Psychological models which explain suicidal thoughts and behaviours

Suicide is included in the ten leading causes of death across different age groups (Bertolote & Fleischmann, 2005) and in the past 45 years, rates have shown a 60% increment, and it is now thought to account for around 1 million deaths worldwide per year (World Health Organization, 2009). A large epidemiological community study in Europe has found that there is a 7.8% lifetime prevalence of suicidal ideation (Bernal et al., 2007).

Sociodemographic factors have been identified as significant predictors of suicidal behaviour including female gender, being of younger age, being divorced or widowed and having low self-esteem (Bernal et al., 2007; Reasoner, 2000). Some work indicates that these predictors apply to both Western and Asian cultures (Thanh, et al., 2006). Social learning theories have been proposed that complete suicide is more common in societies where condemnation of suicide is low (Lester, 1988). Durkheim (1951) hypothesises that this factor may account for the low suicide rates among Catholic, Greek Orthodox and Jewish populations. Linehan (1973) investigated sex differences in society and proposed that if attempted suicide is seen as a “weak” or “feminine” behaviour, then men may be less likely to choose that alternative until emotional dysregulation leads to more lethal attempts. There is a need for greater clarity and detail in relation to individual culture groups (including changes within these groups over a period of time) and to specific religious groups as to the exact reasons for inhibition. The religious corollary also needs attention, that is, when a person who professes to have no formal religious attachments makes decisions on a highly intellectual and apparently objective level.

Psychiatric diagnoses that have been strongly related to suicidality are, major depression, dysthymia, general anxiety disorders (GAD) (Bernal et al., 2007; Weissman et al., 1999) and about 10% have diagnoses of either schizophrenia or substance use disorders (Palmer, Pankratz, & Bostwick, 2005; Rihmer, 2007). Clinical depression is a powerful predictor of suicide attempts and completions (Cheng, Chen, Chen & Jenkins, 2000; Harwitz & Ravizza, 2000). Although considerable research has been carried out into the relationship between depression and suicide, this factor alone is too general to have clinically meaningful predictive value (Drake & Cotton, 1986; Westermeyer, Harrow, & Marengo, 1991; Young et al., 1998). In addition to those psychiatric diagnoses mentioned above, other diagnoses significantly associated with suicide attempts or completions are panic disorder (Friedman, Jones, Chernen & Barlow, 1992), social phobia (Schneier, et al., 1992), borderline personality disorder (Soloff, et al., 2000), antisocial personality disorder (Douglas et al., in press) and gender identity disorder (di Ceglie, 2000). Some of the apparent association of these conditions with suicidal behaviour may be due to their overlap with depression (Cox, Direnfeld, Swinson & Norton, 1994; Hornig & McNally, 1995).

Suicidal behaviour had been traditionally understood within the biomedical-illness model. Consequently, it is unable to see beyond the mental illness tautology, that the suicidal act defines the illness and the illness defines the act. The model is therefore limited as the focus is on the identification of the underlying pathology despite the recognition that pathology alone is not a sufficient explanation for suicidal behaviour (Michel & Valach, 2001; Sheehy & O’Connor, 2002).

For reasons that are unknown, about 5-10% of people who commit suicide have no diagnosable mental disorder at all (Solomon, 2001). It has therefore raised an interest into the predictability and prevention of suicide. A multitude of perspectives and scientific standpoints have been offered to explain the course of suicidal behaviour. Such examples are those that have included sociological, psychiatric, biological and psychological explanations. However, O’Connor & Sheehy, (2000) have suggested that suicidologists have not been successful in predicting and preventing suicide.

Suicide has been theorised as a learned problem-solving technique (Baechler, 1979, 1980; Chiles & Stroshal, 1995; DeCatanzaro, 1981). Baechler’s early model of suicide-as-escape viewed suicide as a mode of problem solving. This model posits that suicidal acts are a learned means of alleviating painful internal states such as feelings of sadness, anxiety, guilt or aversive external problems such as the loss of a loved one. Leenaars reported that escape was a dominant theme found in suicide notes irrespective of nationality or gender (Leenaars, 1999, 2002; O’Connor & Leenaars, 2003) and escape is central to understanding suicidal behaviour (Baumeister, 1990; Williams, 1997, 2001; Williams & Pollock, 2000, 2001).

