Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts Nicholas P. Allan n, Aaron M. Norr, Jay W. Boffa, Daphne Durmaz, Amanda M. Raines, Norman B. Schmidt n Department of Psychology, Florida State University, 1107 W. Call St., Tallahassee, FL 32306-4301, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 31 January 2015 Received in revised form 12 May 2015 Accepted 29 May 2015
Anxiety sensitivity (AS) has recently been linked to suicidality. Specifically, AS cognitive concerns has been implicated as a risk factor, and AS physical concerns as a protective factor, for suicidal ideation and suicide attempts. However, no studies have used structural equation modeling (SEM) to address issues of skewed suicide variables and bifactor modeling of AS to address the high degree of overlap between the lower-order dimensions of AS that limit interpretation of these past findings. AS, suicidal ideation, past suicide attempts, and depression were assessed in a clinical sample of 267 individuals (M age¼ 35.45 years, SD ¼16.53; 52.1% female). The global AS and AS cognitive concerns factors were positively, significantly associated with suicidal ideation, though these effects were nonsignificant controlling for depression. The global AS factor was positively, significantly associated with suicide attempts, controlling for depression. The current study demonstrated that the relations between AS and suicidal ideation are not maintained when accounting for depression, suggesting that the relation between AS and suicidal ideation may be mediated by depression. The positive relation between global AS and suicide attempts is consistent with theories positing suicide attempts as a consequence of an inability to cope with intolerable distress. & 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Suicidal ideation Suicide attempts Anxiety sensitivity Bifactor modeling Depression
1. Introduction According to the most recent report by the World Health Organization, suicide accounts for over one million deaths a year, and is among the top three leading causes of death among individuals 15–44 years old (World Health Organization, 2012), making it a critical public health issue. Existing research indicates that completed suicides are strongly predicted by prior suicide attempts as well as suicidal ideation ( Kessler et al., 1999; Souminen et al., 2004; Weissman et al., 1989). Therefore, it is important to identify and understand potential malleable risk factors (Kraemer et al., 1997) for suicidal behavior to better predict and reduce completed suicides. One risk factor that has been linked to suicide attempts and suicidal ideation is anxiety sensitivity (AS), or the fear of anxiety-related sensations (Reiss et al., 1986). Anxiety sensitivity (AS) is associated with and prospectively predicts the development of both anxiety and mood pathology (Schmidt et al., 2006; Olatunji and Wolitzky-Taylor, 2009), and appears to be malleable through intervention (Schmidt et al., n
Corresponding authors. E-mail addresses: [email protected] (N.P. Allan), [email protected] (N.B. Schmidt).
2007). AS reflects an individual's tendency to respond fearfully to symptoms of anxiety arising from the belief that these symptoms will have negative consequences (Reiss and McNally, 1985). Individuals with elevated AS are likely to interpret benign symptoms of anxiety as potentially harmful or dangerous. AS is composed of three lower-order dimensions: cognitive, physical, and social concerns. AS cognitive concerns reflect the tendency to respond anxiously to feelings of cognitive dyscontrol, such as experiencing racing thoughts. AS physical concerns reflect the tendency to respond anxiously to physiological symptoms of anxiety, such as shortness of breath. AS social concerns reflect the tendency to fear potential negative evaluations resulting from others noticing symptoms of anxiety, such as sweating. In addition to being a risk factor for anxiety and mood pathology broadly, AS also has a robust relationship with suicidality (Schmidt et al., 2001). There is some evidence to suggest that much of this association is driven primarily by the AS cognitive concerns component. Capron and colleagues (Capron et al., 2012b) found in a clinical outpatient sample that the AS cognitive concerns scale was significantly related to suicidal ideation and previous suicide attempts. This positive association between AS cognitive concerns and suicidal ideation was prospectively demonstrated in a large sample of cadets entering basic training at the United States Air Force Academy as well as in a sample of clinic
http://dx.doi.org/10.1016/j.psychres.2015.05.066 0165-1781/& 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
outpatient who screened positive for PTSD (Capron et al., 2012a). Furthermore, the association between AS cognitive concerns and suicidality remain significant after accounting for other relevant risk variables, such as distress tolerance, gender, insomnia, thwarted belongingness, and depressive symptoms (Capron et al., 2013; Oglesby et al., 2015). Although other AS subscales have been found to be related to suicidality, results have been inconsistent. For example, Capron et al. (2012b) found that AS social concerns was positively associated with previous suicide attempts in one study. In addition, negative relations between AS physical concerns and suicidal ideation have been found (Capron et al., 2012a, 2012b), indicating there may be a specific role for AS physical concerns as a protective factor for suicidality. Although relations between AS and suicide have been demonstrated across various samples, there are limitations to the prior studies. First, it is not entirely clear whether the relation between AS and suicidality can be fully accounted for by AS cognitive concerns or whether a global AS factor is also related. As suggested by Mohlman and Zinbarg (2000), AS appears to comprise a global factor in addition to the three lower-order dimensions. As previous work on AS and suicidality has relied on scale scores, it has been impossible to examine both overall AS and the specific dimensions in the same model. One approach that can clarify the true relation between global AS, specific AS dimensions, and suicidality is bifactor modeling (Chen et al., 2012; Reise, 2012). In this approach, common variance, representing AS more generally, can be distinguished from variance specific to the lowerorder factors (i.e., cognitive, physical, and social concerns). In fact, several authors have used this approach to validate and implement a bifactor model of AS (Mohlman and Zinbarg, 2000; Osman et al., 2010; Ebesutani et al., 2013; Allan et al., 2015). Thus, a bifactor model of AS will allow for the investigation of both the global AS factor and specific AS dimensions simultaneously in the prediction of suicidality. Only a single study has examined the relations between the AS factors, using bifactor modeling, and suicidality. Osman et al. (2010) found that suicidality was associated with a global AS factor as well as the specific AS physical and cognitive concerns factors. However, these results should be interpreted with caution as their study was conducted in an undergraduate sample with low reported rates of suicidality (i.e., Suicide Behavior Questionnaire-Revised [SBQ-R; Osman et al., 2001] M¼ 4.43, SD ¼2.11; minimum possible total score on SBQ-R is 4.00). Related to this, a second limitation to be addressed in the previous AS and suicidality literature is the skewed nature of suicidal ideation. Given that suicidality is absent in most individuals, suicidality variables tend to have a significant positive skew and be heavily leptokurtic (Van Orden et al., 2008; Capron et al., 2012b). The issues associated with statistical analysis of significantly skewed data have been well documented (Delucchi and Bostrom, 2004; Grant, 2010). Most previous work on AS and suicidality has either not explicitly dealt with the skewed suicidality variables or conducted a logarithmic transformation (Capron et al., 2013), which does not remedy the issues associated with a significant portion of the sample containing the same value on the suicidality variable (i.e., 0, absence of suicidality; Delucchi and Bostrom, 2004). Modeling suicidal ideation treating the items as categorical indicators of a continuous latent variable will remedy issues associated with a highly skewed dependent scale, as this approach does not require normally distributed indicators (Bollen, 1989; Rhemtulla et al., 2012). 1.1. The current study In the current study, we advanced the literature on AS and suicidality using improved statistical methods in a clinical
outpatient sample. First, by modeling AS as a bifactor we were able to investigate the unique contributions of both the global and specific AS factors to suicidal ideation and past suicide attempts. Second, we improved upon the literature by modeling suicidal ideation as a continuous latent variable with categorical indicators to account for the significantly skewed nature of this variable. We hypothesized that using the bifactor model of AS, we would find a significant positive relation between AS cognitive concerns and suicidal ideation and past suicide attempts as has been found in previous studies (Capron et al., 2012b; Oglesby et al., 2015). Second, we hypothesized that there would be a significant negative relation between the global AS factor and suicidal ideation and past attempts. This hypothesis was based on past studies demonstrating a negative relation between AS physical concerns and suicidality (Capron et al., 2012a) as well as studies indicating that the global AS factor predominantly captures AS physical concerns (Allan et al., 2015). We hypothesized that AS physical and social concerns would not be associated with suicidal ideation or past attempts after accounting for the relations between AS cognitive concerns and the global AS factor. Finally, we hypothesized that the relations that global AS and AS cognitive concerns share with suicidal ideation and past suicide attempts would remain after accounting for depression.
