Psychiatry Research 227 (2015) 302–308

Contents lists available at ScienceDirect

Psychiatry Research journal homepage: www.elsevier.com/locate/psychres

An empirical investigation of suicide schemas in individuals with Posttraumatic Stress Disorder Maria Panagioti a,n, Patricia A. Gooding b, Daniel Pratt b, Nicholas Tarrier c a

Institute of Population Health, Centre for Primary Care, University of Manchester, UK School of Psychological Sciences, University of Manchester, UK c Department of Psychology, Institute of Psychiatry, Kings College London, UK b

art ic l e i nf o

a b s t r a c t

Article history: Received 20 March 2014 Received in revised form 30 January 2015 Accepted 21 February 2015 Available online 28 March 2015

Posttraumatic Stress Disorder (PTSD) has been strongly associated with suicidality. Despite the growing evidence suggesting that suicidality is heightened by the presence of an elaborated suicide schema, investigations of suicide schemas are sparse. Using novel methodologies, this study aimed to compare the suicide schema of PTSD individuals with and without suicidal ideation in the past year. Fifty-six participants with a diagnosis of PTSD (confirmed via the Clinician Administered PTSD Scale) completed questionnaires to assess suicidality, depressive severity and hopelessness. A series of direct and indirect cognitive tasks were used to assess suicide schemas. The pathfinder technique was employed to construct graphical representations of the groups' suicide schemas. The suicidal group reported significantly more severe PTSD symptoms, depressive symptoms, hopelessness and suicidality. The suicide schema of the suicidal group was significantly more extensive compared to the non-suicidal group even after taking into account in the analyses group differences in clinical measures. Moreover, the suicide schemas of the two groups were qualitatively distinct from each other. These findings provide support for contemporary theories of suicide which view suicide schemas as an important indicator of suicide risk. The investigation of schema constructs opens a new avenue of research for understanding suicide. & 2015 Elsevier Ireland Ltd. All rights reserved.

Keywords: Schematic representations PTSD Suicidality Depressive symptoms Hopelessness

1. Introduction People diagnosed with Posttraumatic Stress Disorder (PTSD) often report suicidality including suicidal ideation and suicide attempts (Panagioti et al., 2009, 2012c; Krysinska and Lester, 2010). Recent research suggests that more than 50% of individuals with PTSD experience suicidal ideation and between 20% and 30% attempt suicide (Tarrier and Gregg, 2004; Bernal et al., 2007; Sareen et al., 2007; Panagioti et al., 2012a). Although one of the most widely recognized strategies for preventing future suicides is the early identification and treatment of those who are at the highest risk for suicide, this effort is hampered by the lack of theory-derived and empirically testable models of suicidality (Bolton et al., 2007; Panagioti et al., 2009). In an attempt to overcome this limitation, our research team has proposed the Schematic Appraisal Model of Suicide (SAMS), which is a reconceptualization of Williams' Cry of Pain Model of suicide (CoP) (Williams, 1997; Williams et al., 2005; Johnson et al., 2008). The n Correspondence to: Centre for Primary Care, Institute of Population Health, Williamson Building, Oxford Road, University of Manchester, M13 9PL, UK. Tel.: þ 44 161 306 0665. E-mail address: [email protected] (M. Panagioti).

http://dx.doi.org/10.1016/j.psychres.2015.02.019 0165-1781/& 2015 Elsevier Ireland Ltd. All rights reserved.

SAMS emphasizes the role of two cognitive structures in the development and maintenance of suicidal behaviors, namely, a negatively biased subjective appraisal system and a suicide schema network (Johnson et al., 2008; Pratt et al., 2010). Both, the negative appraisal system and the suicide schema are viewed as dynamic processes which constantly interact and strengthen each other (Johnson et al., 2008; Pratt et al., 2010; Taylor et al., 2011). Suicidal thoughts and acts are thought to emerge as a means of escape from the experience of severe feelings of defeat and entrapment caused by the constant maladaptive interaction of the appraisal system and suicide schema (Taylor et al., 2010a, 2010b; Panagioti et al., 2013). For example, the appraisal of common stressors, such as, negative social interactions, psychiatric symptoms, and personal characteristics in terms of defeat and entrapment activate and strengthen suicide schema which in turn generates more rigid and difficult to overcome negative appraisals (Johnson et al., 2008). Our empirical work so far has focused on the investigation of different aspects of the function of the appraisal system. Our appraisal-related studies which were conducted in a range of clinical and non-clinical populations (i.e., individuals with psychosis, PTSD, students and para-suicidal individuals) have provided strong support to the SAMS model (Taylor et al., 2010; Johnson et al., 2011; Panagioti et al., 2012b; Panagioti et al., 2014).

