beliefs about oneself is not accurate. Alternatively, a one-item measure of self-esteem may not have been sensitive enough to adequately capture self-esteem in this sample.
Another explanation might be related to the issue addressed previously regarding social disconnectedness. If social disconnectedness actually were a part of Burden rather than a part of Belonging, it would make sense that Burden would be more strongly related to loneliness. One further explanation might be that Burden and Belonging were not adequately captured by the INQ in this sample, consistent with the results of the confirmatory factor analysis.
The findings from the exploratory qualitative analysis of Burden among this sample of community-dwelling older adults were consistent with previous investigations of Burden among terminally-ill older adults (e.g., McPherson, Wilson, & Murray, 2007a).
The main theme underlying most of the responses to the open-ended question in the current study was concern about needing help from others to care for oneself, whether that was because of physical health, emotional health, or lack of resources. The responses reflect a relatively low level of perception of Burden among the current sample, and do not necessarily reflect the concept of Burden described by the IPTS. According to the
IPTS, Burden is extremely intense, to the point where a suicidal individual actually believes his death is worth more than his life to others (Van Orden & Conwell, 2011).
This low level intensity of Burden may also reflect the relatively low level of current ! 78 suicide ideation in the sample and explain why the INQ did not capture the concept of
Burden very well in the sample (i.e., there was a floor effect).
In summary, the results of the exploratory PCA, CFA, and validity correlations provided mixed evidence regarding the suitability of using the INQ with a non-clinical sample of community-dwelling older adults. The distinction between burdensomeness and belongingness warrants further investigation.
Suicidal Desire Component of the Interpersonal-Psychological Theory of Suicide
The second aim of the present study was to evaluate the applicability of the suicidal desire component of the IPTS among community-dwelling older adults.
Preliminary analyses indicated that Burden and Belonging were both significantly positively related to current suicide ideation, which is consistent with the hypotheses of the IPTS and provides support for hypotheses 2a and 2b in the current study. These findings suggest that older adults who feel like a burden to others or do not feel like they belong to any valued group, are also likely to experience suicide ideation, consistent with results from previous studies with younger adults (e.g., Van Orden et al., 2006; Van
Orden, Witte, James et al., 2008).
Hierarchical regression analyses showed that Burden was a significant positive predictor of current suicide ideation, above and beyond other known risk and protective factors, whereas Belonging was not a significant predictor. More importantly, the interaction between the two variables was significant in Model 2, Model 3, and when the validity measures (GSIS Loss of Worth and SOBI-P) were used, signifying that the presence of both Burden and Belonging together predicted current suicide ideation above and beyond other known risk and protective factors in late life suicide. The interaction ! 79 was such that individuals with both high perceived burdensomeness and high thwarted belongingness also reported the highest levels of current suicide ideation, suggesting that the joint presence of these two variables is a significant risk factor for suicide ideation.
These findings are also consistent with past research among undergraduates and adult mental health inpatients and outpatients up to age 54 (e.g., Joiner, Van Orden, Witte,
Selby et al., 2009; Van Orden, Witte, Gordon et al., 2008) and provide evidence for hypothesis 2c in the current study. The fact that Model 4 did not have a significant interaction is puzzling, but may point to the somewhat questionable validity of the INQ, as discussed above.
Preliminary analyses also indicated that Burden and Belonging were significantly positively correlated with previous suicide attempt (providing support for hypotheses 2d and 2e in the current study). These findings suggest that older adults who feel like a burden to others or do not feel like a part of any valued group, are more likely to have attempted suicide in the past, consistent with results of previous studies (e.g., Bryan et al.,
2010). These correlations were not as strong as those between Burden, Belonging, and current suicide ideation, which is logical considering the INQ asks for current feelings of burden and belonging and past suicide attempts could have occurred decades ago.
Hierarchical regression analyses showed that Burden and Belonging, for the most part, were not significant independent predictors of past suicide attempt, above and beyond other known risk and protective factors. This finding does not contradict the hypotheses of the IPTS, as the theory indicates that only those individuals with the joint presence of perceived burdensomeness and thwarted belongingness would be more likely to have made a past suicide attempt (Van Orden et al., 2010). The interaction between the ! 80 two variables was significant for Model 2 and Model 4, signifying that the presence of both Burden and Belonging together predicted past suicide attempt above and beyond other known risk and protective factors in late life suicide and providing support for hypothesis 2f in the current study. The interaction was such that individuals with both high perceived burdensomeness and high thwarted belongingness reported the highest agreement with the question regarding past suicide attempt. Findings from previous research regarding the ability of the IPTS to predict past suicide attempt are mixed. For example, Bryan et al. (2010) found that the interaction between burden and belonging did not significantly predict past suicide attempt in a military sample; however, the three-way interaction between burden, belonging, and acquired capability (the third main variable of the IPTS) did significantly predict past suicide attempt. Again, it is puzzling that some of the models had significant interactions, while others did not. This may reflect the somewhat questionable validity of the INQ or simply the low incidence of past suicide attempt in the current sample (2.5%).