2. Methods 2.1. Participants and setting The sample included 267 individuals presenting to an outpatient anxiety clinic to receive psychological services and/or participate in research options. Demographic and information regarding participant's current primary psychiatry diagnosis are provided in Table 1. Gender was fairly evenly distributed (52.1% female) with ages ranging from 18 to 88 (M¼35.45, SD¼ 16.53). The racial composition of the sample was as follows: 63.7% Caucasian, 23.6% African American, 1.5% Asian, 4% Native American, 4% Pacific Islander, and 10.4% other (e.g., biracial) with 11.6% identifying as Hispanic. The majority of the sample attended some college (53.6%), 16.9% obtained a 4-year degree, 12.0% had a high school diploma, 10.1% had a graduate degree, 4.1% trade school degree, and 3.3% had less than a high school education. Regarding primary psychiatric diagnoses, 45.6% of sample met for a primary anxiety disorder, 18.4% mood disorder, 10.9% trauma and stressorrelated disorder, 6.4% substance-related disorder, 2.7% obsessive– compulsive and related disorder, 2.7% other (e.g., somatic disorder) and 13.5% had no primary diagnosis. 2.2. Procedure All individuals included in the current study agreed to participate in Institutional Review Board-approved research being conducted at the clinic. Prior to initiating treatment or research options, all individuals were interviewed by an advanced level clinical psychology graduate student to assess for the presence of psychopathology. All diagnoses were confirmed at a weekly supervision meeting with a licensed clinical psychologist and director of the clinic. Data used in the current investigation were part of a battery of questionnaires given after completion of the diagnostic interview. Participants were excluded from the current study if they were unable to fully complete their diagnostic interview, including their self-report questionnaires. All other participants who were given a structured clinical interview were included in the current study.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
confidence interval [CI] ¼.76, .85), but low for the specific AS Physical Concerns (ωs ¼.32, 95% CI ¼.15, .46), AS Cognitive Concerns (ωs ¼.29, 95% CI ¼ .03,.44), and AS Social Concerns (ωs ¼ .30, 95% CI ¼ .05, .44) factors.1
Table 1 Demographic features and primary axis I diagnoses. Demographics
Age
Mean 35.45
SD 16.53
Gender Male Female
n 128 139
% 47.9 52.1
Race Caucasian African American Asian Native American Pacific Islander Other
170 63 4 1 1 28
63.7 23.6 1.5 0.4 0.4 10.4
Education Graduate degree 4-year degree Associates or some college High school degree Trade school degree Less than high school
27 45 143 32 11 9
10.1 16.9 53.6 12.0 4.1 3.4
Psychiatric diagnosis Anxiety Mood Trauma/stressor-related Alcohol/substance use OCD/related Other No diagnosis
122 49 29 17 7 7 36
3
45.6 18.4 10.9 6.4 2.7 2.7 13.5
Note. SD ¼standard deviation. Other includes biracial or wishing not to report. OCD¼ Obsessive–compulsive disorder.
2.3. Measures 2.3.1. Clinician administered 2.3.1.1. Structured clinical interview for DSM-IV and DSM-5 (SCID). All psychiatric diagnoses were determined using the SCID. Diagnostic interviews were administered by trained doctoral level therapists who completed extensive training in the administration and scoring of the SCID. This training included reviewing SCID training tapes, observing live administrations, and conducting practice interviews with other experienced individuals. During the training process all trainees received feedback until they demonstrated high levels of reliability. Rates of agreement between interviewers within our laboratory have been found to be over 80% with a kappa value of .77 (Timpano and Schmidt, 2012). 2.3.2. Self-report 2.3.2.1. Anxiety sensitivity index-3 (ASI-3). The ASI-3 is an 18-item self-report questionnaire assessing fears of anxious arousal (Taylor et al., 2007) adapted from the Anxiety Sensitivity Index (Reiss et al., 1986). Individuals were instructed to rate how much they agreed with each item on a 5-point Likert-type scale ranging from 0 (very little) to 4 (very much). In addition to an overall total score, the ASI-3 yields three subscales: physical, cognitive, and social concerns. Prior research has demonstrated that the ASI-3 is both a valid and reliable measure of AS (Taylor et al., 2007). In the current investigation, the ASI-3 total and aggregate subscale scores demonstrated good to excellent internal consistency (αs ¼.84 .93). Reliability for the bifactor solution (calculated using McDonald's omega hierarchical [ωh; ωs for the specific factors]; McDonald, 1999; Reise, 2012) was good for the global AS factor (ωh ¼.81, 95%
2.3.2.2. Beck depression inventory-II (BDI-II). The BDI-II is a 21-item self-report questionnaire assessing common symptoms of depression (Beck et al., 1996). In addition to a total score, the BDI-II contains cognitive and affective/somatic subscales (Beck et al., 1996). The BDI-II has been found to be psychometrically sound, with high internal consistency (Beck et al., 1996). In the present study, the cognitive and affective/somatic subscales demonstrated good internal consistency (αs ¼.87,.89; ωhs ¼.88, 95% CI [.86, .90], .88, 95% CI [.86, .90]). 2.3.2.3. Depression symptom inventory-suicide subscale (DSI-SS). The DSI-SS is 4-item self-report questionnaire assessing suicidal ideation during the past two weeks (Metalsky and Joiner, 1997). Items are rated using a 4-point Likert-type scale ranging from 0-3 with higher scores reflecting more severe symptoms. Prior research has demonstrated that the DSI-SS has sound psychometric properties (Joiner and Rudd, 1996). In the present investigation, internal consistency was good (α ¼.88; ωh ¼.91, 95% CI [.88, .93]). 2.3.2.4. Suicide history. Suicide history was assessed by a question on a form seeking medical history asking individuals to indicate prior attempts. 2.4. Data analytic plan AS models were first examined in Mplus version 7 (Muthén and Muthén, 1996 2012) treating items as continuous and using full information maximum likelihood with the Yuan Bentler scaled chi-square (χ2) value to account for missing data and nonnormality. The three-factor correlated-traits model was compared to the bifactor model using the corrected likelihood ratio test. Overall model fit was assessed using the χ2 statistic, for which a nonsignificant value indicates good fit. However, this statistic can be too restrictive, especially with many items per factor (Hu and Bentler, 1999; Mulaik, 2007; Moshagen, 2012). Model fit was therefore additionally examined using the comparative fit index (CFI), Tucker Lewis index (TLI), and root mean square error of approximation (RMSEA) with accompanying 90% confidence interval (CI). CFI and TLI values from above .90 to above .95 and RMSEA values less than .08 to less than .05 are indicative of acceptable to good fit. Further, an RMSEA lower-bound CI below .05 indicates that good fit cannot be ruled out and an upper-bound CI above .10 indicates that poor fit cannot be ruled out. The best-fitting model was then included in a structural equation model (SEM) examining the relations between the AS factors and a suicidal ideation factor (comprised of items from the DSI-SS, treated categorically) and past suicide attempts (0 ¼ not past attempts, 1 ¼one or more past attempts) treated as a manifest variable. Because the outcome variables were categorical, the default robust weighted least squares estimator (WLSMV) was used and model fit was assessed similar to the AS CFA models. Finally, to determine if there were unique relations between AS and the suicide outcome variables controlling for depression, the relations between the AS factors and suicidal ideation and past suicide attempts were examined in a model including a second-order depression factor (with first-order cognitive and affective/somatic 1 All ω's were calculated in Mplus using robust maximum likelihood, treating the data as continuous, and including 1000 bias-corrected bootstrap draw to provide asymmetric confidence intervals.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
significantly better (Δ χ2 ¼68.44, df ¼17, p o.001). All items loaded significantly on the global AS factor as well as on their specific AS cognitive, physical, and social concerns factors.
factors, with item-level data modeled categorically).
3. Results 3.3. Structural equation modeling of the relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts
3.1. Descriptive statistics and correlations Although all variables (with the exception of past suicide attempts) were modeled as latent variables, scale scores were computed to provide means and standard deviations as well as correlations between variables (see Table 2). There were significant correlations among all the variables with the exception of ASI-3 physical concerns and DSI-SS suicidal ideation. There was little missing data (between 0 and 6 participants missing data per item, 11 participants were missing information about past suicide attempt) and all participants were included in the analyses. Within the current sample, 33 participants (12.9%) reported one prior attempt, and 18 (7.1%) reported more than one prior attempt. For the current study, participants were dichotomized into no past attempts (coded as 0) or one or more past attempts (coded as 1). Mean DSI-SS scores were.80 (SD ¼1.61), providing evidence that suicidal ideation would be extremely skewed if it had been treated as continuous. 3.2. Confirmatory factor analysis of anxiety sensitivity CFAs of the three-factor and bifactor model of the ASI-3 were compared. The three-factor model provided acceptable fit to the data (χ2 ¼256.18, df ¼132, p o.05, CFI ¼.95, TLI¼.94, RMSEA ¼.06, 90% CI [.05,.07]) as did the bifactor model (χ2 ¼183.19, df ¼ 117, p o.05, CFI¼ .97, TLI ¼.96, RMSEA ¼ .05, 90% CI [.03,.06]). Comparison of the models indicated that the bifactor model fit Table 2 Descriptive statistics and correlations for ASI-3 subscales, depression subscales, suicidal ideation, and past suicide attempts. 1
2
3
4
5
6
7
8
1. ASI-3 cognitive
2. ASI-3 physical
.61n
3. ASI-3 social
.62n
.50n
4. ASI-3 total
.88n
.83n
.82n
5. BDI-2 cognitive
.56n
.31n
.48n
.52n
6. BDI-2 affective/somatic
.58n
.38n
.43n
.54n
.69n
7. BDI-2 total
.61n
.37n
.48n
.57n
.90n
.94n
8. DSI-SS suicidal ideation
.24n
.06
.17n
.16n
.48n
.34n
.44n
9. Past suicide attempts
.42n
.18n
.27n
.33n
.41n
.36n
.40n
.28n
Mean (% attempts) Standard deviation
9.47 9.39 12.70 31.32 9.29 13.57 22.84 .80 6.67 6.21 5.93
15.72 5.86 7.55
12.27
9
po .05.
3.4. Structural equation modeling of the relations between anxiety sensitivity and depression factors and suicidal ideation and past suicide attempts The SEM examining the relations the global and specific AS factors, and the second-order depression factor (comprised of first-order cognitive and affective/somatic factors) share with the suicidal ideation factor and past suicide attempts provided acceptable fit to the data (χ2 ¼ 244.14, df ¼204, po .05, CFI ¼.98, TLI¼.98, RMSEA ¼ .03, 90% CI [.01, .04]). Model parameters are provided in Table 2. In this model, there was a significant positive relation between the suicidal ideation and depression factors (β ¼ .58, p o.05). There were also significant positive relations between past suicide attempts and the global AS (B ¼.31, po .05) and depression factors (B ¼ .35, p o.01). Controlling for all other variables, a 1 SD increase in the global AS factor was associated with a 30% probability of a past suicide attempt and a 1 SD increase in the depression factor was associated with a 31% probability of a past suicide attempt. This model accounted for 71% of the variance in the suicidal ideation factor and 41% of the variance in past suicide attempts.