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

Despite the progress achieved in the investigation of the appraisal system, the role of suicide schema is under-investigated mainly because the empirical test of “schema” constructs are notoriously difficult (Teasdale and Dent, 1987; Bower and Forgas, 2001; Johnson et al., 2008; Pratt et al., 2010). Suicide schema is defined as a loose network of interconnecting stimulus, response, and emotional information which when activated, triggers thoughts of suicide as a means of escape from defeating and entrapping states (Bower and Forgas, 2001; Johnson et al., 2008). It is assumed that the suicide schema is strengthened each time it is activated, and such activation increases the potential to incorporate a wide range of elements into the schema network, such as, psychiatric symptoms and related emotional and cognitive states (Johnson et al., 2008; Pratt et al., 2010). Based on the differential activation model and the suggestion that some people are more susceptible to experience networks of selfreferent negative thoughts as a consequence of small mood changes, one tentative hypothesis is that the extensiveness and elaboration of the suicide schema will differ from individual to individual (Teasdale and Dent, 1987). Using a novel approach, Pratt and colleagues conducted the first empirical test of the direct and indirect features of the suicide schema in individuals with psychoses. This study showed that individuals with a history of suicide attempts had more extensive suicide schemas compared to individuals without a history of suicide attempts even after adjusting for the effects of comorbid depression, anxiety and hopelessness (Pratt et al., 2010). The overarching aim of this study was to conduct the first empirical investigation of the suicide schema in individuals with PTSD. On the grounds of theoretical suggestions and empirical evidence from psychoses (Johnson et al., 2008; Pratt et al., 2010), it was hypothesized that individuals with PTSD who experienced suicidal ideation in the past year would generate more extensive and elaborated suicide schema networks compared to those who did not experience suicidal ideation in the past year. Consistent with the Pratt et al.'s (2010) paradigm, we also examined whether the differences in the groups' suicide schemas remained after adjusting for the effects of PTSD symptom severity, depressive symptom severity, hopelessness and lifetime suicidality.

2. Methods 2.1. Participants Participants were recruited using adverts (i.e., newspaper advertising, online advertising in the University of Manchester [UK], posters in mental health services based in Manchester, such as, Victim Support and the Rape Crisis Center) asking for people who had experienced a traumatic event (i.e., crime, physical threat, serious accident, military combat, natural disaster, terrorist attack, diagnosed with a lifethreatening illness) in the past, and have been affected by it, to volunteer. Potential participants were sent by post or email a self-report measure, the Posttraumatic Stress Diagnostic Scale (PDS) (Foa et al., 1997) to assess whether they met the inclusion criteria for the study. Those participants who returned the PDS scale and met the inclusion criteria of the study proceeded to the full assessment. Potential participants had to fulfill the following inclusion criteria to be included in the study: (1) they had to have experienced a serious traumatic event and meet criterion A1 of the PDS (Foa et al., 1997); (2) be aged between 18 and 65 years; (3) fulfill the criteria for a lifetime diagnosis of PTSD confirmed by the Clinical Administrated PTSD scale (CAPS) for DSM IV (Blake et al., 1995); (4) have experienced at least one PTSD symptom in past month with Z1 frequency and Z 2 intensity scores determined by the CAPS; and (5) have a thorough grasp of the

1 The criterion A of the PDS scale consists of four questions which assess if the person experienced or witnessed an event that involved actual or threatened death or serious injury or a threat to the physical integrity of self or others and if the person felt intense fear, helplessness, or horror as a consequence of that event. This inclusion criterion was used in order to ensure that all the prospective participants had been exposed to a traumatic experience which was severe enough to meet the criterion A of the PDS scale.