In summary, the evidence for the hypothesized positive relationships between
Burden, Belonging, current suicide ideation, and past suicide attempt was strong. The results provide partial support for the hypothesized interaction between Burden and
Belonging. Given the sample had low reported rates of current suicide ideation and past suicide attempt and there were still significant findings, it appears that the suicidal desire component of the IPTS is applicable to community-dwelling older adults.
Limitations and Strengths
Several limitations of the current study should be noted. First, the study used a convenience sample mostly made up of well-educated European Americans, with a ! 81 relatively low incidence of suicide ideation or past suicide attempt. This restricts the generalizability of the findings to other culturally diverse and clinical populations, and the lack of variability in the data may have contributed to the mixed findings regarding the structure and validity of the INQ. Future research should investigate the IPTS variables among older adults of diverse racial/ethnic backgrounds, different socio- economic status’, and geographical locations (e.g., rural vs. urban areas), as Burden and
Belonging may operate differently in these groups. Nevertheless, the results from the current study can be used to compare with clinical samples of older adults in future studies. In addition, the high response rate, large sample size, and equal proportion of women and men in the present study are improvements over the samples of other studies investigating the same variables among older adults. Also, the sample was not made up of treatment-seeking individuals (e.g., presenting in a primary care or outpatient mental health clinic); thus, the study may have reached individuals who may not otherwise participate in psychological studies.
A second limitation is the exclusive use of self-report data. Although mailed questionnaires were intended to enhance the anonymity of data collection (over in-person data collection), it is possible that participants underreported thoughts of suicide, previous suicide attempts, or experiences of perceived burdensomeness and thwarted belongingness. Using a combination of self-report questionnaires, collateral reports, and behavioral measures might strengthen the conclusions of future studies. Strengths of the data include the examination of personality disorder features and several other relevant covariates in the analysis, as well as the investigation of the interaction between perceived burdensomeness and thwarted belongingness, all of which are an improvement ! 82 over previous studies. A third limitation is that this study only investigated correlational relationships between variables, so causality between variables cannot be determined definitively. Longitudinal studies tracking perceived burdensomeness and thwarted belongingness over time or intervention studies targeting these perceptions would help to determine causality.
Research Implications and Future Research
As mentioned previously, the results of the exploratory PCA raised the question of whether the INQ items related to social disconnection fit best on the Burden or
Belonging subscale. In the present study, those items clearly fell on to the Burden scale, but other versions of the INQ include them on the Belonging subscale (Joiner et al., 2009;
Van Orden, 2009; K. A. Van Orden, personal communication, March 2, 2011; Van
Orden, Witte, Gordon et al., 2008). Unfortunately, neither the validity correlations nor the
CFA in the current study provided a clear answer to this question. Recently, the joint presence of perceived burdensomeness and thwarted belongingness have been conceptualized as social disconnection (Van Orden & Conwell, 2011), further obscuring the answer. Future research should seek to clarify the relationship between Burden,
Belonging, and social disconnection, specifically whether burden and belonging are distinct variables that should be measured separately, or if they are actually subcomponents of another variable altogether. Perhaps one solution might be to use the
INQ as if it were unidimensional, given that it is the joint presence of burden and belonging that is of most interest in the IPTS. Longitudinal research that examines how perceived burdensomeness and thwarted belongingness each change over time as suicide ideation fluctuates (e.g., for someone who begins psychotherapy) might be useful to ! 83 further clarify these concepts. Longitudinal research could also help to further establish reliability of the INQ, such as test-retest reliability, and other types of validity of the INQ, such as predictive validity, which could be helpful in suicide prevention efforts.
Given that few studies have investigated the relationship between suicide ideation and perceived burdensomeness among non-terminally ill older adults (with the exception of Cukrowicz et al., 2011), more research in this area is definitely needed. Research regarding protective factors and the IPTS among older adults might be useful in order to better understand perceived burdensomeness. For example, for those older adults who are currently receiving care from family members or friends, but who do not feel like a burden to others, why is this so? How are they coping, in a way that suicidal individuals may not be? Is there a certain “cut off” point at which someone feels so burdensome that suicide becomes an attractive option? Future research may also examine the quality of the relationship between older adults who receive care and their caregivers; perhaps, levels of burdensomeness may vary with the quality of a relationship. The interaction between perceived burdensomeness and thwarted belongingness is also poorly understood at present. For instance, how do different levels of Burden and Belonging relate to each other? If one is relatively high and the other is relatively low, would suicide ideation still develop? Finally, the IPTS suggests that active suicide desire will only develop if individuals believe their levels of thwarted belongingness and perceived burdensomeness are unchangeable (Van Orden et al., 2010). The question remains to be answered regarding how hopeless must one feel about his perceived interpersonal status for the desire for suicide to develop. ! 84
Future research should also evaluate acquired capability for suicide among older adults. This IPTS variable, not examined in the current study, refers to the capability to carry out the act of suicide, which includes habituation to pain and sense of fearlessness about death that is learned over time (Van Orden et al., 2010). In other words, individuals who have experienced “painful and provocative experiences” (Van Orden et al., 2010, p.
