The clinical assessment of violence in the context of psychosis: taking a phenomenological stance

Summary

Although most people experiencing psychosis are not violent, a diagnosis of a psychotic disorder is associated with an increased likelihood of violence. Some progress has been made in delineating the nature of this association, but it remains unclear whether specific types of psychotic experience make a specific contribution to the propensity for violence. Just as the phenomenological approach has produced a fuller understanding of psychotic experiences (that can inform improved aetiological and interventional frameworks), the authors assert that such an approach (with its closer attention to the full extent of the patient's subjectivity) has the potential to advance our understanding of the relationship between psychosis and violent behaviour in a way that has clinical applicability. This article examines this potential by overlaying approaches to the phenomenology of psychosis with a framework for the subjectivity of violence to demonstrate how a fuller explanatory formulation for violent behaviour can be derived.

LEARNING OBJECTIVES

After reading this article you will be able to:

• • recognise the importance of adopting a phenomenological approach during the clinical assessment of violence and psychosis

• • appreciate the importance of taking into account the broad nature of altered experiences when assessing violence and psychosis

• • understand key concepts and terminology within phenomenological literature which are relevant to understanding the relationship between violence and psychosis.

Most people with mental illness are not violent and are more commonly the victims of violence (Buchanan 2019). Nevertheless, those with a psychotic illness are more likely than the general population to engage in violent behaviour (Fazel 2009; Whiting 2022), particularly during a first episode (Nielsson 2012). One in five adults with psychosis comes into contact with the criminal justice system (Yee 2020) and almost 40% of all homicides committed by individuals with a psychotic disorder occur before initial treatment (Flynn 2021), suggesting that prompt identification and understanding of the associated risks is a key element to not only early intervention but also long-term management.

Psychosis is a term used to describe impairments in reality testing that may present as brief, isolated episodes or can form part of a life-long condition (Wilson 2022). This article will focus on schizophrenia, a diagnostic entity that relies on the presence of distinct positive and negative symptoms; however, the nature of psychosis cannot fully be appreciated using strict diagnostic criteria alone (Pienkos 2019). Early phenomenological descriptions of schizophrenia had been increasingly cast aside in favour of more structured approaches to diagnosis, notably in the form of the Diagnostic and Statistical Manual of Mental Disorders (DSM) and International Classification of Diseases (ICD) (Andreasen 1997). Eugen Bleuler first described a condition associated with distorted affect, associations, ambivalence and autism in 1911, using the term schizophrenia (Kuhn 2004). He suggested that individuals who appeared to be suffering from this condition shared common difficulties in decision-making, disordered thinking and a preoccupation with their own thoughts as well as markedly reduced emotional responses. Although the more recent burgeoning academic literature devoted to the phenomenological nature of schizophrenia has been very welcome, little attention has yet been given to what this approach has to offer to understanding violence in psychosis.

This article aims to introduce the reader to the potential advantages of adopting a phenomenological approach when working alongside patients who have acted violently during an episode of psychosis or who are considered to have the potential to do so.

Violence and psychotic psychopathology

Attempts have been made to delineate aspects of psychosis that may be specific predictors of violence. Early studies suggested that violence is most commonly associated with delusional beliefs (Taylor 1985) and command hallucinations (McNiel 2000). Although there is robust evidence from multiple studies of a correlation between violence and persecutory delusions (Lamsma 2015), as with many risk factors their prevalence in clinical practice is so high as to limit their utility for clinical decisions in a single instance. However, there has been clinical and academic interest in more specific types of symptom. In particular, threat-control override (TCO) symptoms were postulated to have a unique association with the risk of violence. TCO symptoms comprise delusional beliefs held by the patient of being threatened in some way and losing control of thoughts, feelings and actions to an external agent. Despite initial empirical support for an association between TCO symptoms and violence (Link 1998), subsequent studies are less clear, indicating that there may be a link with the severity of violent acts but not risk of violence overall (Stompe 2004).

Thus, the evidence favouring the role of specific psychotic symptoms in predicting violence is mixed and it is unclear how significant the association is. This may be a consequence both of the complex distal and proximal pathways leading to violence and of a tendency to focus more on experiences that have achieved prominence in diagnostic criteria. Although the use of bounded definitions of aberrant experiences, such as hallucinations and delusions, is an important part of the diagnostic process, these definitions are only one way of describing the patient's psychopathology. Furthermore, even if associations are found in group-level studies, it is still necessary to consider the unique nature of the patient's experiences and their potential role in explaining behaviour (rather than retrofitting an explanatory model to every case).