303

English language (this was necessary for participation in the assessment interview and for the understanding of the questionnaire items). In addition, participants had to provide informed consent and be willing to come into the University of Manchester to carry out the study. Participants were excluded if they suffered from dementia, organic brain disorder or an active psychotic disorder. Two individuals were excluded because they reported suffering active psychosis. No participant reported suffering from dementia or organic brain disorder. All the participants were already in contact with mental health services (either they were previously/currently receiving treatment or were placed in the waiting list). 2.2. Assessments and measures 2.2.1. Clinician administered PTSD scale The CAPS (Blake et al., 1995) was used to confirm a PTSD diagnosis or to assess the number and severity of PTSD symptoms. The total CAPS severity score was computed by adding the intensity and frequency scores for each of the PTSD symptoms. A current or lifetime diagnosis of PTSD was assigned according to guidance (i.e., item frequency¼1 and intensity¼ 2, for at least one PTSD criterion B, three C, and two D symptoms, and total severity 465) (Weathers et al., 1999). Previous research has found that the Cronbach's alpha coefficient ranges from 0.85 to 0.87 for the three symptom clusters and 0.94 for the total CAPS score (Blake et al., 1995). The alpha coefficient for the total CAPS severity score was 0.93 in this sample. 2.2.2. Beck depression inventory II The BDI-II (Beck et al., 1996a) comprises 21 items which measure the severity of depressive symptoms (range 0–63) in the past two weeks. The BDI has high internal consistency (Cronbach's alpha coefficient of 0.86 for psychiatric patients and 0.81 for non-psychiatric individuals) and concurrent validity with respect to clinical ratings and the Hamilton Psychiatric Rating Scale for Depression (HRSD) for psychiatric (0.72 and 0.73, respectively) and non-psychiatric individuals (0.60 and 0.74 respectively) (Beck et al., 1996a, 1996b; Dozois et al., 1998; Richter et al., 1998). In this sample, the alpha coefficient was 0.94. 2.2.3. Beck hopelessness scale The BHS (Beck et al., 1974) consists of 20 true or false items assessing the prevalence of thoughts and beliefs about feelings of hopelessness in the past week (e.g., “My future seems dark to me”). The scale has been found to have an alpha coefficient of 0.93 (Holden and Fekken, 1988) and in this study it was 0.92. 2.2.4. Suicidal behaviors questionnaire-revised The SBQ-R (Osman et al., 2001) is a four-item measure which assesses the level of suicidality experienced by the participants. The first item measures levels of lifetime suicidality including thoughts and attempts; the second item assesses the frequency of suicidal thoughts in the past year; the third item measures the communication of the intent to commit suicide; and the fourth item assesses the likelihood of committing suicide in the future. The total score ranges from 3 to 18 with higher scores indicating greater levels of suicidality (Osman et al., 2001). The alpha coefficient was 0.87 in the present sample. 2.2.5. Suicide schema tasks Based on the structural approach (Goldsmith et al., 1991), two different types of cognitive tasks (fluency tasks and sort task) were used to measure the content and structure of suicide schema. We have used these tasks previously to measure suicide schema among individuals with psychoses (Pratt et al., 2010). The fluency tasks aimed to assess whether individuals with suicidality differed from those without suicidality in their ability to generate information (Macleod et al., 1993; MacLeod et al., 1997). Participants were informed that they will be provided with a concept and that they have to orally generate as many distinct words, thoughts, feelings or phrases related to this concept within one minute. To ensure that participants understood the task they were provided with an example. Additionally, the Animal Category Fluency Task (ACF) (Pratt et al., 2010) was administered first as a means of familiarizing the participants with the cognitive tasks and ensuring that the two groups were equally competent in generating information. Subsequently, participants were presented with the Suicide Category Fluency Task (SCF) (Pratt et al., 2010) in which they repeated the fluency task using the concept “Suicide” instead of “Animals”. The participants' responses to the concepts of “animals” and “suicide” were recorded in the order they were produced, and the total number of words generated was calculated. The Suicide Category Sort Task (SCT) aimed to complement the suicide fluency task because the latter is dependent on the availability of the components of the schema to both introspection and articulation (Olsen and Rueter, 1987). Participants were given 10 concepts and they were instructed to sort them from the most related to suicide to the least related to suicide. The development and selection process of these 10 concepts and also more details about all the cognitive tasks implemented in this study are described elsewhere (Pratt et al., 2010). Briefly, after reviewing the academic and clinical literature, a questionnaire with 100 concepts with varying degree of relevance to suicide was developed and administered to 12 volunteers who rated the relevance of these 100 concepts to suicide in a 0–3 scale.

304

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

In order to achieve variability, the 10 most related and 10 least related words to suicide were excluded from the list. Therefore, the 10 words for the SCS task comprised of three words randomly selected from those rated 11th to 20th in relatedness, four words from those rated 46th to 55th, and three words from those rated 81st to 90th in relatedness. Participants were allowed up to 5 min to complete the suicide category sort task.2 2.3. Procedure Participants were initially administrated the CAPS and afterwards they completed the self-report measures; BDI, BHS and SBQE in the order detailed. Next, participants were presented with the three cognitive tasks; ACF, SCF and then SCS. The CAPS interview (baseline) and all the self-report questionnaires were administrated by the first author in one session. The first author is a chartered psychologist and had previously received training for administering CAPS. The research session was conducted in a private room and opportunities for questions and breaks were provided. Ethical approval was obtained from the relevant NHS research ethics committee before this study commenced. 2.4. Data analysis Research evidence suggests that a lifetime diagnosis of PTSD and subthreshold PTSD symptoms are associated with comparable levels of distress and suicidality with a current diagnosis of PTSD (Marshall et al., 2001; Zlotnick et al., 2002; Jakupcak et al., 2011). Moreover, a recent study showed that there is a significant discordance between different versions of DSM in assigning a diagnosis of PTSD (Hoge et al., 2014). These findings support the view that focusing on individuals who experience a continuum of psychiatric symptoms might have greater clinical relevance than restricting focus to specific diagnostic groups (McGovern and Turkington, 2001; Broman-Fulks et al., 2006; van Os et al., 2009). On these grounds, individuals with a current diagnosis of PTSD and those with a lifetime diagnosis of PTSD were treated as a single group in the analyses. The data analysis was based on the idea that the order of words named by the participants in the cognitive tasks would be indicative of their semantic distance within the person's cognitive schema (Crowe and Prescott, 2003). Hence, it was anticipated that semantically similar words would be closer to each other in the generated lists compared to semantically dissimilar words (Neely, 1991). A matrix of proximity data was developed from the responses generated by the three cognitive tasks in which the number in each cell represented the distance between two words. Prescott et al.'s (2006) Mean Cumulative Frequency (MCF) metric was then utilized to transform the proximity matrices into an average matrix for the suicidal group and non-suicidal group. Next, the pathfinder technique (Pathfinder for Windows, Version 6.2) (Schvaneveldt, 1990) was used to generate graphical representations of the proximity data obtained from each group. The consistency of the Pathfinder data was assessed by calculating a measure of Coherence which is equivalent to the concept of “internal consistency” ( 40.20 value ¼ lack of consistency and meaningless outcomes; 0.20–0.60 value ¼acceptable; 40.60 ¼ high). Moreover, the pathfinder networks generated by the two groups were compared. Based on Goldsmith et al.'s (1991) recommendation, the correlations of the MCF metric for each pair of items within the group's networks (higher correlations¼higher network similarity) and the Closeness statistic (i.e., the proportion of shared links for matching items across two networks; higher value¼ greater network similarity) were computed to assess network similarity (Goldsmith et al., 1991). Additionally, as recommended by Prescott et al. (2006), network difference was further assessed by performing a multivariate analysis of variance (MANOVA) in which the MCF metric and standard deviation of the MCF metric were the dependent variables and group was the independent variable.