587), such as repeated suicide attempts, non-suicidal self-injury, physical abuse, or combat trauma, develop a sense of numbness to pain with repeated exposure. According to the IPTS, this habituation, coupled with the desire for suicide, is what ultimately leads to suicide completion. Gordon et al. (2010) recently found support for the reinforcing properties of repeated deliberate self-harm (also known as non-suicidal self-injury), as hypothesized by the IPTS. Among young and middle aged individuals recruited from an outpatient clinic (N = 39, M age = 23.87 years, SD = 7.14, age range = 18 – 43) and a university-affiliated mood disorders research lab (N = 67, M age = 19.22 years, SD =
3.01, no age range reported), those with a higher number of deliberate self-harm episodes reported feeling less afraid and distressed, as well as more calm and attentive, after their most recent episode (Gordon et al., 2010). These findings point to the habituation to pain that develops with repeated self-injury. Another study found that individuals who had attempted suicide (N = 15, M age = 28.60 years, SD = 11.98, no age range reported) reported higher levels of fearlessness and pain insensitivity and more painful and provocative life events than suicide ideators (N = 15, M age = 33.07, SD = 14.03) or controls (N = 14, M age = 20.36, SD = 2.17; Smith, Cukrowicz, Poindexter, Hobson &
Cohen, 2010). Interestingly, no group differences were found for physiological data (i.e.,
EMG reactivity to suicide-related photos), indicating that perceptions of fear and pain ! 85 sensitivity may be more related to suicide attempts than physiological arousal (Smith et al., 2010). To date, no published studies have investigated acquired capability among older adults in the context of the IPTS. Future research should also examine the psychometric properties of the Acquired Capability for Suicide Scale (Joiner, Van Orden,
Witte, & Rudd, 2009) among samples of older adults hypothesized to have elevated levels of capability for suicide (e.g. combat veterans, individuals with recent suicide attempts).
Clinical Implications and Future Research
One of the goals of the IPTS is to determine more sensitive and specific predictors of completed suicide, in an effort to improve the assessment and prevention of suicidal behaviors (Joiner, 2005). Joiner, Van Orden, Witte, and Rudd (2009) proposed that perceived burdensomeness and thwarted belongingness should be targets in treatment for suicidal individuals for two reasons: (1) they are states, not traits, and are amenable to change and (2) according to the IPTS, suicide will not occur in the absence of desire for suicide (i.e., the joint presence of perceived burdensomeness and thwarted belongingness), regardless of whether or not an individual has acquired the capability for suicide. A recent article on the IPTS and late life suicide pointed out that many interventions designed to treat suicidal older adults already focus on increasing connectedness (Van Orden & Conwell, 2011). For instance, Heisel, Duberstein, Talbot,
King, and Tu (2009) found that a 16-week course of modified Interpersonal
Psychotherapy was effective in reducing suicide ideation, death ideation, and depression in a small sample of suicidal older adults (N = 11, M age = 69.4 years, SD = 4.9, age range = 60 – 78). They posited that interpersonal difficulties have a bidirectional ! 86 relationship with suicide risk, such that each exacerbates the other (Heisel et al., 2009).
One of the main interventions during the middle phase of treatment was to enhance sources of positive support and reduce sources of negative support. They also noted that participants in the trial perceived a strong relationship with their therapist (Heisel et al.,
2009). These findings, along with the findings from the current study, are consistent with the idea that thwarted belongingness is a fruitful target for change in psychotherapy with suicidal older adults.
Evidence suggests that perceived burdensomeness would be an appropriate target for treatment of suicidal older adults as well. For example, a recent study of 106 older adults (M age = 70.9 years, SD = 7.6, no age range reported) recruited from a university- affiliated primary care clinic found that perceived burdensomeness mediated the relationship between depressive symptoms and suicide ideation (Jahn, Cukrowicz,
Linton, Prabhu, 2011). This study is one of the first to be published that specifically examined the relationship between perceived burdensomeness and suicide ideation among a community-dwelling sample of older adults and suggests that perceptions of burden on others may be a useful point of intervention with suicidal older adults. One limitation of the study was their use of a two-item scale to measure Burden. One further limitation was the use of only hopelessness and gender as covariates in their regression analysis. Nevertheless, the results from Jahn et al. (2011) and the current study indicate that future research should examine the effectiveness of interventions that target cognitions related to perceptions of burdensomeness among suicidal older adults.
The clinical utility of the INQ should also be investigated among older adults in order to improve the assessment of suicide risk. Bryan (2010) recently conducted a study ! 87 with 219 deployed military personnel (M age 27.75 years, SD = 7.71, no age range reported) using a 10-item version of the INQ. Results indicated perceived burdensomeness and thwarted belongingness were related, but distinct components (r =
.53) and were correlated with suicide ideation, global mental health, insomnia, and
PTSD. More importantly, a score of 1 for perceived burdensomeness and a score of 3.2 for thwarted belongingness on the 10-item INQ were the most useful cutoff scores for detection of current suicide ideation among military personnel (Bryan, 2010). These findings suggest that the INQ can improve clinicians’ ability to rule-in and rule-out current suicide ideation among clients. More specifically, the presence of any perceived burdensomeness may signify increased suicide risk. This could be particularly valuable for screening for suicide risk among older adults, considering late-life suicide is associated with less expression of suicide ideation (Fiske & Arbore, 2000 – 2001). The current study found significant relationships between Burden and Belonging and other known risk factors for late-life suicide (e.g., depression, hopelessness), providing further support for the usefulness of assessing the IPTS variables in at-risk older adults.