4. Discussion
20.0%
1.61
Note. ASI-3 ¼anxiety sensitivity index-3. BDI-2 ¼Beck depression inventory-2. DSISS ¼ depressive symptom inventory-suicidal scale. n
The SEM examining the relations past suicidal ideation shares with the global and specific AS factors as well as the suicidal ideation factor and past suicide attempts provided acceptable fit to the data (χ2 ¼ 244.14, df ¼ 204, p o.05, CFI¼.98, TLI¼.98, RMSEA¼ .03, 90% CI [.01,.04]). Standardized factor loadings for the AS and suicidal ideation factors as well as standardized model parameters (unstandardized probit regression parameters are reported for past suicide attempts) are provided in Fig. 1. There were significant positive relations between the suicidal ideation factor and the global AS (β ¼.22, p o.05) and AS cognitive concerns factors (β ¼ .26, po .05). There were also significant positive relations between past suicide attempts and the slobal AS factor (B ¼.50, po .001), such that, controlling for the specific AS factors, a standard deviation (SD) increase in the global AS factor was associated with a 36% probability of having a past suicide attempt. This model accounted for 13% of the variance in the suicidal Ideation factor and 33% of the variance in past suicide attempts. Results were also conducted including gender as a covariate and the relations between the AS factors and suicidal ideation and past suicide attempts were unchanged. Further, gender was not significantly associated with either outcome variable. Therefore, only the model excluding gender was reported.
In the current study, global AS and AS cognitive concerns were only related to suicidal ideation when depression was not controlled for, whereas global AS was related to suicide attempts, controlling for depression. The use of latent variables and bifactor modeling in the current study allowed us to avoid problems in past studies regarding skewed suicide outcome variables and suppression effects due to the high degree of overlap between the AS subdimensions. Whereas prior studies reported a positive relation between AS cognitive concerns and suicidal ideation, they also reported a negative relation between AS physical concerns and suicidal ideation in several studies (Capron et al., 2012a,
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 1. Structural equation model examining the relations between the global AS and specific factors and the suicidal ideation factor (i.e., S. ideation) and past suicide attempts (i.e., S. attempts). Standardized parameter estimates are provided for relations with the suicidal ideation factor and unstandardized parameter estimates are provided for the relations with past suicide attempts. Significant path estimates are represented by solid lines and nonsignificant path estimates are represented by dashed lines.
Table 3 Structural equation model parameters examining the relations between the AS and depression factors and the suicidal ideation factor and past suicidal ideation. Suicidal ideation factor Factors
β
SE
Global AS Cognitive Physical Social Depression
.09 .46 .28 .29 .58n
.17 .17 .11 .12 .17
Past suicide attempts R
2
B
SE
.31n .11 .14 .06 .35nn
.14 .15 .10 .12 .13
.71
R2 .41
Note. SE ¼Standard error. nn n
p o.01. po .05.
2012b), implying that high AS physical concerns might act as a protective factor for the development of suicidal ideation. However, given that prior bifactor studies indicate that global AS encompasses the core features of AS, including most of the variance attributable to AS physical concerns (Allan et al., 2015), the positive relation between the global AS factor and suicidal ideation contradicts these prior studies. The bivariate relations in these prior studies support the findings of the current study and suggests suppression effects were accounting for the significant negative relation between AS physical concerns and suicidal ideation (controlling for AS cognitive concerns), as bivariate correlations between AS physical concerns and suicidal ideation and suicidality were positive and significant across most studies (although all bivariate correlations were positive, not all were significantly so (Capron et al., 2012a, 2012b, 2014). Therefore, it appears that general sensitivity to anxious arousal as well as sensitivity specific to cognitive concerns are related to suicidal ideation.(Table 3) In contrast to prior studies (Capron et al., 2012a, 2012b), the
relations between the AS factors and suicidal ideation were no longer significant when accounting for depression symptoms. This discrepancy between the current study and past studies is likely due to the use of latent variables in the current study. This approach reduced the level of measurement error in these constructs, and allowed for suicidal ideation to be modeled as a normally distributed continuous variable (Bollen, 1989). Whereas AS was not associated with suicidal ideation when controlling for depression, prior studies have found evidence that AS is concurrently associated with and prospectively predictive of depression (Grant et al., 2007; Zavos et al., 2012; Allan et al., 2014). Given the causal links between AS and depression, and between depression and suicidal ideation, it is possible that the relation between and suicidal ideation is mediated through depression. Whereas AS was not uniquely related to suicidal ideation, global AS was uniquely associated with suicide attempts. Although there are few studies directly examining global AS and suicide attempts, controlling for depression, Capron et al. (2012b) found higher mean levels of AS in individuals with a suicide attempt compared to individuals without an attempt, although they did not control for depression. Osman et al. (2010) found significant relations between global AS as measured by a bifactor and a measure of suicidality including a question asking whether an individual had attempted or thought about attempting suicide in the past, controlling for depression. However, that study was conducted in college-age undergraduates, and it was not possible to distinguish between suicidal ideation and past suicide attempts. Given the empirical and theoretical links between AS and panic disorder (PD; Olatunji and Wolitzky-Taylor, 2009; Allan et al., 2015), studies exploring the relations between PD and suicide attempts might provide some insight into the relation between AS and suicide attempts. Several recent epidemiological studies provide converging evidence of a link between PD and suicide attempts (Nock et al., 2010; Katz et al., 2011). As Nock et al. (2010)
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
6
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
discusses regarding PD, the relation between AS and suicide attempts is consistent with Shneidman's (1993) theory of psychache or intolerable psychological/emotional distress, from which escape is sought. A defining feature of this theory is that individual differences in thresholds for tolerating psychache reflect thresholds for suicidal behavior. Therefore, AS might operate within Shneidman's (1993) theory as contributing to establishing an individual's tolerance level. There are important clinical implications of the current findings. Recent studies have not only demonstrated that AS is a malleable risk factor that can be targeted through brief (one-session) interventions (Keough and Schmidt, 2012; Schmidt et al., 2014), but also that reductions in AS through these interventions lead to subsequent reductions in psychopathology, including depression, and suicidal ideation (Schmidt et al., 2014). Indeed, a recent intervention targeting AS cognitive concerns not only reduced AS, but also subsequently reduced depression and suicidal ideation through reductions in AS cognitive concerns. The current findings suggest targeting both AS physical and cognitive concerns might show an even greater impact on mitigating future suicidal ideation and attempts. There are several limitations to consider when interpreting the current results. The relations between AS and suicide were examined using self-report measures, leaving open the possibility of method bias influencing these results. Other methods of measuring AS, such as direct measures through fear-challenge paradigms (Schmidt and Mallott, 2006) or neurobiological indicators (Sehlmeyer et al., 2010) could usefully be employed to further clarify the relations between AS and suicidal ideation and attempts. The sample size of the current study in conjunction with the SEM approach did not allow for examination of possible important external correlates. In particular, it would be important to understand whether the relations between AS and suicidal ideation and attempts differ by gender or across distinct psychopathological diagnoses. However, it should be noted that not only was gender not a significant predictor of suicidal ideation or attempts when included in the current study, inclusion of gender did not substantively change the relations that AS and depression shared with suicidal ideation or attempts. Another limitation is the use of concurrent measures to examine the relations between AS and suicide. Although difficult in conceptualization and execution, carefully crafted studies are needed to determine whether AS is longitudinally associated with suicidal ideation and attempts. Finally, because the current study was conducted in a clinical sample that was not selected for suicide attempts, it is unclear how these findings might generalize to community samples or samples comprising individuals selected for past suicide attempts. The limitations of the current study notwithstanding, there are several valuable contributions of the current study. The use of latent variable modeling in conjunction with a bifactor approach to modeling AS allowed us to determine the most integral aspects of the AS/suicidality relation. In particular, the approach used in the current study was able to correct the misconception that AS physical concerns operated as a protective factor for suicidal ideation and suicide attempts. Whereas the current study demonstrated that global AS was the driving force behind the relations between AS and suicide attempts when the cognitive concerns factor was modeled separately from a general sensitivity to anxiety, this does not suggest that studies examining the relations between AS cognitive concerns and suicide variables are invalid. Rather, these relations are likely to be distorted if other highly correlated variables (i.e., AS physical concerns) are included in modeling approaches other than the bifactor approach. Finally, these results are in support of Shneidman's theory of intolerable distress. Future studies should be conducted to further explore the relations between AS and this model of suicide.
Contributors Nicholas P. Allan developed the research questions, conducted data analysis, and wrote substantial portions of the introduction and discussion. Aaron M. Norr wrote substantial portions of the manuscript, including most of the introduction and helped hone and develop the research questions. Jay W. Boffa wrote substantial portions of the manuscript, including parts of the introduction. Daphne Durmaz wrote substantial portions of the manuscript, including parts of the introduction. Amanda M. Raines wrote substantial portions of the manuscript, including the methods and reviewed and edited the manuscript. Norman B. Schmidt reviewed and edited the manuscript and helped develop the research questions. All authors contributed to and have approved the final manuscript.
Conflict of interest All authors report no conflicts of interest.
Funding Information This investigation was supported in part by the Military Suicide Research Consortium (MSRC), an effort supported by the Office of the Assistant Secretary of Defense for Health Affairs under Award no. (W81XWH-10-2-0181). Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the MSRC or the Department of Defense.