3. Results 3.1. Demographic and clinical characteristics of the sample Overall, 56 individuals completed this study. Of those, two participants were excluded from the analyses because they did not complete any of the three schema tasks. Hence, the analyses were 2 It should be noted that since the material was initially designed to be used in individuals with psychoses, we replaced the “psychoses related concepts” by “trauma/PTSD concepts” such as “flashbacks”, “abuse”, “trauma”. Following this, the revised material was re-administered to 12 volunteers. The relatedness of all the common concepts was equivalent in both tests and, therefore, used the same concepts as in Pratt et al. (2010) with the exception of using “flashbacks” instead of “psychoses” (the relatedness of these two words was also equivalent). This choice allowed a better comparability of the outcomes across the two studies.

based on 54 individuals (mean age¼28.7, S.D. ¼10.8). The sample consisted mainly of white (n¼ 40, 76.9%) and unmarried (n ¼35, 67.3%) women (n ¼42, 80.8%). Thirty-two (61.5%) participants met the CAPS criteria for a current diagnosis of PTSD. The remaining 22 (38.5%) were in the remission phase; they all met the CAPS criteria for a lifetime PTSD diagnosis but currently experienced a range of PTSD symptoms (M¼ 5.48, S.D. ¼3.32; range¼2–9) which did not account for full current diagnosis of PTSD. The average number of PTSD symptoms experienced by all participants were 12.5 (S.D. ¼4.59). All 22 participants experienced at least one symptom associated directly with a current PTSD diagnosis, namely reexperiencing or avoidance symptoms (Rosen et al., 2008). Thirtyfour (71.2%) participants endorsed some type of suicidality, such as, suicidal ideation, plans or attempts in their lifetime, whereas 28 (51.9%) individuals experienced suicidal thoughts in the past year. Table 1 presents the demographic and clinical characteristics of participants who experienced suicidal ideation in the past year based on the second item of the SBQR (suicidal group) compared with participants who did not experience suicidal ideation in the past year (non-suicidal group). As shown, the suicidal group was significantly more likely to have crime/sexual crime as an index trauma and to report more severe PTSD symptoms, depressive symptoms, and hopelessness compared to the non-suicidal group. All the participants who reported lifetime suicide attempts (n ¼9) also reported suicidal ideation in the past year and the vast majority of individuals (21 out of 25) who reported lifetime suicidal ideation/plans also reported suicidal ideation in the past year. The two groups did not differ significantly in terms of demographic characteristics including age, gender, marital status and ethnicity.

3.2. Category fluency task Participants named a total of 253 distinct animal concepts in the ACF task, the most frequent of which were cat and dog named by 53 (96%) and 49 (92%) participants, respectively. No significant difference was found between the mean number of animal concepts generated by the suicidal group (M ¼17.71, S.D. ¼5.87) compared to the non-suicidal group (M ¼16.72, S.D. ¼6.89; t[52] ¼ 0.832, P ¼0.409). This result suggests that both groups were equally cognitively competent to complete cognitive fluency tasks. A total of 484 distinct suicide words or phrases were named by all participants in the SCF task. Death (37%), depression (37%), sadness (37%), loneliness (28%), knife (27%) and pills (27%) were the most commonly named concepts. The suicidal group listed a significantly greater number suicide concepts (M ¼14.56, S.D.¼ 3.92) compared to the non-suicidal group (M¼ 11.00, S.D.¼ 3.79; t[52] ¼3.32, P¼ 0.002). This difference in the mean number of suicide concepts named by the two groups, remained significant (F(1,48) ¼ 3.37, P ¼0.047) even after controlling for statistically significant group differences on clinical characteristics including severity of PTSD symptoms, depressive symptoms, hopelessness and lifetime suicidality. Next, graphical representations of the participants' underlying suicide schema networks were generated using the pathfinder technique. Figs. 1 and 2 present the resulted networks. The degree of coherence was high in the suicidal group data (0.76) and acceptable in non-suicidal group data (0.46) and, therefore, meaningful interpretations can be drawn from the data. The most important difference between the two schematic networks lay in their content/organization. The suicidal group's network primarily consisted of methods/means of suicide whereas the non-suicidal group consisted of concepts related to causes/feelings (depressed, sad, unhappy) and consequences (death) of suicide.

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

305

Table 1 Demographic and clinical characteristics of the sample. Suicidal ideation (suicidal group)

No suicidal ideation (non-suicidal group)

N¼ 28

%

N¼26

%

x2

d.f.