Before examining the potential explanatory utility of taking a phenomenological approach to the assessment of a patient's mental state at the time of an act of violence, we must emphasise that the explanatory formulation's concern with the altered subjectivity of psychosis should not be at the expense of an interest in other key areas. These include (a) the impact on decisions to act violently of those circumscribed manifestations of psychosis that are given prominence in diagnostic frameworks, (b) aspects of subjectivity relevant to violence that are not necessarily a manifestation of psychosis (Table 1) (Nathan 2020) and (c) wider and more distal contextual factors (Box 1). In relation to the last point, we believe it is particularly important to be mindful of developmental pathways to violence in psychosis. Early life adversity increases vulnerability to later life violence and psychosis (Storvestre 2020). As well as considering how the experience of adversity may have shaped general psychological response patterns or the nature of psychotic experiences, the clinician should also pay heed to the interaction between these processes, i.e. how, on the one hand, the psychological adaptations to early adversity may influence evaluations of aberrant experiences and how, on the other, aberrant experiences may influence the current meaning of previous trauma. In the way that the phenomenological stance discourages the imposition of boundaries (to disaggregate mental experiences and prioritise those components prioritised by diagnostic systems), we stress that the issues raised in this article should not be seen as standing apart from other aspects of understanding the patient.

TABLE 1 Elements of subjectivity proximal to a violent act

BOX 1 Distal contextual factors associated with violence in psychosis

Substance misuse

Homelessness

Unemployment

Relationship problems

Financial difficulties

Social isolation

Early life adversity

Childhood/adolescent conduct problems

Phenomenological approach to understanding violence in psychosis

Background

Although it is important to consider the potential role of commonly recognised types of psychopathology when exploring the relationship between psychosis and violence in clinical practice, focusing only on these experiences will not do justice to the likely complex processes at play. Phenomenological analysis reveals that what are assumed to be discrete types of psychopathology are often not experienced in such a clear-cut way (Broome 2012). Moreover, such analysis highlights the ‘true human relevance’ as well as demonstrating that anomalous experiences are usually not confined to those types of psychopathology that have made it onto the diagnostic criteria lists. There has, in some areas, been a shift in the direction of studies exploring psychosis, with more emphasis on individual psychotic phenomena as opposed to confined diagnosis (Pienkos 2019).

To describe how a wider interest in the patient's subjectivity is likely to inform the clinician's understanding of why they carried out a violent act while subject to psychotic experiences, we have used a pragmatic prototypical model of the time course of psychosis (from the prodrome, through generalised impaired reality testing, and to experiences akin to those ‘psychotic symptoms’ highlighted within diagnostic criteria). However, it is important to recognise that these phases are not distinct and psychotic experiences are not unique to a specific point in the time course.

It is not our intention to present a comprehensive summary of either the literature describing the phenomenology of psychosis or of all the possible explanatory processes that could connect psychotic phenomena with violent actions. We have selected certain types of presentation to provide illustrative examples of how the phenomenological stance can be used in clinical practice when exploring the relationship between aberrant experiences in psychosis and co-occurring violent behaviour. To do this, we have drawn from some key published phenomenological analyses of psychosis and from our clinical experience of working with patients experiencing psychosis who have enacted violence. We then examine the relationship between psychotic phenomena and violent actions through a framework of subjectivity proximal to the violent act (Anderson 2022). The detail of this framework is described elsewhere (Nathan 2020), but in brief it considers the flow of mental processes from a trigger, through attention, meaning, preparedness, evaluation and inhibition/activation, and to the final step of the violent act itself (Table 1).

The prodromal phase

During the prodromal phase, prior to the formation of fixed delusional beliefs, individuals often experience what Karl Jaspers described as an ‘uncanny’ feeling that ‘something suspicious is afoot’ (Bortolotti 2018). The world around them is somehow different, taking on new meaning and there is a change in their relationship with the world which can be described as a delusional atmosphere or delusional mood. This can be a brief or prolonged transition, extending over months or years. Jaspers refers to an uncomfortable tension that invades the individual due to the uncertainty of this experience (Bortolotti 2018).