References Allan, N.P., Albanese, B.J., Short, N.A., Raines, A.M., Schmidt, N.B., 2015. Support for the general and specific bifactor model factors of anxiety sensitivity. Personal. Individ. Differ. 74, 78–83. Allan, N.P., Capron, D.W., Raines, A.M., Schmidt, N.B., 2014. Unique relations among anxiety sensitivity factors and anxiety, depression, and suicidal ideation. J. Anxiety Disord. 28, 266–275. Beck, A.T., Steer, R.A., Brown, G.K., 1996. Beck Depression Inventory Manual, 2nd ed. Psychological Corporation, San Antonio, TX. Bollen, K., 1989. Structural Equation Models. Wiley, New York, NY. Capron, D.W., Cougle, J.R., Ribeiro, J.D., Joiner, T.E., Schmidt, N.B., 2012a. An interactive model of anxiety sensitivity relevant to suicide attempt history and future suicidal ideation. J. Psychiatr. Res. 46, 174–180. Capron, D.W., Fitch, K., Medley, A., Blagg, C., Mallott, M., Joiner, T., 2012b. Role of anxiety sensitivity subfactors in suicidal ideation and suicide attempt history. Depress. Anxiety 29, 195–201. Capron, D.W., Lamis, D. a, Schmidt, N.B., 2014. Test of the depression distress amplification model in young adults with elevated risk of current suicidality. Psychiatry Res. 219, 531–535. Capron, D.W., Norr, A.M., Macatee, R.J., Schmidt, N.B., 2013. Distress tolerance and anxiety sensitivity cognitive concerns: testing the incremental contributions of affect dysregulation constructs on suicidal ideation and suicide attempt. Behav. Ther. 44, 349–358. Chen, F.F., Hayes, A., Carver, C.S., Laurenceau, J.P., Zhang, Z., 2012. Modeling general and specific variance in multifaceted constructs: a comparison of the bifactor model to other approaches. J. Personal. 80, 219–251. Delucchi, K.L., Bostrom, A., 2004. Methods for analysis of skewed data distributions in psychiatric clinical studies: working with many zero values. Am. J. Psychiatry 161, 1159–1168. Ebesutani, C., McLeish, A.C., Luberto, C.M., Young, J., Maack, D.J., 2013. A bifactor model of anxiety sensitivity: analysis of the Anxiety Sensitivity Index-3. J. Psychopathol. Behav. Assess. Grant, D.M., Beck, J.G., Davila, J., 2007. Does anxiety sensitivity predict symptoms of panic, depression, and social anxiety? Behav. Res. Ther. 45, 2247–2255. Grant, S., 2010. What's normal anyway? Normal and non-normal distributions in psychiatry. Acta Neuropsychiatr. 22, 305–310. Hu, L., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct. Equ. Model. Multidiscip. J. 6, 1–55. Joiner, T.E., Rudd, M.D., 1996. Disentangling the interrelations between hopelessness, loneliness, and suicidal ideation. Suicide Life Threat. Behav. 26, 19–26. Katz, C., Yaseen, Z.S., Mojtabai, R., Cohen, L.J., Galynker, I.I., 2011. Panic as an
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎ independent risk factor for suicide attempt in depressive illness: findings from the National Epidemiological Survey on Alcohol and Related Conditions (NESARC). J. Clin. Psychiatry 72, 1628–1635. Keough, M.E., Schmidt, N.B., 2012. Refinement of a brief anxiety sensitivity reduction intervention. J. Consult. Clin. Psychol. 80, 766–772. Kessler, Ronald C., Borges, G., Walters, E.E., 1999. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch. Gen. Psychiatry 56, 617–626. Kraemer, H.C., Kazdin, A.E., Offord, D.R., Kessler, R.C., Jensen, P.S., Kupfer, D.J., 1997. Coming to terms with the terms of risk. Arch. Gen. Psychiatry 54, 337–343. McDonald, R.P., 1999. Test theory: A unified treatment. Erlbaum, Mahwah, NJ. Metalsky, G.I., Joiner, T.E., 1997. The hopelessness depression symptom questionnaire. Cogn. Ther. Res. 21, 359–384. Mohlman, J., Zinbarg, R.E., 2000. The structure and correlates of anxiety sensitivity in older adults. Psychol. Assess. 12, 440–446. Moshagen, M., 2012. The model size effect in SEM: inflated goodness-of-fit statistics are due to the size of the covariance matrix. Struct. Equ. Model. Multidiscip. J. 19, 86–98. Mulaik, S., 2007. There is a place for approximate fit in structural equation modelling. Personal. Individ. Differ. 42, 883–891. Muthén, L.K., Muthén, B.O., 1996 2012. Mplus User's Guide, seventh ed. Muthén & Muthén, Los Angeles, CA. Nock, M.K., Hwang, I., Sampson, N.A., Kessler, R.C., 2010. Mental disorders, comorbidity and suicidal behavior: results from the National Comorbidity Survey Replication. Mol. Psychiatry 15, 868–876. Oglesby, M.E., Capron, D.W., Raines, A.M., Schmidt, N., 2015. Anxiety sensitivity cognitive concerns predict suicide risk. Psychiatry Res. 226, 252–256. Olatunji, B.O., Wolitzky-Taylor, K.B., 2009. Anxiety sensitivity and the anxiety disorders: a meta-analytic review and synthesis. Psychol. Bull. 135, 974–999. Osman, A., Bagge, C.L., Gutierrez, P.M., Konick, L.C., Kopper, B.A., Barrios, F.X., 2001. The Suicidal Behaviors Questionnaire-Revised (SBQ-R): validation with clinical and nonclinical samples. Assessment 8, 443–454. Osman, A., Gutierrez, P.M., Smith, K., Fang, Q., Lozano, G., Devine, A., 2010. The Anxiety Sensitivity Index-3: analyses of dimensions, reliability estimates, and correlates in nonclinical samples. J. Personal. Assess. 92, 45–52. Reise, S.P., 2012. Invited paper: the rediscovery of bifactor measurement models. Multivar. Behav. Res. 47, 667–696. Reiss, S., McNally, R.J., 1985. Expectancy model of fear. In: Theoretical Issues in Behavior Therapy. Academic Press, San Diego, CA, pp. 107–121. Reiss, S., Peterson, R.A., Gursky, D.M., McNally, R.J., 1986. Anxiety sensitivity, anxiety frequency and the prediction of fearfulness. Behav. Res. Ther. 24, 1–8. Rhemtulla, M., Brosseau-Liard, P.É., Savalei, V., 2012. When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychol. Methods 17,
7
354–373. Schmidt, N.B., Capron, D.W., Raines, A.M., Allan, N.P., 2014. Randomized clinical trial evaluating the efficacy of a brief intervention targeting anxiety sensitivity cognitive concerns. J. Consult. Clin. Psychol. Schmidt, N.B., Eggleston, A.M., Woolaway-Bickel, K., Fitzpatrick, K.K., Vasey, M.W., Richey, J.A., 2007. Anxiety Sensitivity Amelioration Training (ASAT): a longitudinal primary prevention program targeting cognitive vulnerability. J. Anxiety Disord. 21, 302–319. Schmidt, N.B., Mallott, M., 2006. Evaluating anxiety sensitivity and other fundamental sensitivities predicting anxiety symptoms and fearful responding to a biological challenge. Behav. Res. Ther. 44, 1681–1688. Schmidt, N.B., Zvolensky, M.J., Maner, J.K., 2006. Anxiety sensitivity: prospective prediction of panic attacks and Axis I pathology. J. Psychiatr. Res. 40, 691–699. Schmidt, Norman B., Woolaway-Bickel, K., Bates, M., 2001. Evaluating panic-specific factors in the relationship between suicide and panic disorder. Behav. Res. Ther. 39, 635–649. Sehlmeyer, C., Konrad, C., Zwitserlood, P., Arolt, V., Falkenstein, M., Beste, C., 2010. ERP indices for response inhibition are related to anxiety-related personality traits. Neuropsychologia 48, 2488–2495. Shneidman, E.S., 1993. Suicide as psychache. J. Nerv. Ment. Dis. 181, 145–147. Souminen, K., Isometsӓ, E., Suokas, J., Haukka, J., Achte, K., Lönnqvist, J., 2004. Completed suicide after a suicide attempt: a 37-year follow-up study. Am. J. Psychiatry 161, 562–563. Taylor, S., Zvolensky, M.J., Cox, B.J., Deacon, B., Heimberg, R.G., Ledley, D.R., Abramowitz, J.S., Holaway, R.M., Sandin, B., Stewart, S.H., Coles, M., Eng, W., Daly, E. S., Arrindell, W.A., Bouvard, M., Cardenas, S.J., 2007. Robust dimensions of anxiety sensitivity: Development and initial validation of the Anxiety Sensitivity Index-3. Psychol. Assess. 19, 176–188. Timpano, Kiara R., Schmidt, N.B., 2012. The relationship between self-control deficits and hoarding: A multimethod investigation across three samples. J. Abnorm. Psychol. 122, 13–25. 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. J. Consult. Clin. Psychol. 76, 72–83. Weissman, M.M., Klerman, G.L., Markowitz, J.S., Oullette, R., 1989. Suicidal ideation and suicide attempts in panic disorder and attacks. N. Engl. J. Med. 321, 1209–1214. World Health Organization, 2012. Public health action for the prevention of suicide: a framework. Zavos, H.M.S., Rijsdijk, F.V., Eley, T.C., 2012. A longitudinal, genetically informative, study of associations between anxiety sensitivity, anxiety and depression. Behav. Genet. 42, 592–602.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts Nicholas P. Allan n, Aaron M. Norr, Jay W. Boffa, Daphne Durmaz, Amanda M. Raines, Norman B. Schmidt n Department of Psychology, Florida State University, 1107 W. Call St., Tallahassee, FL 32306-4301, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 31 January 2015 Received in revised form 12 May 2015 Accepted 29 May 2015
Anxiety sensitivity (AS) has recently been linked to suicidality. Specifically, AS cognitive concerns has been implicated as a risk factor, and AS physical concerns as a protective factor, for suicidal ideation and suicide attempts. However, no studies have used structural equation modeling (SEM) to address issues of skewed suicide variables and bifactor modeling of AS to address the high degree of overlap between the lower-order dimensions of AS that limit interpretation of these past findings. AS, suicidal ideation, past suicide attempts, and depression were assessed in a clinical sample of 267 individuals (M age¼ 35.45 years, SD ¼16.53; 52.1% female). The global AS and AS cognitive concerns factors were positively, significantly associated with suicidal ideation, though these effects were nonsignificant controlling for depression. The global AS factor was positively, significantly associated with suicide attempts, controlling for depression. The current study demonstrated that the relations between AS and suicidal ideation are not maintained when accounting for depression, suggesting that the relation between AS and suicidal ideation may be mediated by depression. The positive relation between global AS and suicide attempts is consistent with theories positing suicide attempts as a consequence of an inability to cope with intolerable distress. & 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Suicidal ideation Suicide attempts Anxiety sensitivity Bifactor modeling Depression
1. Introduction According to the most recent report by the World Health Organization, suicide accounts for over one million deaths a year, and is among the top three leading causes of death among individuals 15–44 years old (World Health Organization, 2012), making it a critical public health issue. Existing research indicates that completed suicides are strongly predicted by prior suicide attempts as well as suicidal ideation ( Kessler et al., 1999; Souminen et al., 2004; Weissman et al., 1989). Therefore, it is important to identify and understand potential malleable risk factors (Kraemer et al., 1997) for suicidal behavior to better predict and reduce completed suicides. One risk factor that has been linked to suicide attempts and suicidal ideation is anxiety sensitivity (AS), or the fear of anxiety-related sensations (Reiss et al., 1986). Anxiety sensitivity (AS) is associated with and prospectively predicts the development of both anxiety and mood pathology (Schmidt et al., 2006; Olatunji and Wolitzky-Taylor, 2009), and appears to be malleable through intervention (Schmidt et al., n
Corresponding authors. E-mail addresses: [email protected] (N.P. Allan), [email protected] (N.B. Schmidt).