Conclusion
The results from the present study add to the current, albeit limited, body of knowledge regarding the relationships between perceived burdensomeness, thwarted belongingness, and suicidality in older adults. Given the elevated risk for suicide among older adults and the increasing average age of individuals in the United States, the need for a better understanding of this phenomenon is critical. Since beliefs about thwarted belongingness and perceived burdensomeness are potentially modifiable, the IPTS may eventually be used to improve the assessment and treatment of older adults at risk for ! 88 suicide. Enhanced knowledge of the connection between emotional well-being and attitudes towards relationships among older adults, including perceptions about closeness to others and the degree to which they are a burden are on others, is a worthwhile pursuit in the effort to prevent late-life suicide.
REFERENCES
Akechi, T., Okuyama, T., Sugawara, Y., Nakano, T., Shima, Y., & Uchitomi, Y. (2004). Suicidality in terminally ill Japanese patients with cancer: Prevalence, patient perceptions, contributing factors, and longitudinal changes. Cancer, 100(1), 183– 191. doi: 10.1002/cncr.11890
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.
Anestis, M. D., Bryan, C. J., Cornette, M. M., & Joiner, T. E. (2009). Understanding suicidal behavior in the military: An evaluation of Joiner’s interpersonal- psychological theory of suicidal behavior in two case studies of active duty post- deployers. Journal of Mental Health Counseling, 31(1), 60–75.
Bailey, M., & McLaren, S. (2005). Physical activity alone and with others as predictors of sense of belonging and mental health in retirees. Aging and Mental Health, 9(1), 82–90. doi: 10.1080/13607860512331334031
Beck, A. T., Weissman, A., Lester, D., & Trexler, L. (1974). The measurement of pessimism: The Hopelessness Scale. Journal of Consulting and Clinical Psychology, 42(6), 861–865. doi:10.1037/h0037562
Bowling, A. (1997). Measuring health: A review of quality of life measurement scales. (2nd ed.). Philadelphia, PA: Open University Press.
Borges, G., Angst, J., Nock, M. K., Ruscio, A. M., Walters, E. E., & Kessler, R. (2006). A risk index for 12-month suicide attempts in the National Comorbidity Survey Replication (NCS-R). Psychological Medicine, 36, 1747–1757. doi:10.1017/S0033291706008786
Brenner, L. A., Gutierrez, P. M., Cornette, M. M., Betthauser, L. M., Bahraini, N., & Staves, P. J. (2008). A qualitative study of potential suicide risk factors in returning combat veterans. Journal of Mental Health Counseling, 30(3), 211–225.
Bryan, C. J. (2010). The clinical utility of a brief measure of perceived burdensomeness and thwarted belongingness for the detection of suicidal military personnel. Journal of Clinical Psychology, 66, 1–12. doi:10.1002/jclp.20726 ! 90
Bryan, C. J., Morrow, C. E., Anestis, M. D., & Joiner, T. E. (2010). A preliminary test of the interpersonal-psychological theory of suicidal behavior in a military sample. Personality and Individual Differences, 48, 347–350. doi: 10.1016/j.paid.2009.10.023
Cahill, E., Lewis, L. M., Barg, F. K., & Bogner, H. R. (2009). You don’t want to burden them: Older adults’ views on family involvement in care. Journal of Family Nursing, 15(3), 295–317. doi: 10.1177/1074840709337247
Centers for Disease Control and Prevention, National Centers for Injury Prevention and Control. (2010). Web-based Injury Statistics Query and Reporting System (WISQARS) Injury Mortality Report, 1999 – 2006. Retrieved from http://www.cdc.gov/injury/wisqars/index.html
Chatard, A., Selimbegovic, L., N’Dri Konan, P. (2009). Self-esteem and suicide rates in 55 nations. European Journal of Personality, 23, 19–32. doi: 10.1002/per.701
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159. doi: 10.1037/0033-2909.112.1.155
Conner, K. R., Britton, P. C., Sworts, L. M., & Joiner, T. E. (2007). Suicide attempts among individuals with opiate dependence: The critical role of belonging. Addictive Behaviors, 32, 1395–1404. doi: 10.1016/j.addbeh.2006.09.012
Conwell, Y., Duberstein, P. R., & Caine, E. D. (2002). Risk factors for suicide in later life. Biological Psychiatry, 52, 193–204. doi:10.1016/S0006-3223(02)01347-1
Coolidge, F. L., Segal, D. L., Cahill, B. S., & Simenson, J. T. (2010). Psychometric properties of a brief inventory for the screening of personality disorders: The SCATI. Psychology and Psychotherapy: Theory, Research, and Practice, 83, 395–405. doi:10.1348/147608310X486363
Coyle, N., & Sculco, L. (2004). Expressed desire for hastened death in seven patients living with advanced cancer: A phenomenologic inquiry. Oncology Nursing Forum, 31(1), 699–706. doi: 10.1188/04.ONF.699-709