Baumeister (1990) put forward his theory of Suicide as Escape from Self. He suggests that suicide is the action of escaping from painful self-awareness of certain symbolic interpretations or implications about the self. Baumeister draws upon Baechler’s early model of suicide-as-escape (Baechler, 1979, 1980) but suggests this model was lacking in significant detail, so proposed a new, more elaborate version that consists of six steps. Firstly, there is a severe experience (e.g. interpersonal strife) that results in an individual falling short of their expectations and standards. Secondly, they attribute blame (for this shortfall) internally and this leads to step three, negative self-awareness and feelings of inadequacy, incompetence or guilt, i.e. low self-esteem. The fourth step is to escape the resultant negative affect by attempting to avoid meaningful thought and engage in cognitive deconstruction. This leads to disinhibition, which in turn, increases their willingness to attempt suicide therefore the sixth point is elevating suicide risk. Several studies with college students (O’Connor & O’Connor, 2003; Tassava & Ruderman, 1999) and clinical patients (Dean & Range, 1999; Hunter & O’Connor, 2003) have also produced findings consistent with escape theory. It therefore remains an extremely useful explanatory framework.

Williams broadened the focus beyond escape theory and took account of how theories fit with social, biological and genetic facts (Williams, 1997, 2001; Williams & Pollock, 2000, 2001). ‘Arrested flight’ was a phenomenon (Gilbert, 1989; Gilbert & Allan, 1998) which describes a situation where an animal, e.g. a bird, is defeated but flight is blocked and cannot escape. It is the state of entrapment, where the motivation to take flight is blocked, not the defeat itself, that is dangerous (MacLean, 1990). Williams & Pollock (2000, 2001), Gilbert & Allan, (1998) and Goldstein & Willner, (2002) have argued that there is a similar reaction in humans that could explain depression and suicidal behaviour. It has been suggested that suicidal ideation arises from the feeling of entrapment, but it is when the individual fails to find alternative ways to solve their problems that this can be exacerbated into suicidal behaviour (Williams, Barnhofer, Crane, & Beck, 2005).Gilbert & Allan (1998) investigated the relationship between defeat and entrapment in students and depressed patients and found that both variables explained substantial variance in depression. Furthermore, Goldstein & Willner (2002) investigated the causal relationships between perceptions of defeat, entrapment and depressed mood. They concluded that perceived defeat and entrapment were consequences of depressed mood.

Williams (1997, 2001) and Williams & Pollock (2000, 2001) shifted the focus from depression per se to suicidal behaviour. The ‘Cry of Pain’ model of suicide is founded on an evolutionary approach to understanding suicidal behaviour in depression. Consistent with the ‘arrested flight’ phenomenon, they proposed that suicidal behaviour is reactive and ‘the cry’ to a situation has six components; 1) the presence of stressors; 2) the appraisal of stressors and their consequences, in terms of defeat; 3) information-processing biases, negative memory schema (Pollock & Williams, 2001; Williams & Broadbent, 1986), and deficits in problem-solving strategies (Pollock & Williams, 1998; Schotte & Clum, 1987), all of which may magnify perceptions of entrapment through negative perceptions or responses of self, others, and circumstances; 4) arrested flight resulting in increasingly intractable feelings of entrapment; 5) a perceived absence of rescue factors and concomitant feelings of social isolation (Hewitt, et al., 1998; Hunter & O’Connor, 2003), and 6) imitation models and access to methods of suicide determines whether or not someone acts upon this activation or impulse. When these components are present they can subsequently lead to feelings of hopelessness (Williams & Pollock, 2001). In some studies, hopelessness has been a better predictor of suicide than depression itself (Beck, Brown & Steer, 1989; Wetzel, 1976).

The Cry of Pain model has introduced a number of important concepts. It has a strong empirical basis demonstrating the importance of defeat and entrapment affecting mood and behaviour. It has incorporated psychobiological and evolutionary factors, and has established the importance for the perceptions of defeat and entrapment in mediating stress and depression. For example, Gilbert & Allan (1998) found a strong relationship between self-rated defeat, entrapment, and depression. The Cry of Pain model also emphasizes the potential interactions between emotions and cognitions leading to suicide (Williams, 1997). Recognising that not all depressive patients are suicidal, it implicates the role of entrapment and hopelessness in exacerbating feelings of defeat and depression in the development of suicidality. Finally, the research generated by the Cry of Pain model has not only increased understanding of suicide, but has stimulated exploration of new avenues for treatment and interventions.

On the other hand, a good working model should have clearly defined components that are exclusive and testable. In these areas, the Cry of Pain model may be limited. Firstly, the feeling of defeat has not been fully defined and, for example, it is not clear whether defeat is purely a feeling of failure and giving up, or whether it also comprises some feelings of future entrapment (Gilbert, et al., 2002; Rohde, 2001).