P value

Gender Female Male

23 5

82 18

21 5

81 19

0.17

1

0.897

Ethnicity White Black Asian

21 3 4

75 11 14

19 0 7

74 0 26

3.849

2

0.146

Marital status Single Co-habituating

21 7

75 25

19 7

74 26

0.26

1

0.876

Type of trauma Crime/sexual crime Accident Other

15 7 6

54 25 21

4 15 7

15 58 27

9.29

2

0.010

Time of trauma Less than 1 year ago 1–3 Years ago 3–5 Years ago

9 16 3

32 57 11

9 15 2

35 57 8

0.87

2

0.892

Lifetime suicidality Suicidal ideation/plans Suicide attempts

21 9

75 32

4 0

15 0

10.03 10.87

1 1

o 0.01 o 0.001

Age PTSD severity (CAPS) Depression (BDI-II) Hopelessness (BHS)

M 30.43 55.97 21.61 9.36

(S.D.) (11.85) (21.36) (11.08) (3.76)

M 28.96 29.69 8.00 3.35

(S.D.) (10.37) (8.87) (4.74) (2.17)

t 1.14 4.62 5.79 7.12

d.f. 52 52 52 52

P 0.257 o 0.001 o 0.001 o 0.001

Note. CAPS, Clinician Administered PTSD Scale; BDI-II, Beck Depression Inventory II BHS: Beck Hopelessness Scale.

Fig. 1. Pathfinder network of the SCF items for the non-suicidal group.

The networks of the two groups only shared one common item (depressed) and therefore, the comparison of their similarity/ distinctness by examining the significance of the correlation between the inter-item distances and the Closeness statistic was not possible (as these comparisons can only be applied when the networks share the same items). Similarly, the use of a MANOVA to compare the variability in the inter-item distance measures for pairs of items from the SCF task was considered meaningless because the two networks only shared one item.

Fig. 2. Pathfinder network of the SCF items for the suicidal group.

3.3. Suicide sort task The pathfinder technique was used to generate graphical representations of the SCT data generated by the 54 participants. Figs. 3 and 4 present the resultant networks. A high degree of coherence was found in both suicidal and non-suicidal groups (0.85 and 0.86, respectively). Contrary to the representation generated by the CFT task, the schematic representations

306

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

4. Discussion

Fig. 3. Pathfinder network of the SCT items for the non-suicidal group.

Fig. 4. Pathfinder network of the SCT items for the suicidal group.

generated by the SCT task did not differ across the two study groups. Two clusters can be distinguished in both networks, one consisting of the concepts of suffering, hopeless, self-hate, relieving pain and self-esteem and the second consisted of the concepts of beliefs, sinful, personality, and death. The significant correlation between the inter-item distances (r ¼0.68, P o0.001) and the Closeness statistic (0.39, Po 0.01) obtained by the comparison of the two networks confirmed their similarity. A MANOVA was also performed to compare the variability in the inter-item distance measures for pairs of items from the SCS task. No significant multivariate (F(2, 87) ¼0.858, P ¼0.427) or univariate differences were identified for the MCF metric (suicidal: M¼ 0.467, S.D. ¼0.113; non-suicidal: M ¼0.467, S.D. ¼0.114;F(1,88)¼0.01, P ¼0.998) and the standard deviation of the MCF metric (suicidal: M ¼0.26, S.D. ¼ 0.017; non-suicidal: M¼ 0.24, S.D. ¼ 0.020; F(1,88) ¼1.656, P ¼0.201) further suggested that the two networks did not differ substantially.

The primary aim of this study was to investigate the elaboration of suicide schema in individuals with PTSD who reported suicidal ideation in the past year (suicidal group) compared to individuals with PTSD who did not report suicidal ideation (non-suicidal group) in the past year. Consistent with our initial hypotheses, the suicidal group named significantly more suicide-related words in the SFC task compared to the non-suicidal group. This difference in the number of words generated by the two groups remained significant after controlling for the effects PTSD symptom severity, depressive symptom severity, hopelessness, and lifetime suicidality. These findings suggest that individuals who experience suicidal ideation in the past year have more elaborate suicide schemas compared to individuals who do not engage in suicidality independently of the presence/severity of mental health issues that have been identified as important predictors of suicidality in people with PTSD (Panagioti et al., 2012a,b). Moreover, this study replicated the findings of a similar study conducted in individuals with psychosis (Pratt et al., 2010). In agreement with the views of contemporary theories of suicide such as Cry of Pain and SAMS(Williams, 1997; Williams et al., 2005; Bolton et al., 2007; Johnson et al., 2008; Pratt et al., 2010), the current findings suggest that suicide schema appears to form part of a generic and transdiagnostic mechanism of suicide which is implicated in the initiation and establishment of suicidal thoughts and behaviors as a mechanism of action under stressful situations. The suicide schema networks of the two groups produced by the Pathfinder program, revealed an interesting pattern of findings. Whereas the size and elaboration of the two networks did not differ substantially, their content was fundamentally distinct. In particular, the network of the suicidal group incorporated, almost exclusively, means and aspects of the actual act of suicide whereas the network of the non-suicidal group mainly incorporated emotional concepts related to causes and consequences of suicide. This difference could be explained in the light of the hypothesis that people who think about suicide are more likely to gradually progress towards considering methods of translating their suicidal thoughts into suicide acts (Hawton et al., 1998; Joiner et al., 2003; Galfalvy et al., 2006; Oquendo et al., 2006). The examination of this hypothesis using longitudinal designs could be fruitful future research direction. Within SAMS, suicidal individuals are characterized by a rigid and distorted appraisal system which prevents them from widening their perspective and considering the emotional drivers (feelings loneliness, sadness and helplessness) of their negative thoughts and the irreversible consequences of their suicide act (i.e., death) (Johnson et al., 2008). However, it should be noted that the Pratt et al. (2010) study found significant group differences in terms of the size and elaboration of the Pathfinder networks but failed to exhibit similar group differences in relation to the content of the suicide networks. One reason for this inconsistency is that Pratt and colleagues administered the schema tasks in a highly suicidal sample (approximately two thirds of whom had attempted suicide at least once in their lifetime). They divided their groups according to the presence/absence of a suicide attempt history and, therefore, a significant proportion of the non-suicidal group might have experienced suicidal ideation but never attempted suicide (Pratt et al., 2010). The investigation of the characteristics of the suicide schemas in groups of individuals with varying degrees of suicidality (i.e., no suicidal ideation; suicidal ideation; single suicide attempts; multiple suicide attempts) by a future larger study would be a fruitful approach to further understand the function of suicide schema. Moreover, although the differences between suicidal and non-suicidal individuals on the fluency task remained significant after controlling for a number of key clinical factors in the analyses, the types of affective temperament might be another