Cukrowicz, K. C., Cheavens, J. S., Van Orden, K. A., Ragain, R. M., & Cook, R. L. (2011, March 14). Perceived burdensomeness and suicide ideation in older adults. Psychology and Aging. Advance online publication. doi: 10.1037/a0021836
Davidson, C. L., Wingate, L. R., Rasmussen, K. A., & Slish, M. L. (2009). Hope as a predictor of interpersonal suicide risk. Suicide and Life-Threatening Behavior, 39(5), 499–507. doi: 10.1521/suli.2009.39.5.499 ! 91
Duberstein, P. R., & Conwell, Y. (1997). Personality disorders and completed suicide: A methodological and conceptual review. Clinical Psychology: Science and Practice, 4(4), 359-376. doi:10.1111/j.1468-2850.1997.tb00127.x
Duberstein, P. R., Conwell, Y., Conner, K. R., Eberly, S., Evinger, J. S., & Caine, E. D. (2004). Poor social integration and suicide: Fact or artifact? Psychological Medicine, 34, 1331–1337. doi: 10.1017/S0033291704002600
Field, A. (2005). Discovering statistics using SPSS (2nd ed.). London: SAGE Publications.
Filiberti, A., Ripamonti, C., Totis, A., Ventafridda, V., De Conno, F., Contiero, P., & Tamburini, M. (2001). Characteristics of terminal cancer patients who committed suicide during a home palliative care program. Journal of Pain and Symptom Management, 22(1), 544–553. doi:10.1016/S0885-3924(01)00295-0
Fiske, A., & Arbore, P. (2000 – 2001). Future directions in late life suicide prevention. Omega, 42, 37–53.
Foster, T. (2003). Suicide note themes and suicide prevention. International Journal of Psychiatry in Medicine, 33(4), 323–331. doi: 10.2190/T210-E2V5-A5M0-QLJU
Garza, M. J., & Pettit, J. W. (2010). Perceived burdensomeness, familism, and suicidal ideation among Mexican women: Enhancing understanding of risk and protective factors. Suicide and Life-Threatening Behavior, 40(6), 561–573. doi:10.1521/suli.2010.40.6.561
Gordon, K. H., Selby, E. A., Anestis, M. D., Bender, T. W., Witte, T. K., Braithwaite, S., … Joiner, T. E. (2010). The reinforcing properties of repeated deliberate self- harm. Archives of Suicide Research, 14(4), 329–341. doi: 10.1080/13811118.2010.524059 Greene, S. M. (1981). Levels of measured hopelessness in the general population. British Journal of Clinical Psychology, 20, 11–14.
Hagerty, B. M., Lynch-Sauer, J., Patusky, K. L., Bouwsema, M., & Collier, P. (1992). Sense of belonging: A vital mental health concept. Archives of Psychiatric Nursing, 6(3), 172–177. doi:10.1016/0883-9417(92)90028-H
Hagerty, B. M., & Patusky, K. (1995). Developing a measure of sense of belonging. Nursing Research, 44(1), 9–13. doi:10.1097/00006199-199501000-00003
Harrison, K. E., Dombrovski, A. Y., Morse, J. Q., Houck, P., Schlernitzauer, M., Reynolds, C. F., III, & Szanto, K. (2010). Alone? Perceived social support and chronic interpersonal difficulties in suicidal elders. International Psychogeriatrics, 22(3), 445–454. doi: 10.1017/S1041610209991463 ! 92
Heisel, M. J. (2006). Suicide and its prevention among older adults. Canadian Journal of Psychiatry, 51, 143–154.
Heisel, M. J., & Duberstein, P. R. (2005). Suicide prevention in older adults. Clinical Psychology: Science and Practice, 12(3), 242–259. doi:10.1093/clipsy/bpi030
Heisel, M. J., Duberstein, P. R., Talbot, N. L., King, D. A., & Tu, X. M. (2009). Adapting interpersonal psychotherapy for older adults at risk for suicide: Preliminary findings. Professional Psychology: Research and Practice, 40(2), 156–164. doi: 10.1037/a0014731
Heisel, M. J., & Flett, G. L. (2005). A psychometric analysis of the Geriatric Hopelessness Scale (GHS): Towards improving assessment of the construct. Journal of Affective Disorders, 87, 211–220. doi: 10.1016/j.jad.2005.03.016
Heisel, M. J., & Flett, G. L. (2006). The development and initial validation of the Geriatric Suicide Ideation Scale. American Journal of Geriatric Psychiatry, 14, 742–751. doi:10.1097/01.JGP.0000218699.27899.f9
Hill, R. D., Gallagher, D., Thompson, L. W., & Ishida, T. (1988). Hopelessness as a measure of suicidal intent in the depressed elderly. Psychology and Aging, 3(3), 230–232. doi:10.1037/0882-7974.3.3.230
Hughes, M. E., Waite, L., J., Hawkley, L. C., & Cacioppo, J. T. (2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Research on Aging, 26(6), 655–672. doi: 10.1177/0164027504268574
Jahn, D. R., Cukrowicz, K. C., Linton, K., & Prabhu, F. (2011). The mediating effect of perceived burdensomeness on the relation between depressive symptoms and suicide ideation in a community sample of older adults. Aging and Mental Health, 15(2), 214–220. doi: 10.1080/13607863.2010.501064
June, A., Segal, D. L., Coolidge, F. L., & Klebe, K. (2009). Religiousness, social support and reasons for living in African American and European American older adults: An exploratory study. Aging and Mental Health, 13, 753-760. doi:10.1080/13607860902918215
Joiner, T. E. (2005). Why people die by suicide. Cambridge, MA: Harvard University Press.