There is also an uncertainty surrounding the third component of the model. Whilst it is sometimes referred to as ‘hopelessness’ (Williams, 1997), other studies suggest that it is actually ‘no rescue’ (Hunter & O’Connor, 2003). Studies researching ‘no rescue’ have therefore measured it as level social support and have shown it to be an important factor in explaining suicide (O’Connor, 2003; Weber, Metha, & Nelsen, 1997). Studies researching the third component as ‘hopelessness’ have measured it using the Beck Hopelessness Scale which claims to measure cognitive, affective, and motivational aspects (Beck, Weissman, Lester, & Trexler, 1974). There will therefore be inconsistencies when using the model.

A third objection concerns a lack of clarity in separating hopelessness, depression, and suicidal ideation (Shahar, Bareket, Rudd, & Joiner, 2006). The Cry of Pain literature has failed to define hopelessness in a manner that separates it cognitively from entrapment. Although entrapment is related to arrested flight in the animal literature or frustrated escape in humans, it is probable to say that hopelessness may develop from the perception of blocked escape. As Van Heeringen et al., (2000) state, the higher the perceived entrapment the higher the level of hopelessness felt by an individual.

A fourth concern is, is hopelessness or no rescue necessary at all. Gilbert (1989) noted that if a defeated and trapped bird were to be rescued and removed from the situation then after time it would eventually recover. Williams & Pollock (2001) state this observation as a reason to believe that there is a third factor which determines how long lasting the reaction is. However, from this, it is reasonable to assume that removal from the situation simply affected the bird’s conditions of entrapment. It was not expecting rescue and once removed it began to recover. Hope and the belief in rescue are not based on actual removal from the defeating situation or the present solution of their problems, but the belief and expectation that they will be rescued. Therefore it has been suggested that the third factor could be conceptualized as ‘relief of the feeling of entrapment’ (Johnson, Gooding & Tarrier, 2008).

The Schematic Appraisal Model of Suicide (SAMS) developed by Johnson, Gooding & Tarrier (2008) is a new approach which builds on previous models. It has three main components; negative information-processing biases, the presence of suicide schema, and an appraisal system.

Firstly, information-processing biases plays a role in maintaining negative emotional states in those who have psychotic symptoms which include misattribution, incorrect inference, catastrophising, misperception of threat, and selective attention (Bentall, et al., 2001; Freeman et al., 2002, 2007; Startup, Freeman, & Garety, 2007). It is suggested that these biases operate in encoding information into memory and in accessing suicide-relevant information.

Secondly, suicidal ideation is the result of the activation of a suicide schema which becomes established as a plan of action for escape. The schema is a loose network of interconnecting stimulus, response, and emotional stored information which when activated at any point will trigger thoughts of suicide as an escape strategy (Christianson & Engelberg, 2006). It is believed that the stronger and more elaborate the suicide schema, the greater the potential to initiate suicide behaviour.

Thirdly, appraisals are valuation judgements that involve attention, interpretation, and inference. Thus, beliefs and attitudes are incorporated within the appraisal system. Appraisal is seen as an important factor because it influences the degree to which events and experience are viewed as stressful, challenging, or an opportunity (Roe et al., 2006).

Instead of describing the pathway into suicidality, the SAMS (Johnson et al., 2008) focuses on the individual’s appraisal system and processes thought to affect this. The SAMS suggests that two types of appraisal are relevant to suicidality. First, it suggests that situation appraisals are important and when stressful events are appraised as defeating and entrapping. Secondly, the model suggests a key role for a self-appraisal construct, which is thought to impact on other cognitive processes. The model suggests that positive self-appraisals may directly impact the situation appraisals system. However, the SAMS does not explain the mechanisms by which positive self appraisals may have their impact. Few studies have directly investigated the way in which negative appraisals affect suicidality (Esposito & Clum, 2002) and even fewer have examined the role of positive appraisals (Esposito & Clum, 2002). Johnson, et al., (2009) aimed to test the SAMS proposal and found that positive self-appraisals interacted with stressful life events, such that those with moderate or high levels of positive self-appraisals did not experience increased suicidality even at elevated levels of stress. The second finding was that broadminded coping, reappraising emotion regulation style and survival and coping related reasons for living did not moderate the impact of stressful life events. These findings support the SAMS, and this is the first study to explore the SAMS concept of the self-appraisal (Johnson, et al., 2009).

The SAMS model is as successful as it avoids the lack of clarity in the current Cry of Pain model. It is consistent with current paradigms of developing cognitive models of psychopathology, and specifically in explaining psychotic disorders and associated problems. It is coherent and testable, and it is more likely to produce clear guidelines for intervention unlike the Cry of Pain model.