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

factor that could exert an important influence on the results. Cyclothymic-depressive temperament has been found to be a key predictor of suicidality (Pompili et al., 2009, 2012). Future studies are encouraged to consider the influence of temperament patterns on the extensiveness and content of suicide schemas. This study has three key limitations. The first limitation relates to the sample size. Since we used a novel approach to analyze our data we based a power calculation on a previously recommended “rule of thumb” which suggests three participants for each item of the constructed model (Kruskal and Wish, 1978; Davison, 1983). Ten concepts were administered in the SCT task and therefore we aimed for 30 participants per group (Paulsen et al., 1996; Prescott et al., 2006). The number of participants in the non-suicide group (n¼26) who provided complete data in the SCT was slightly smaller than the recommended rule and for this reason we recommend the replication of this study using a larger sample size. A second limitation is that this study failed to identify any significant group differences in the SCT task. Similarly to this study, Pratt et al. (2010) also failed to identify differences in the SCT task among individuals with psychosis. The adoption of some key methodological amendments by future studies, such as, the use of a larger sample size and a greater number of items (15–20 items) would help to ascertain whether the insignificant SCT findings were a true finding. Third, the statistical comparison of the two networks resulting from the SCF task using Pathfinder was not possible because the application of this technique requires that the two networks share the same items (Prescott et al., 2006). Nevertheless, the dissimilarity and incomparability of the networks produced by the two groups provides support to the hypothesis that individuals with PTSD who report suicidal ideation in the past year differ radically from individuals with PTSD who do not report suicidal ideation in the way that they conceptualize suicide. In conclusion, this is the second study to examine suicide schemas in the literature which underscores the novel contribution that this study makes. Our findings suggest that the presence of an elaborate suicide schema differentiates suicidal from nonsuicidal populations with PTSD. These findings lend support to contemporary models of suicide that emphasize the importance of suicide schema towards understanding suicide. Although encouraging, the findings of this study are preliminary, and therefore, further research into suicide schemas is recommended. References Beck, A.T., Steer, R.A., Ball, R., Ranieri, W.F., 1996a. Comparison of Beck Depression Inventories-IA and -II in psychiatric outpatients. Journal of Personality Assessment 67, 588–597. Beck, A.T., Steer, R.A., Brown, G.K., 1996b. Manual for Beck Depression Inventory-II. Psychological Corporation, San Antonio, TX. Beck, A.T., Weissman, A., Lester, D., Trexler, L., 1974. The measurement of pessimism: the Hopelessness Scale. Journal of Consulting and Clinical Psychology 42, 861–865. Bernal, M., Haro, J.M., Bernert, S., Brugha, T., de Graaf, R., Bruffaerts, R., Lépine, J.P., de Girolamo, G., Vilagut, G., Gasquet, I., Torres, J.V., Kovess, V., Heider, D., Neeleman, J., Kessler, R., Alonso, J., 2007. Risk factors for suicidality in Europe: results from the ESEMED study. Journal of Affective Disorders 101, 27–34. Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Gusman, F.D., Charney, D.S., Keane, T.M., 1995. The development of a clinician-administered PTSD scale. Journal of Traumatic Stress 8, 75–90. Bolton, C., Gooding, P., Kapur, N., Barrowclough, C., Tarrier, N., 2007. Developing psychological perspectives of suicidal behaviour and risk in people with a diagnosis of schizophrenia: we know they kill themselves but do we understand why? Clinical Psychology Review 27, 511–536. Bower, G.H., Forgas, J.P., 2001. Mood and social memory. Handbook of Affect and Social Cognition, 95–120. Broman-Fulks, J.J., Ruggiero, K.J., Green, B.A., Kilpatrick, D.G., Danielson, C.K., Resnick, H.S., Saunders, B.E., 2006. Taxometric investigation of PTSD: data from two nationally representative samples. Behavior Therapy 37, 364–380. Crowe, S.J., Prescott, T.J., 2003. Continuity and change in the development of category structure: insights from the semantic fluency task. International Journal of Behavioral Development 27, 467–479. Davison, M.L., 1983. Multidimensional Scaling. Wiley, New York.