Joiner, T. E., Hollar, D., & Van Orden, K. (2006). On Buckeyes, Gators, Super Bowl Sunday, and the Miracle on Ice: “Pulling together” is associated with lower suicide rates. Journal of Social and Clinical Psychology, 25(2), 179–195. doi:10.1521/jscp.2006.25.2.179
! 93
Joiner, T. E., Pettit, J. W., Walker, R. L., Voelz, Z. R., Cruz, J., Rudd, M. D., & Lester, D. (2002). Perceived burdensomeness and suicidality: Two studies on the suicide notes of those attempting and those completing suicide. Journal of Social and Clinical Psychology, 21(5), 531–545. doi:10.1521/jscp.21.5.531.22624
Joiner, T. E., Van Orden, K. A., Witte, T. K., Rudd, M. D. (2009). The interpersonal theory of suicide: Guidance for working with suicidal clients. Washington, DC: American Psychological Association.
Joiner, T. E., Van Orden, K. A., Witte, T. K., Selby, E. A., Ribeiro, J. D., Lewis, R., & Rudd, M. D. (2009). Main predictions of the interpersonal-psychological theory of suicidal behavior: Empirical tests in two samples of young adults. Journal of Abnormal Psychology, 118(3), 634–646. doi:10.1037/a0016500
Kelly, B., Burnett, P., Pelusi, D., Badger, S. Varghese, F., & Robertson, M. (2002). Terminally ill cancer patients’ wish to hasten death. Palliative Medicine, 16, 339– 345. doi:10.1191/0269216302pm538oa
Kessler, R. C., Berglund, P., Borges, G., Nock, M., & Wang, P. S. (2005). Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990-1992 to 2001-2003. Journal of the American Medical Association, 293(20), 2487–2495. doi:10.1001/jama.293.20.2487
Kissane, M., & McLaren, S. (2006). Sense of belonging as a predictor of reasons for living in older adults. Death Studies, 30, 243–258. doi:10.1080/07481180500493401
Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16, 606–613. doi:10.1046/j.1525-1497.2001.016009606.x
Linehan, M. M., & Laffaw, J. A. (1982). Suicidal behaviors among clients of an outpatient clinic versus the general population. Suicide and Life-Threatening Behavior, 12, 234–239.
Luoma, J. B., Martin, C. E., & Pearson, J. L. (2002). Contact with mental health and primary care providers before suicide: A review of the evidence. American Journal of Psychiatry, 159, 909–916. doi:10.1176/appi.ajp.159.6.909
Maris, R. W., Berman, A. L., & Silverman, M. M. (Eds.) (2000). Comprehensive textbook of suicidology. New York, NY: The Guilford Press.
Martin, A., Rief, W., Klaiberg, A., & Braehler, E. (2006). Validity of the Brief Patient Health Questionnaire Mood Scale (PHQ-9) in the general population. General Hospital Psychiatry, 28, 71–77. doi:10.1016/j.genhosppsych.2005.07.003 ! 94
Marty, M. A., Segal, D. L., & Coolidge, F. L. (2010). Relationships among dispositional coping strategies, suicidal ideation, and protective factors against suicide in older adults. Aging and Mental Health, 14(8), 1015–1023.
McHorney, C. A. (1996). Measuring and monitoring general health status in elderly persons: Practical and methodological issues in using the SF-36 Health Survey. The Gerontologist, 36(5), 571–583.
McLaren, S., & Challis, C. (2009). Resilience among men farmers: The protective roles of social support and sense of belonging in the depression-suicidal ideation relation. Death Studies, 33, 262–276.