2007). AS reflects an individual's tendency to respond fearfully to symptoms of anxiety arising from the belief that these symptoms will have negative consequences (Reiss and McNally, 1985). Individuals with elevated AS are likely to interpret benign symptoms of anxiety as potentially harmful or dangerous. AS is composed of three lower-order dimensions: cognitive, physical, and social concerns. AS cognitive concerns reflect the tendency to respond anxiously to feelings of cognitive dyscontrol, such as experiencing racing thoughts. AS physical concerns reflect the tendency to respond anxiously to physiological symptoms of anxiety, such as shortness of breath. AS social concerns reflect the tendency to fear potential negative evaluations resulting from others noticing symptoms of anxiety, such as sweating. In addition to being a risk factor for anxiety and mood pathology broadly, AS also has a robust relationship with suicidality (Schmidt et al., 2001). There is some evidence to suggest that much of this association is driven primarily by the AS cognitive concerns component. Capron and colleagues (Capron et al., 2012b) found in a clinical outpatient sample that the AS cognitive concerns scale was significantly related to suicidal ideation and previous suicide attempts. This positive association between AS cognitive concerns and suicidal ideation was prospectively demonstrated in a large sample of cadets entering basic training at the United States Air Force Academy as well as in a sample of clinic
http://dx.doi.org/10.1016/j.psychres.2015.05.066 0165-1781/& 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
outpatient who screened positive for PTSD (Capron et al., 2012a). Furthermore, the association between AS cognitive concerns and suicidality remain significant after accounting for other relevant risk variables, such as distress tolerance, gender, insomnia, thwarted belongingness, and depressive symptoms (Capron et al., 2013; Oglesby et al., 2015). Although other AS subscales have been found to be related to suicidality, results have been inconsistent. For example, Capron et al. (2012b) found that AS social concerns was positively associated with previous suicide attempts in one study. In addition, negative relations between AS physical concerns and suicidal ideation have been found (Capron et al., 2012a, 2012b), indicating there may be a specific role for AS physical concerns as a protective factor for suicidality. Although relations between AS and suicide have been demonstrated across various samples, there are limitations to the prior studies. First, it is not entirely clear whether the relation between AS and suicidality can be fully accounted for by AS cognitive concerns or whether a global AS factor is also related. As suggested by Mohlman and Zinbarg (2000), AS appears to comprise a global factor in addition to the three lower-order dimensions. As previous work on AS and suicidality has relied on scale scores, it has been impossible to examine both overall AS and the specific dimensions in the same model. One approach that can clarify the true relation between global AS, specific AS dimensions, and suicidality is bifactor modeling (Chen et al., 2012; Reise, 2012). In this approach, common variance, representing AS more generally, can be distinguished from variance specific to the lowerorder factors (i.e., cognitive, physical, and social concerns). In fact, several authors have used this approach to validate and implement a bifactor model of AS (Mohlman and Zinbarg, 2000; Osman et al., 2010; Ebesutani et al., 2013; Allan et al., 2015). Thus, a bifactor model of AS will allow for the investigation of both the global AS factor and specific AS dimensions simultaneously in the prediction of suicidality. Only a single study has examined the relations between the AS factors, using bifactor modeling, and suicidality. Osman et al. (2010) found that suicidality was associated with a global AS factor as well as the specific AS physical and cognitive concerns factors. However, these results should be interpreted with caution as their study was conducted in an undergraduate sample with low reported rates of suicidality (i.e., Suicide Behavior Questionnaire-Revised [SBQ-R; Osman et al., 2001] M¼ 4.43, SD ¼2.11; minimum possible total score on SBQ-R is 4.00). Related to this, a second limitation to be addressed in the previous AS and suicidality literature is the skewed nature of suicidal ideation. Given that suicidality is absent in most individuals, suicidality variables tend to have a significant positive skew and be heavily leptokurtic (Van Orden et al., 2008; Capron et al., 2012b). The issues associated with statistical analysis of significantly skewed data have been well documented (Delucchi and Bostrom, 2004; Grant, 2010). Most previous work on AS and suicidality has either not explicitly dealt with the skewed suicidality variables or conducted a logarithmic transformation (Capron et al., 2013), which does not remedy the issues associated with a significant portion of the sample containing the same value on the suicidality variable (i.e., 0, absence of suicidality; Delucchi and Bostrom, 2004). Modeling suicidal ideation treating the items as categorical indicators of a continuous latent variable will remedy issues associated with a highly skewed dependent scale, as this approach does not require normally distributed indicators (Bollen, 1989; Rhemtulla et al., 2012). 1.1. The current study In the current study, we advanced the literature on AS and suicidality using improved statistical methods in a clinical
outpatient sample. First, by modeling AS as a bifactor we were able to investigate the unique contributions of both the global and specific AS factors to suicidal ideation and past suicide attempts. Second, we improved upon the literature by modeling suicidal ideation as a continuous latent variable with categorical indicators to account for the significantly skewed nature of this variable. We hypothesized that using the bifactor model of AS, we would find a significant positive relation between AS cognitive concerns and suicidal ideation and past suicide attempts as has been found in previous studies (Capron et al., 2012b; Oglesby et al., 2015). Second, we hypothesized that there would be a significant negative relation between the global AS factor and suicidal ideation and past attempts. This hypothesis was based on past studies demonstrating a negative relation between AS physical concerns and suicidality (Capron et al., 2012a) as well as studies indicating that the global AS factor predominantly captures AS physical concerns (Allan et al., 2015). We hypothesized that AS physical and social concerns would not be associated with suicidal ideation or past attempts after accounting for the relations between AS cognitive concerns and the global AS factor. Finally, we hypothesized that the relations that global AS and AS cognitive concerns share with suicidal ideation and past suicide attempts would remain after accounting for depression.
2. Methods 2.1. Participants and setting The sample included 267 individuals presenting to an outpatient anxiety clinic to receive psychological services and/or participate in research options. Demographic and information regarding participant's current primary psychiatry diagnosis are provided in Table 1. Gender was fairly evenly distributed (52.1% female) with ages ranging from 18 to 88 (M¼35.45, SD¼ 16.53). The racial composition of the sample was as follows: 63.7% Caucasian, 23.6% African American, 1.5% Asian, 4% Native American, 4% Pacific Islander, and 10.4% other (e.g., biracial) with 11.6% identifying as Hispanic. The majority of the sample attended some college (53.6%), 16.9% obtained a 4-year degree, 12.0% had a high school diploma, 10.1% had a graduate degree, 4.1% trade school degree, and 3.3% had less than a high school education. Regarding primary psychiatric diagnoses, 45.6% of sample met for a primary anxiety disorder, 18.4% mood disorder, 10.9% trauma and stressorrelated disorder, 6.4% substance-related disorder, 2.7% obsessive– compulsive and related disorder, 2.7% other (e.g., somatic disorder) and 13.5% had no primary diagnosis. 2.2. Procedure All individuals included in the current study agreed to participate in Institutional Review Board-approved research being conducted at the clinic. Prior to initiating treatment or research options, all individuals were interviewed by an advanced level clinical psychology graduate student to assess for the presence of psychopathology. All diagnoses were confirmed at a weekly supervision meeting with a licensed clinical psychologist and director of the clinic. Data used in the current investigation were part of a battery of questionnaires given after completion of the diagnostic interview. Participants were excluded from the current study if they were unable to fully complete their diagnostic interview, including their self-report questionnaires. All other participants who were given a structured clinical interview were included in the current study.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
confidence interval [CI] ¼.76, .85), but low for the specific AS Physical Concerns (ωs ¼.32, 95% CI ¼.15, .46), AS Cognitive Concerns (ωs ¼.29, 95% CI ¼ .03,.44), and AS Social Concerns (ωs ¼ .30, 95% CI ¼ .05, .44) factors.1
Table 1 Demographic features and primary axis I diagnoses. Demographics
Age
Mean 35.45
SD 16.53
Gender Male Female
n 128 139
% 47.9 52.1
Race Caucasian African American Asian Native American Pacific Islander Other
170 63 4 1 1 28
63.7 23.6 1.5 0.4 0.4 10.4
Education Graduate degree 4-year degree Associates or some college High school degree Trade school degree Less than high school
27 45 143 32 11 9
10.1 16.9 53.6 12.0 4.1 3.4
Psychiatric diagnosis Anxiety Mood Trauma/stressor-related Alcohol/substance use OCD/related Other No diagnosis
122 49 29 17 7 7 36
3
45.6 18.4 10.9 6.4 2.7 2.7 13.5
Note. SD ¼standard deviation. Other includes biracial or wishing not to report. OCD¼ Obsessive–compulsive disorder.