307

Dozois, D.J.A., Dobson, K.S., Ahnberg, J.L., 1998. A psychometric evaluation of the Beck Depression Inventory-II. Psychological Assessment 10, 83–89. Foa, E.B., Cashman, L., Jaycox, L., Perry, K., 1997. The validation of a self-report measure of posttraumatic stress disorder: the Posttraumatic Diagnostic Scale. Psychological Assessment 9, 445–451. Galfalvy, H., Oquendo, M.A., Carballo, J.J., Sher, L., Grunebaum, M.F., Burke, A., Mann, J.J., 2006. Clinical predictors of suicidal acts after major depression in bipolar disorder: a prospective study. Bipolar Disorders 8, 586–595. Goldsmith, T.E., Johnson, P.J., Acton, W.H., 1991. Assessing structural knowledge. Journal of Educational Psychology 83, 88–96. Hawton, K., Arensman, E., Wasserman, D., Hulten, A., Bille-Brahe, U., Bjerke, T., Crepet, P., Deisenhammer, E., Kerkhof, A., De Leo, D., Michel, K., Ostamo, A., Philippe, A., Querejeta, I., Salander-Renberg, E., Schmidtke, A., Temesvary, B., 1998. Relation between attempted suicide and suicide rates among young people in Europe. Journal of Epidemiology and Community Health 52, 191–194. Hoge, C.W., Riviere, L.A., Wilk, J.E., Herrell, R.K., Weathers, F.W., 2014. The prevalence of post-traumatic stress disorder (PTSD) in US combat soldiers: a head-to-head comparison of DSM-5 versus DSM-IV-TR symptom criteria with the PTSD checklist. The Lancet Psychiatry 1, 269–277. Holden, R.R., Fekken, G.C., 1988. Test–retest reliability of the Hopelessness Scale and its items in a University population. Journal of Clinical Psychology 44, 40–43. Jakupcak, M., Hoerster, K.D., Varra, A., Vannoy, S., Felker, B., Hunt, S., 2011. Hopelessness and Suicidal ideation in Iraq and Afghanistan War Veterans reporting subthreshold and threshold posttraumatic stress disorder. Journal of Nervous and Mental Disease 199, 272–275. Johnson, J., Gooding, P., Tarrier, N., 2008. Suicide risk in schizophrenia: explanatory models and clinical implications, the schematic appraisal model of suicide (SAMS). Psychology and Psychotherapy – Theory Research and Practice 81, 55–77. Johnson, J., Gooding, P.A., Wood, A.M., Taylor, P.J., Tarrier, N., 2011. Trait reappraisal amplifies subjective defeat, sadness, and negative affect in response to failure versus success in nonclinical and psychosis populations. Journal of Abnormal Psychology 120, 922–934. Joiner Jr., T.E., Steer, R.A., Brown, G., Beck, A.T., Pettit, J.W., Rudd, M.D., 2003. Worstpoint suicidal plans: a dimension of suicidality predictive of past suicide attempts and eventual death by suicide. Behavaviour Research and Therapy 41, 1469–1480. Kruskal, J.B., Wish, M., 1978. Multidimensional Scaling. Sage, Newberry Park, CA. Krysinska, K., Lester, D., 2010. Post-traumatic stress disorder and suicide risk: a systematic review. Archives of Suicide Research 14, 1–23. MacLeod, A.K., Pankhania, B., Lee, M., Mitchell, D., 1997. Parasuicide, depression and the anticipation of positive and negative future experiences. Psychological Medicine 27, 973–977. Macleod, A.K., Rose, G.S., Williams, J.M.G., 1993. Components of hopelessness about the future in parasuicide. Cognitive Therapy and Research 17, 441–455. Marshall, R.D., Olfson, M., Hellman, F., Blanco, C., Guardino, M., Struening, E.L., 2001. Comorbidity, impairment, in subthreshold and suicidality subthreshold PTSD. American Journal of Psychiatry 158, 1467–1473. McGovern, J., Turkington, D., 2001. ‘Seeing the wood from the trees’: a continuum model of psychopathology advocating cognitive behaviour therapy for schizophrenia. Clinical Psychology and Psychotherapy 8, 149–175. Neely, J.H., 1991. Semantic priming effects in visual word recognition: a selective review of current findings and theories. In: Besner, D., Humphreys, G.W. (Eds.), Basic Processes in Reading: Visual Word Recognition. Lawrence Erlbaum Associates, Hillsdale, NJ, pp. 264–336. Olsen, J.R., Rueter, H.H., 1987. Extracting expertise from experts: methods for knowledge acquisition. Expert Systems 4, 152–168. Oquendo, M.A., Currier, D., Mann, J.J., 2006. Prospective studies of suicidal behavior in major depressive and bipolar disorders: what is the evidence for predictive risk factors? Acta Psychiatrica Scandinavica 114, 151–158. 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. Panagioti, M., Gooding, P., Tarrier, N., 2009. Post-traumatic stress disorder and suicidal behavior: a narrative review. Clinical Psychology Review 29, 471–482. Panagioti, M., Gooding, P., Taylor, P.J., Tarrier, N., 2013. A model of suicidal behavior in posttraumatic stress disorder (PTSD): the mediating role of defeat and entrapment. Psychiatry Research 209, 55–59. Panagioti, M., Gooding, P.A., Tarrier, N., 2012a. An empirical investigation of the effectiveness of the broad-minded affective coping procedure (BMAC) to boost mood among individuals with posttraumatic stress disorder (PTSD). Behaviour Research and Therapy 50, 589–595. Panagioti, M., Gooding, P.A., Tarrier, N., 2012b. Hopelessness, defeat, and entrapment in posttraumatic stress disorder their association with suicidal behavior and severity of depression. Journal of Nervous and Mental Disease 200, 676–683. Panagioti, M., Gooding, P., Taylor, P., Tarrier, N., 2012c. Negative self-appraisals and suicidal behavior among trauma victims experiencing PTSD symptoms: the mediating role of defeat and entrapment. Depression and Anxiety 29, 187–194. Panagioti, M., Gooding, P.A., Taylor, P.J., Tarrier, N., 2014. Perceived social support buffers the impact of PTSD symptoms on suicidal behavior: implications into suicide resilience research. Comprehensive Psychiatry 55, 104–112. Paulsen, J.S., Romero, R., Chan, A., Davis, A.V., Heaton, R.K., Jeste, D.V., 1996. Impairment of the semantic network in schizophrenia. Psychiatry Research 63, 109–121.