McLaren, S., Gomez, R., Bailey, M., & Van Der Horst, R. K. (2007). The association of depression and sense of belonging with suicidal ideation among older adults: Applicability of resiliency models. Suicide and Life-Threatening Behavior, 37(1), 89–102. doi:10.1521/suli.2007.37.1.89
McPherson, C., Wilson, K. G., & Murray, M. A. (2007a). Feeling like a burden: Exploring the perspectives of patients at the end of life. Social Science and Medicine, 64, 417–427. doi:10.1016/j.socscimed.2006.09.013
McPherson, C., Wilson, K. G., & Murray, M. A. (2007b). Feeling like a burden to others: A systematic review focusing on the end of life. Palliative Medicine, 21, 115– 128. doi:10.1177/0269216307076345
Mireault, M., & de Man, A. F. (1996). Suicidal ideation among the elderly: Personal variables, stress, and social support. Social Behavior and Personality, 24(4), 385– 392. doi:10.2224/sbp.1996.24.4.385
Miret, M., Nuevo, R., Morant, C., Sainz-Corton, E., Jimenez-Arriero, M. A., Lopez-Ibor, J. J., … Ayuso-Mateos, J. L. (2010). Differences between younger and older adults in the structure of suicidal intent and its correlates. American Journal of Geriatric Psychiatry, 18(9), 839–847. doi:10.1097/JGP.0b013e3181d145b0
Nademin, E., Jobes, D. A., Pflanz, S. E., Jacoby, A. M., Ghahramanlou-Holloway, M., Campise, R., … Johnson, L. (2008). An investigation of interpersonal- psychological variables in Air Force suicides: A controlled-comparison study. Archives of Suicide Research, 12, 309–326. doi:10.1080/13811110802324847
Neufeld, E., O’Rourke, N., & Donnelly, M. (2010). Enhanced measurement sensitivity of hopeless ideation among older adults at risk of self-harm: Reliability and validity of Likert-type responses to the Beck Hopelessness Scale. Aging and Mental Health, 14(6), 752–756.
! 95
Nock, M. K., & Favazza, A. R. (2009). Nonsuicidal self-injury: Definition and classification. In M. K. Nock (Ed.), Understanding nonsuicidal self-injury: Origins, assessment, and treatment (pp. 9–18). Washington, DC: American Psychological Association.
Pinquart, M., & Sorensen, S. (2003). Differences between caregivers and noncaregivers in psychological health and physical health: A meta-analysis. Psychology and Aging, 18(2), 250–267. doi:10.1037/0882-7974.18.2.250
Roberts, J. K. (1999, January). Basic concepts of confirmatory factor analysis. Paper presented at the annual meeting of the Southwest Educational Research Association, San Antonio, TX.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self- esteem: Construct validation of a single-item measure of the Rosenberg Self- Esteem Scale. Personality and Social Psychology Bulletin, 27, 151–161. doi: 10.1177/0146167201272002
Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J. (2002). Global self-esteem across the life span. Psychology and Aging, 17(3), 423–434. doi: 10.1037//0882-7974.17.3.423
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Segal, D. L., Marty, M. A., Meyer, W. J., & Coolidge, F. L. (2010). Personality traits, suicidal ideation, and resilience to suicide among older adults. Manuscript submitted for publication.
Smith, P. N., Cukrowicz, K. C., Poindexter, E. K., Hobson, V., & Cohen, L. M. (2010). The acquired capability for suicide: A comparison of suicide attempters, suicide ideators, and non-suicidal controls. Depression and Anxiety, 27, 871–877. doi: 10.1002/da20701
Thompson, B. (2004). Exploratory and confirmatory factor analysis: Understanding concepts and applications. Washington, DC: American Psychological Association.
Trout, D. L. (1980). The role of social isolation in suicide. Suicide and Life Threatening Behavior, 10(1), 10–23.
You, S., Van Orden, K. A., & Conner, K. R. (2011). Social connections and suicidal thoughts and behavior. Psychology of Addictive Behaviors, 25(1), 180–184. doi: 10.1037/a0020936 ! 96
Van Der Horst, R. K., & McLaren, S. (2005). Social relationships as predictors of depression and suicidal ideation in older adults. Aging and Mental Health, 9(6), 517–525. doi:10.1080/13607860500193062
Van Orden, K. A. (2009). Construct validity of the Interpersonal Needs Questionnaire (Doctoral dissertation, Florida State University). Retrieved from http://www.psy.fsu.edu/~joinerlab/
Van Orden, K. A., & Conwell, Y. (2011). Suicides in late life. Current Psychiatry Reports, 13(3), 234–241. doi: 10.1007/s11920-011-0193-3
Van Orden, K. A., Lynam, M. E., Hollar, D., & Joiner, T. E. (2006). Perceived burdensomeness as an indicator of suicidal symptoms. Cognitive Therapy & Research, 30, 457–467. doi:10.1007/s10608-006-9057-2
Van Orden, K. A., Merrill, K. A., & Joiner, T. E. (2005). Interpersonal-psychological precursors to suicidal behavior: A theory of attempted and completed suicide. Current Psychiatry Reviews, 1, 187–196. doi:10.2174/1573400054065541
Van Orden, K. A., Witte, T. K., Cukrowicz, K. C., Braithwaite, S. R., Selby, E. A., & Joiner, T. E. (2010). The interpersonal theory of suicide. Psychological Review, 117(2), 575–600. doi:10.1037/a0018697
Van Orden, K. A., Witte, T. K., Gordon, K. H., Bender, T. W., & Joiner, T. E. (2008). Suicidal desire and the capability for suicide: Tests of the interpersonal- psychological theory of suicidal behavior among adults. Journal of Consulting and Clinical Psychology, 76(1), 72–83. doi:10.1037/0022-006X.76.1.72