2.3. Measures 2.3.1. Clinician administered 2.3.1.1. Structured clinical interview for DSM-IV and DSM-5 (SCID). All psychiatric diagnoses were determined using the SCID. Diagnostic interviews were administered by trained doctoral level therapists who completed extensive training in the administration and scoring of the SCID. This training included reviewing SCID training tapes, observing live administrations, and conducting practice interviews with other experienced individuals. During the training process all trainees received feedback until they demonstrated high levels of reliability. Rates of agreement between interviewers within our laboratory have been found to be over 80% with a kappa value of .77 (Timpano and Schmidt, 2012). 2.3.2. Self-report 2.3.2.1. Anxiety sensitivity index-3 (ASI-3). The ASI-3 is an 18-item self-report questionnaire assessing fears of anxious arousal (Taylor et al., 2007) adapted from the Anxiety Sensitivity Index (Reiss et al., 1986). Individuals were instructed to rate how much they agreed with each item on a 5-point Likert-type scale ranging from 0 (very little) to 4 (very much). In addition to an overall total score, the ASI-3 yields three subscales: physical, cognitive, and social concerns. Prior research has demonstrated that the ASI-3 is both a valid and reliable measure of AS (Taylor et al., 2007). In the current investigation, the ASI-3 total and aggregate subscale scores demonstrated good to excellent internal consistency (αs ¼.84 .93). Reliability for the bifactor solution (calculated using McDonald's omega hierarchical [ωh; ωs for the specific factors]; McDonald, 1999; Reise, 2012) was good for the global AS factor (ωh ¼.81, 95%
2.3.2.2. Beck depression inventory-II (BDI-II). The BDI-II is a 21-item self-report questionnaire assessing common symptoms of depression (Beck et al., 1996). In addition to a total score, the BDI-II contains cognitive and affective/somatic subscales (Beck et al., 1996). The BDI-II has been found to be psychometrically sound, with high internal consistency (Beck et al., 1996). In the present study, the cognitive and affective/somatic subscales demonstrated good internal consistency (αs ¼.87,.89; ωhs ¼.88, 95% CI [.86, .90], .88, 95% CI [.86, .90]). 2.3.2.3. Depression symptom inventory-suicide subscale (DSI-SS). The DSI-SS is 4-item self-report questionnaire assessing suicidal ideation during the past two weeks (Metalsky and Joiner, 1997). Items are rated using a 4-point Likert-type scale ranging from 0-3 with higher scores reflecting more severe symptoms. Prior research has demonstrated that the DSI-SS has sound psychometric properties (Joiner and Rudd, 1996). In the present investigation, internal consistency was good (α ¼.88; ωh ¼.91, 95% CI [.88, .93]). 2.3.2.4. Suicide history. Suicide history was assessed by a question on a form seeking medical history asking individuals to indicate prior attempts. 2.4. Data analytic plan AS models were first examined in Mplus version 7 (Muthén and Muthén, 1996 2012) treating items as continuous and using full information maximum likelihood with the Yuan Bentler scaled chi-square (χ2) value to account for missing data and nonnormality. The three-factor correlated-traits model was compared to the bifactor model using the corrected likelihood ratio test. Overall model fit was assessed using the χ2 statistic, for which a nonsignificant value indicates good fit. However, this statistic can be too restrictive, especially with many items per factor (Hu and Bentler, 1999; Mulaik, 2007; Moshagen, 2012). Model fit was therefore additionally examined using the comparative fit index (CFI), Tucker Lewis index (TLI), and root mean square error of approximation (RMSEA) with accompanying 90% confidence interval (CI). CFI and TLI values from above .90 to above .95 and RMSEA values less than .08 to less than .05 are indicative of acceptable to good fit. Further, an RMSEA lower-bound CI below .05 indicates that good fit cannot be ruled out and an upper-bound CI above .10 indicates that poor fit cannot be ruled out. The best-fitting model was then included in a structural equation model (SEM) examining the relations between the AS factors and a suicidal ideation factor (comprised of items from the DSI-SS, treated categorically) and past suicide attempts (0 ¼ not past attempts, 1 ¼one or more past attempts) treated as a manifest variable. Because the outcome variables were categorical, the default robust weighted least squares estimator (WLSMV) was used and model fit was assessed similar to the AS CFA models. Finally, to determine if there were unique relations between AS and the suicide outcome variables controlling for depression, the relations between the AS factors and suicidal ideation and past suicide attempts were examined in a model including a second-order depression factor (with first-order cognitive and affective/somatic 1 All ω's were calculated in Mplus using robust maximum likelihood, treating the data as continuous, and including 1000 bias-corrected bootstrap draw to provide asymmetric confidence intervals.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
significantly better (Δ χ2 ¼68.44, df ¼17, p o.001). All items loaded significantly on the global AS factor as well as on their specific AS cognitive, physical, and social concerns factors.
factors, with item-level data modeled categorically).
3. Results 3.3. Structural equation modeling of the relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts
3.1. Descriptive statistics and correlations Although all variables (with the exception of past suicide attempts) were modeled as latent variables, scale scores were computed to provide means and standard deviations as well as correlations between variables (see Table 2). There were significant correlations among all the variables with the exception of ASI-3 physical concerns and DSI-SS suicidal ideation. There was little missing data (between 0 and 6 participants missing data per item, 11 participants were missing information about past suicide attempt) and all participants were included in the analyses. Within the current sample, 33 participants (12.9%) reported one prior attempt, and 18 (7.1%) reported more than one prior attempt. For the current study, participants were dichotomized into no past attempts (coded as 0) or one or more past attempts (coded as 1). Mean DSI-SS scores were.80 (SD ¼1.61), providing evidence that suicidal ideation would be extremely skewed if it had been treated as continuous. 3.2. Confirmatory factor analysis of anxiety sensitivity CFAs of the three-factor and bifactor model of the ASI-3 were compared. The three-factor model provided acceptable fit to the data (χ2 ¼256.18, df ¼132, p o.05, CFI ¼.95, TLI¼.94, RMSEA ¼.06, 90% CI [.05,.07]) as did the bifactor model (χ2 ¼183.19, df ¼ 117, p o.05, CFI¼ .97, TLI ¼.96, RMSEA ¼ .05, 90% CI [.03,.06]). Comparison of the models indicated that the bifactor model fit Table 2 Descriptive statistics and correlations for ASI-3 subscales, depression subscales, suicidal ideation, and past suicide attempts. 1
2
3
4
5
6
7
8
1. ASI-3 cognitive
2. ASI-3 physical
.61n
3. ASI-3 social
.62n
.50n
4. ASI-3 total
.88n
.83n
.82n
5. BDI-2 cognitive
.56n
.31n
.48n
.52n
6. BDI-2 affective/somatic
.58n
.38n
.43n
.54n
.69n
7. BDI-2 total
.61n
.37n
.48n
.57n
.90n
.94n
8. DSI-SS suicidal ideation
.24n
.06
.17n
.16n
.48n
.34n
.44n
9. Past suicide attempts
.42n
.18n
.27n
.33n
.41n
.36n
.40n
.28n
Mean (% attempts) Standard deviation
9.47 9.39 12.70 31.32 9.29 13.57 22.84 .80 6.67 6.21 5.93
15.72 5.86 7.55
12.27
9
po .05.
3.4. Structural equation modeling of the relations between anxiety sensitivity and depression factors and suicidal ideation and past suicide attempts The SEM examining the relations the global and specific AS factors, and the second-order depression factor (comprised of first-order cognitive and affective/somatic factors) share with the suicidal ideation factor and past suicide attempts provided acceptable fit to the data (χ2 ¼ 244.14, df ¼204, po .05, CFI ¼.98, TLI¼.98, RMSEA ¼ .03, 90% CI [.01, .04]). Model parameters are provided in Table 2. In this model, there was a significant positive relation between the suicidal ideation and depression factors (β ¼ .58, p o.05). There were also significant positive relations between past suicide attempts and the global AS (B ¼.31, po .05) and depression factors (B ¼ .35, p o.01). Controlling for all other variables, a 1 SD increase in the global AS factor was associated with a 30% probability of a past suicide attempt and a 1 SD increase in the depression factor was associated with a 31% probability of a past suicide attempt. This model accounted for 71% of the variance in the suicidal ideation factor and 41% of the variance in past suicide attempts.