308

M. Panagioti et al. / Psychiatry Research 227 (2015) 302–308

Pompili, M., Innamorati, M., Rihmer, Z., Gonda, X., Serafini, G., Akiskal, H., Amore, M., Niolu, C., Sher, L., Tatarelli, R., Perugi, G., Girardi, P., 2012. Cyclothymic-depressiveanxious temperament pattern is related to suicide risk in 346 patients with major mood disorders. Journal of Affective Disorders 136, 405–411. Pompili, M., Rihmer, Z., Innamorati, M., Lester, D., Girardi, P., Tatarelli, R., 2009. Assessment and treatment of suicide risk in bipolar disorders. Expert Review of Neurotherapeutics 9, 109–136. Pratt, D., Gooding, P., Johnson, J., Taylor, P., Tarrier, N., 2010. Suicide schemas in nonaffective psychosis: an empirical investigation. Behaviour Research and Therapy 48, 1211–1220. Prescott, T.J., Newton, L.D., Mir, N.U., Woodruff, P.W.R., Parks, R.W., 2006. A new dissimilarity measure for finding semantic structure in category fluency data with implications for understanding memory organization in schizophrenia. Neuropsychology 20, 685–699. Richter, P., Werner, J., Heerlein, A., Kraus, A., Sauer, H., 1998. On the validity of the Beck Depression Inventory. Psychopathology 31, 160–168. Rosen, G.M., Spitzer, R.L., McHugh, P.R., 2008. Problems with the post-traumatic stress disorder diagnosis and its future in DSM–V. The British Journal of Psychiatry 192, 3–4. Sareen, J., Cox, B.J., Stein, M.B., Afifi, T.O., Fleet, C., Asmundson, G.J.G., 2007. Physical and mental comorbidity, disability, and suicidal behavior associated with posttraumatic stress disorder in a large community sample. Psychosomatic Medicine 69, 242–248. Schvaneveldt, R.W., 1990. Pathfinder Associative Networks: Studies in Knowledge Organization. Ablex, Norwood, NJ. Tarrier, N., Gregg, L., 2004. Suicide risk in civilian PTSD patients. Social Psychiatry and Psychiatric Epidemiology 39, 655–661.

Taylor, P.J., Gooding, P., Wood, A.M., Tarrier, N., 2011. The role of defeat and entrapment in depression, anxiety, and suicide. Psychological Bulletin 137, 391–420. Taylor, P.J., Gooding, P.A., Wood, A.M., Johnson, J., Pratt, D., Tarrier, N., 2010a. Defeat and entrapment in schizophrenia: the relationship with suicidal ideation and positive psychotic symptoms. Psychiatry Research 178, 244–248. Taylor, P.J., Wood, A.M., Gooding, P., Tarrier, N., 2010b. Appraisals and suicidality: the mediating role of defeat and entrapment. Archives of Suicide Research 14, 236–247. Teasdale, J.D., Dent, J., 1987. Cognitive vulnerability to depression – an investigation of 2 hypotheses. British Journal of Clinical Psychology 26, 113–126. van Os, J., Linscott, R.J., Myin-Germeys, I., Delespaul, P., Krabbendam, L., 2009. A systematic review and meta-analysis of the psychosis continuum: evidence for a psychosis proneness-persistence-impairment model of psychotic disorder. Psychological Medicine 39, 179–195. Weathers, F.W., Ruscio, A.M., Keane, T.M., 1999. Psychometric properties of nine scoring rules for the Clinician-Administered Posttraumatic Stress Disorder Scale. Psychological Assessment 11, 124–133. Williams, J.M.G., 1997. Cry of Pain. Penguin, London. Williams, J.M.G., Crane, C., Barnhofer, T., Duggan, D., 2005. Psychology and suicidal behaviour: elaborating the entrapment model. In: Hawton, K. (Ed.), Prevention and Treatment of Suicidal Behaviour: From Science to Practice. Oxford University Press, Oxford, pp. 71–89. Zlotnick, C., Franklin, C.L., Zimmerman, M., 2002. Does “subthreshold” posttraumatic stress disorder have any clinical relevance? Comprehensive Psychiatry 43, 413–419.