Van Orden, K., A., Witte, T., K., James, L. M., Castro, Y., Gordon, K. H., Braithwaite, S. R., … Joiner, T. E. (2008). Suicidal ideation in college students varies across semesters: The mediating role of belongingness. Suicide and Life Threatening Behavior, 38(4), 427–435. doi:10.1521/suli.2008.38.4.427
Waern, M., Rubenowitz, E., & Wilhelmson, K. (2003). Predictors of suicide in the old elderly. Gerontology, 49, 328–334. doi:10.1159/000071715
Ware, J. E., & Sherbourne, C. D. (1992). The MOS 36-item Short-Form Health Survey (SF-36): Conceptual framework and item selection. Medical Care, 30(6), 473– 483. doi:10.1097/00005650-199206000-00002
Wilson, K. G., Curran, D., & McPherson, C. J. (2005). A burden to others: A common source of distress for the terminally ill. Cognitive Behavior Therapy, 34(2), 115– 123. doi:10.1080/16506070510008461
! 97
Wilson, K. G., Scott, J. F., Graham, I. D., Kozak, J. F., Chater, S., Viola, R. A., …Curran, D., (2000). Attitudes of terminally ill patients toward euthanasia and physician-assisted suicide. Archives of Internal Medicine, 160, 2454–2460. doi:10.1001/archinte.160.16.2454 98
APPENDIX A
INSTITUTIONAL REVIEW BOARD APPROVAL
University of Colorado at Colorado Springs
Institutional Review Board ______1420 Austin Bluffs Parkway P.O. Box 7150 Colorado Springs, CO 80933-7150 (719) 255-4150; [email protected]
Memorandum
TO: Meghan Marty, M.A., Ed.M.
FROM: Sandy K. Wurtele, Ph.D. IRB Chair
DATE: September 24, 2010
RE: Pikes Peak Region Late Life Wellness Study (IRB #10-159) Expedited UCCS Speedtype (-) Proposal Number (-)
Thank you for submitting your Request for IRB Review. Your protocol has been approved for one year, with an expiration date of 9-24-11. NOTE: Please include the IRB number and the expiration date on your informed consent document.
Once human subject research has been approved, it is the Principal Investigator’s (PI) responsibility to report changes in research activity related to the project. The PI must provide the IRB with all protocol and consent form amendments and revisions. IRB must approve these changes prior to their implementation. All advertisements recruiting study subjects must also receive prior approval by the IRB. The PI must promptly inform the IRB of all adverse and serious adverse events to subjects. If the project is to continue beyond the expiration date noted above, the PI must submit a Request for Renewal prior to that date. Failure to comply with these federally mandated responsibilities may result in suspension or termination of the project.
Thank you for your concern about human subject protection issues, and good luck with your research.
Cc: Office of Sponsored Research Dan Segal, Ph.D.
99
APPENDIX B
INFORMED CONSENT FORM
IRB Approval #10-159, Exp. 9/24/11
Informed Consent Form Pikes Peak Region Late Life Wellness Study
You are invited to participate in a research project conducted by Meghan Marty, graduate student in the University of Colorado's Department of Psychology.
You will be asked to complete a series of questionnaires that will ask you questions regarding your beliefs about yourself and your relationships with others. We will also ask you some basic demographic questions such as your age, gender, and racial background. Completing the survey should take approximately 30 minutes of your time.
The potential risks associated with this study are minor. You may experience some emotional discomfort due to the personal nature of the survey’s questions. We expect the project to benefit you by encouraging you to think of your personal values and how you relate to others.
If you have decided to participate in this project, please understand that your participation is voluntary and that you have the right to withdraw your consent or discontinue participation at any time with no penalty. You also have the right to refuse to answer any question(s) for any reason with no penalty.
In addition, your individual privacy will be maintained in all publications or presentations resulting from this study. Only code numbers will be used to indentify forms and data. No individual scores will be given out after you complete the questionnaires. Any publications or presentations that come from this study will report group averages only and no identifying data of individual participants will be reported. No persons other than qualified members of the research team will have access to your anonymous responses. All data will be stored in a locked cabinet.
If you have any questions regarding this project, you may contact Meghan Marty by telephone at 719-357-6347, by e-mail at [email protected], or at 1420 Austin Bluffs Pkwy, Colorado Springs, CO 80918. You may also contact Dr. Daniel Segal, the faculty advisor for this project, at 719-255-4176 or by e-mail at [email protected].
If you have questions regarding your rights as a research participant or any concerns regarding this project, you may report them -- confidentially, if you wish -- to Dr. Sandy Wurtele, the UCCS Chairperson of the Institutional Review Board at 719-255-4150.
By completing and returning the questionnaire packet, you are consenting to participate in this study. Please keep this page for your records.