4. Discussion
20.0%
1.61
Note. ASI-3 ¼anxiety sensitivity index-3. BDI-2 ¼Beck depression inventory-2. DSISS ¼ depressive symptom inventory-suicidal scale. n
The SEM examining the relations past suicidal ideation shares with the global and specific AS factors as well as the suicidal ideation factor and past suicide attempts provided acceptable fit to the data (χ2 ¼ 244.14, df ¼ 204, p o.05, CFI¼.98, TLI¼.98, RMSEA¼ .03, 90% CI [.01,.04]). Standardized factor loadings for the AS and suicidal ideation factors as well as standardized model parameters (unstandardized probit regression parameters are reported for past suicide attempts) are provided in Fig. 1. There were significant positive relations between the suicidal ideation factor and the global AS (β ¼.22, p o.05) and AS cognitive concerns factors (β ¼ .26, po .05). There were also significant positive relations between past suicide attempts and the slobal AS factor (B ¼.50, po .001), such that, controlling for the specific AS factors, a standard deviation (SD) increase in the global AS factor was associated with a 36% probability of having a past suicide attempt. This model accounted for 13% of the variance in the suicidal Ideation factor and 33% of the variance in past suicide attempts. Results were also conducted including gender as a covariate and the relations between the AS factors and suicidal ideation and past suicide attempts were unchanged. Further, gender was not significantly associated with either outcome variable. Therefore, only the model excluding gender was reported.
In the current study, global AS and AS cognitive concerns were only related to suicidal ideation when depression was not controlled for, whereas global AS was related to suicide attempts, controlling for depression. The use of latent variables and bifactor modeling in the current study allowed us to avoid problems in past studies regarding skewed suicide outcome variables and suppression effects due to the high degree of overlap between the AS subdimensions. Whereas prior studies reported a positive relation between AS cognitive concerns and suicidal ideation, they also reported a negative relation between AS physical concerns and suicidal ideation in several studies (Capron et al., 2012a,
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 1. Structural equation model examining the relations between the global AS and specific factors and the suicidal ideation factor (i.e., S. ideation) and past suicide attempts (i.e., S. attempts). Standardized parameter estimates are provided for relations with the suicidal ideation factor and unstandardized parameter estimates are provided for the relations with past suicide attempts. Significant path estimates are represented by solid lines and nonsignificant path estimates are represented by dashed lines.
Table 3 Structural equation model parameters examining the relations between the AS and depression factors and the suicidal ideation factor and past suicidal ideation. Suicidal ideation factor Factors
β
SE
Global AS Cognitive Physical Social Depression
.09 .46 .28 .29 .58n
.17 .17 .11 .12 .17
Past suicide attempts R
2
B
SE
.31n .11 .14 .06 .35nn
.14 .15 .10 .12 .13
.71
R2 .41
Note. SE ¼Standard error. nn n
p o.01. po .05.
2012b), implying that high AS physical concerns might act as a protective factor for the development of suicidal ideation. However, given that prior bifactor studies indicate that global AS encompasses the core features of AS, including most of the variance attributable to AS physical concerns (Allan et al., 2015), the positive relation between the global AS factor and suicidal ideation contradicts these prior studies. The bivariate relations in these prior studies support the findings of the current study and suggests suppression effects were accounting for the significant negative relation between AS physical concerns and suicidal ideation (controlling for AS cognitive concerns), as bivariate correlations between AS physical concerns and suicidal ideation and suicidality were positive and significant across most studies (although all bivariate correlations were positive, not all were significantly so (Capron et al., 2012a, 2012b, 2014). Therefore, it appears that general sensitivity to anxious arousal as well as sensitivity specific to cognitive concerns are related to suicidal ideation.(Table 3) In contrast to prior studies (Capron et al., 2012a, 2012b), the
relations between the AS factors and suicidal ideation were no longer significant when accounting for depression symptoms. This discrepancy between the current study and past studies is likely due to the use of latent variables in the current study. This approach reduced the level of measurement error in these constructs, and allowed for suicidal ideation to be modeled as a normally distributed continuous variable (Bollen, 1989). Whereas AS was not associated with suicidal ideation when controlling for depression, prior studies have found evidence that AS is concurrently associated with and prospectively predictive of depression (Grant et al., 2007; Zavos et al., 2012; Allan et al., 2014). Given the causal links between AS and depression, and between depression and suicidal ideation, it is possible that the relation between and suicidal ideation is mediated through depression. Whereas AS was not uniquely related to suicidal ideation, global AS was uniquely associated with suicide attempts. Although there are few studies directly examining global AS and suicide attempts, controlling for depression, Capron et al. (2012b) found higher mean levels of AS in individuals with a suicide attempt compared to individuals without an attempt, although they did not control for depression. Osman et al. (2010) found significant relations between global AS as measured by a bifactor and a measure of suicidality including a question asking whether an individual had attempted or thought about attempting suicide in the past, controlling for depression. However, that study was conducted in college-age undergraduates, and it was not possible to distinguish between suicidal ideation and past suicide attempts. Given the empirical and theoretical links between AS and panic disorder (PD; Olatunji and Wolitzky-Taylor, 2009; Allan et al., 2015), studies exploring the relations between PD and suicide attempts might provide some insight into the relation between AS and suicide attempts. Several recent epidemiological studies provide converging evidence of a link between PD and suicide attempts (Nock et al., 2010; Katz et al., 2011). As Nock et al. (2010)
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
6
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎
discusses regarding PD, the relation between AS and suicide attempts is consistent with Shneidman's (1993) theory of psychache or intolerable psychological/emotional distress, from which escape is sought. A defining feature of this theory is that individual differences in thresholds for tolerating psychache reflect thresholds for suicidal behavior. Therefore, AS might operate within Shneidman's (1993) theory as contributing to establishing an individual's tolerance level. There are important clinical implications of the current findings. Recent studies have not only demonstrated that AS is a malleable risk factor that can be targeted through brief (one-session) interventions (Keough and Schmidt, 2012; Schmidt et al., 2014), but also that reductions in AS through these interventions lead to subsequent reductions in psychopathology, including depression, and suicidal ideation (Schmidt et al., 2014). Indeed, a recent intervention targeting AS cognitive concerns not only reduced AS, but also subsequently reduced depression and suicidal ideation through reductions in AS cognitive concerns. The current findings suggest targeting both AS physical and cognitive concerns might show an even greater impact on mitigating future suicidal ideation and attempts. There are several limitations to consider when interpreting the current results. The relations between AS and suicide were examined using self-report measures, leaving open the possibility of method bias influencing these results. Other methods of measuring AS, such as direct measures through fear-challenge paradigms (Schmidt and Mallott, 2006) or neurobiological indicators (Sehlmeyer et al., 2010) could usefully be employed to further clarify the relations between AS and suicidal ideation and attempts. The sample size of the current study in conjunction with the SEM approach did not allow for examination of possible important external correlates. In particular, it would be important to understand whether the relations between AS and suicidal ideation and attempts differ by gender or across distinct psychopathological diagnoses. However, it should be noted that not only was gender not a significant predictor of suicidal ideation or attempts when included in the current study, inclusion of gender did not substantively change the relations that AS and depression shared with suicidal ideation or attempts. Another limitation is the use of concurrent measures to examine the relations between AS and suicide. Although difficult in conceptualization and execution, carefully crafted studies are needed to determine whether AS is longitudinally associated with suicidal ideation and attempts. Finally, because the current study was conducted in a clinical sample that was not selected for suicide attempts, it is unclear how these findings might generalize to community samples or samples comprising individuals selected for past suicide attempts. The limitations of the current study notwithstanding, there are several valuable contributions of the current study. The use of latent variable modeling in conjunction with a bifactor approach to modeling AS allowed us to determine the most integral aspects of the AS/suicidality relation. In particular, the approach used in the current study was able to correct the misconception that AS physical concerns operated as a protective factor for suicidal ideation and suicide attempts. Whereas the current study demonstrated that global AS was the driving force behind the relations between AS and suicide attempts when the cognitive concerns factor was modeled separately from a general sensitivity to anxiety, this does not suggest that studies examining the relations between AS cognitive concerns and suicide variables are invalid. Rather, these relations are likely to be distorted if other highly correlated variables (i.e., AS physical concerns) are included in modeling approaches other than the bifactor approach. Finally, these results are in support of Shneidman's theory of intolerable distress. Future studies should be conducted to further explore the relations between AS and this model of suicide.
Contributors Nicholas P. Allan developed the research questions, conducted data analysis, and wrote substantial portions of the introduction and discussion. Aaron M. Norr wrote substantial portions of the manuscript, including most of the introduction and helped hone and develop the research questions. Jay W. Boffa wrote substantial portions of the manuscript, including parts of the introduction. Daphne Durmaz wrote substantial portions of the manuscript, including parts of the introduction. Amanda M. Raines wrote substantial portions of the manuscript, including the methods and reviewed and edited the manuscript. Norman B. Schmidt reviewed and edited the manuscript and helped develop the research questions. All authors contributed to and have approved the final manuscript.
Conflict of interest All authors report no conflicts of interest.
Funding Information This investigation was supported in part by the Military Suicide Research Consortium (MSRC), an effort supported by the Office of the Assistant Secretary of Defense for Health Affairs under Award no. (W81XWH-10-2-0181). Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the MSRC or the Department of Defense.
References Allan, N.P., Albanese, B.J., Short, N.A., Raines, A.M., Schmidt, N.B., 2015. Support for the general and specific bifactor model factors of anxiety sensitivity. Personal. Individ. Differ. 74, 78–83. Allan, N.P., Capron, D.W., Raines, A.M., Schmidt, N.B., 2014. Unique relations among anxiety sensitivity factors and anxiety, depression, and suicidal ideation. J. Anxiety Disord. 28, 266–275. Beck, A.T., Steer, R.A., Brown, G.K., 1996. Beck Depression Inventory Manual, 2nd ed. Psychological Corporation, San Antonio, TX. Bollen, K., 1989. Structural Equation Models. Wiley, New York, NY. Capron, D.W., Cougle, J.R., Ribeiro, J.D., Joiner, T.E., Schmidt, N.B., 2012a. An interactive model of anxiety sensitivity relevant to suicide attempt history and future suicidal ideation. J. Psychiatr. Res. 46, 174–180. Capron, D.W., Fitch, K., Medley, A., Blagg, C., Mallott, M., Joiner, T., 2012b. Role of anxiety sensitivity subfactors in suicidal ideation and suicide attempt history. Depress. Anxiety 29, 195–201. Capron, D.W., Lamis, D. a, Schmidt, N.B., 2014. Test of the depression distress amplification model in young adults with elevated risk of current suicidality. Psychiatry Res. 219, 531–535. Capron, D.W., Norr, A.M., Macatee, R.J., Schmidt, N.B., 2013. Distress tolerance and anxiety sensitivity cognitive concerns: testing the incremental contributions of affect dysregulation constructs on suicidal ideation and suicide attempt. Behav. Ther. 44, 349–358. Chen, F.F., Hayes, A., Carver, C.S., Laurenceau, J.P., Zhang, Z., 2012. Modeling general and specific variance in multifaceted constructs: a comparison of the bifactor model to other approaches. J. Personal. 80, 219–251. Delucchi, K.L., Bostrom, A., 2004. Methods for analysis of skewed data distributions in psychiatric clinical studies: working with many zero values. Am. J. Psychiatry 161, 1159–1168. Ebesutani, C., McLeish, A.C., Luberto, C.M., Young, J., Maack, D.J., 2013. A bifactor model of anxiety sensitivity: analysis of the Anxiety Sensitivity Index-3. J. Psychopathol. Behav. Assess. Grant, D.M., Beck, J.G., Davila, J., 2007. Does anxiety sensitivity predict symptoms of panic, depression, and social anxiety? Behav. Res. Ther. 45, 2247–2255. Grant, S., 2010. What's normal anyway? Normal and non-normal distributions in psychiatry. Acta Neuropsychiatr. 22, 305–310. Hu, L., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct. Equ. Model. Multidiscip. J. 6, 1–55. Joiner, T.E., Rudd, M.D., 1996. Disentangling the interrelations between hopelessness, loneliness, and suicidal ideation. Suicide Life Threat. Behav. 26, 19–26. Katz, C., Yaseen, Z.S., Mojtabai, R., Cohen, L.J., Galynker, I.I., 2011. Panic as an
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
N.P. Allan et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎ independent risk factor for suicide attempt in depressive illness: findings from the National Epidemiological Survey on Alcohol and Related Conditions (NESARC). J. Clin. Psychiatry 72, 1628–1635. Keough, M.E., Schmidt, N.B., 2012. Refinement of a brief anxiety sensitivity reduction intervention. J. Consult. Clin. Psychol. 80, 766–772. Kessler, Ronald C., Borges, G., Walters, E.E., 1999. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch. Gen. Psychiatry 56, 617–626. Kraemer, H.C., Kazdin, A.E., Offord, D.R., Kessler, R.C., Jensen, P.S., Kupfer, D.J., 1997. Coming to terms with the terms of risk. Arch. Gen. Psychiatry 54, 337–343. McDonald, R.P., 1999. Test theory: A unified treatment. Erlbaum, Mahwah, NJ. Metalsky, G.I., Joiner, T.E., 1997. The hopelessness depression symptom questionnaire. Cogn. Ther. Res. 21, 359–384. Mohlman, J., Zinbarg, R.E., 2000. The structure and correlates of anxiety sensitivity in older adults. Psychol. Assess. 12, 440–446. Moshagen, M., 2012. The model size effect in SEM: inflated goodness-of-fit statistics are due to the size of the covariance matrix. Struct. Equ. Model. Multidiscip. J. 19, 86–98. Mulaik, S., 2007. There is a place for approximate fit in structural equation modelling. Personal. Individ. Differ. 42, 883–891. Muthén, L.K., Muthén, B.O., 1996 2012. Mplus User's Guide, seventh ed. Muthén & Muthén, Los Angeles, CA. Nock, M.K., Hwang, I., Sampson, N.A., Kessler, R.C., 2010. Mental disorders, comorbidity and suicidal behavior: results from the National Comorbidity Survey Replication. Mol. Psychiatry 15, 868–876. Oglesby, M.E., Capron, D.W., Raines, A.M., Schmidt, N., 2015. Anxiety sensitivity cognitive concerns predict suicide risk. Psychiatry Res. 226, 252–256. Olatunji, B.O., Wolitzky-Taylor, K.B., 2009. Anxiety sensitivity and the anxiety disorders: a meta-analytic review and synthesis. Psychol. Bull. 135, 974–999. Osman, A., Bagge, C.L., Gutierrez, P.M., Konick, L.C., Kopper, B.A., Barrios, F.X., 2001. The Suicidal Behaviors Questionnaire-Revised (SBQ-R): validation with clinical and nonclinical samples. Assessment 8, 443–454. Osman, A., Gutierrez, P.M., Smith, K., Fang, Q., Lozano, G., Devine, A., 2010. The Anxiety Sensitivity Index-3: analyses of dimensions, reliability estimates, and correlates in nonclinical samples. J. Personal. Assess. 92, 45–52. Reise, S.P., 2012. Invited paper: the rediscovery of bifactor measurement models. Multivar. Behav. Res. 47, 667–696. Reiss, S., McNally, R.J., 1985. Expectancy model of fear. In: Theoretical Issues in Behavior Therapy. Academic Press, San Diego, CA, pp. 107–121. Reiss, S., Peterson, R.A., Gursky, D.M., McNally, R.J., 1986. Anxiety sensitivity, anxiety frequency and the prediction of fearfulness. Behav. Res. Ther. 24, 1–8. Rhemtulla, M., Brosseau-Liard, P.É., Savalei, V., 2012. When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychol. Methods 17,
7
354–373. Schmidt, N.B., Capron, D.W., Raines, A.M., Allan, N.P., 2014. Randomized clinical trial evaluating the efficacy of a brief intervention targeting anxiety sensitivity cognitive concerns. J. Consult. Clin. Psychol. Schmidt, N.B., Eggleston, A.M., Woolaway-Bickel, K., Fitzpatrick, K.K., Vasey, M.W., Richey, J.A., 2007. Anxiety Sensitivity Amelioration Training (ASAT): a longitudinal primary prevention program targeting cognitive vulnerability. J. Anxiety Disord. 21, 302–319. Schmidt, N.B., Mallott, M., 2006. Evaluating anxiety sensitivity and other fundamental sensitivities predicting anxiety symptoms and fearful responding to a biological challenge. Behav. Res. Ther. 44, 1681–1688. Schmidt, N.B., Zvolensky, M.J., Maner, J.K., 2006. Anxiety sensitivity: prospective prediction of panic attacks and Axis I pathology. J. Psychiatr. Res. 40, 691–699. Schmidt, Norman B., Woolaway-Bickel, K., Bates, M., 2001. Evaluating panic-specific factors in the relationship between suicide and panic disorder. Behav. Res. Ther. 39, 635–649. Sehlmeyer, C., Konrad, C., Zwitserlood, P., Arolt, V., Falkenstein, M., Beste, C., 2010. ERP indices for response inhibition are related to anxiety-related personality traits. Neuropsychologia 48, 2488–2495. Shneidman, E.S., 1993. Suicide as psychache. J. Nerv. Ment. Dis. 181, 145–147. Souminen, K., Isometsӓ, E., Suokas, J., Haukka, J., Achte, K., Lönnqvist, J., 2004. Completed suicide after a suicide attempt: a 37-year follow-up study. Am. J. Psychiatry 161, 562–563. Taylor, S., Zvolensky, M.J., Cox, B.J., Deacon, B., Heimberg, R.G., Ledley, D.R., Abramowitz, J.S., Holaway, R.M., Sandin, B., Stewart, S.H., Coles, M., Eng, W., Daly, E. S., Arrindell, W.A., Bouvard, M., Cardenas, S.J., 2007. Robust dimensions of anxiety sensitivity: Development and initial validation of the Anxiety Sensitivity Index-3. Psychol. Assess. 19, 176–188. Timpano, Kiara R., Schmidt, N.B., 2012. The relationship between self-control deficits and hoarding: A multimethod investigation across three samples. J. Abnorm. Psychol. 122, 13–25. 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. J. Consult. Clin. Psychol. 76, 72–83. Weissman, M.M., Klerman, G.L., Markowitz, J.S., Oullette, R., 1989. Suicidal ideation and suicide attempts in panic disorder and attacks. N. Engl. J. Med. 321, 1209–1214. World Health Organization, 2012. Public health action for the prevention of suicide: a framework. Zavos, H.M.S., Rijsdijk, F.V., Eley, T.C., 2012. A longitudinal, genetically informative, study of associations between anxiety sensitivity, anxiety and depression. Behav. Genet. 42, 592–602.
Please cite this article as: Allan, N.P., et al., Examining the unique relations between anxiety sensitivity factors and suicidal ideation and past suicide attempts. Psychiatry Research (2015), http://dx.doi.org/10.1016/j.psychres.2015.05.066i