Thus, phenomenological descriptions of delusional mood include changes in the way individuals attend to the world as well as alterations to the meaning they attribute to the world they encounter. The assessing clinician needs to be mindful that the inherently enigmatic and nebulous qualities of delusional mood (Sul 2022) can make it difficult for the patient to understand this experience and to use everyday language to describe it. Nevertheless, the depth and pervasiveness of the sense of portent may lead the person to behave in uncharacteristic ways, including on occasion in the form of violence. Thus, rather than immediately adopting a line of questioning that seeks to elicit unambiguous descriptions of clearly delineated mental phenomena, the assessor should allow space in the assessment for the patient to openly explore the quality of their own experiences and to describe what they find in their own words. In those cases where there has been violence during a phase of psychosis-related altered experiences, the clinician should then study this phenomenological account for processes that may enrich the explanatory formulation of the violence. For example, given both that a decision to act violently can often be traced back to a particular form of attention to, and interpretation of, environmental cues (Nathan 2020) and that delusional mood is often experienced as an altered form of attending to the environment and attributing meaning to those environmental elements that are attended to, the assessment of this aspect of delusional mood in explaining behaviour may prove particularly informative.

Adhering to the spirit of the phenomenological approach, we would discourage the assessor from approaching the assessment armed with preferred mechanisms that explain violence in psychosis and then trying to reshape the patient's experience to fit one of those mechanisms. However, that is not to say it is unhelpful to have an awareness of the ways subjectivity can change so as to influence behaviour. Assessment of changes in the quality of patients’ experiences can be facilitated by familiarity with the relevant phenomenological literature. There is not space here to present a comprehensive review of different phenomenological models. However, to illustrate how this literature can assist in understanding violent behaviour in the context of generalised changes in subjectivity (such as that included within the descriptive term delusional mood), we have chosen the self-disorder (or ipseity disorder) model. This is an evidence-based conceptualisation of schizophrenic experiences that provides a more fine-grained elucidation of the nature of such changes (Sass 2018). A core component of this model describes how aspects of oneself that are ordinarily experienced implicitly become the focus of explicit attention (hyperreflexivity). Linked to this is the reduction in the subjective sense of self (diminished self-presence).

In examining a patient's account of violence during a phase of psychosis, the clinician should be vigilant to the potential relevance to the index behaviour of changes in the quality of experience and agency associated with ipseity disturbance. For instance, bearing in mind the established association between dissociation and violence (Moskowitz 2004; Daisy 2014), the clinician should be alert to the possibility that the dissociative quality of diminished self-presence (Sass 2018) may be part of the explanatory formulation for an act of violence occurring in the context of delusional mood. The subjective sense of separateness from the past ‘violent self’ can, in our clinical experience, compromise the patient's ability to fully reflect and feel ownership of a past violent act, which in turn may interfere with their insight into the potential for future violence and with their motivation to engage in offence-specific work. Further, given the role of affect in violent behaviour (Nathan 2020), the patient's recollection of any change in affect accompanying the delusional mood should be explored. Although it is often experienced as a tense discomfort, alternatively it can have a positive valence such as excitement (Ritunnano 2022). As well as revealing the nature of the affective change, the patient's account of their subjectivity may offer insights into how this aspect of their subjectivity influenced their actions. Possible mechanisms include responding disproportionately to a discordant encounter, given a background hypervigilance and/or heightened perceived sense of threat. Alternatively, the patient may act in an uncharacteristic way in the hope that it will attenuate the unbearable tension, or even amplify an already excited state.

Loss of insight and impaired reality testing

Hallucinations and delusions are often cited as characteristic examples of the psychotic detachment from reality. However, aside from, and often before, the development of experiences analogous to circumscribed symptom descriptors (such as hallucinations and delusions), a more general loss of contact with the exterior (often accompanied by an ascendancy of the inner world) has been held to be a core feature of schizophrenia (Pienkos 2019). Individuals describe an enduring emotional and mental struggle during the initial development of psychotic symptoms. While contending with this disorientating uncertainty, the person's capacity to negotiate day-to-day occurrences is often compromised. Although this can lead to apparent disinterested introspection, very occasionally the individual may respond with an uncharacteristic outburst of aggression. In this event, the assessing clinician may find it helpful to consider the emotional correlates and their relationship to any aggression. Feelings of fear, shame and guilt have been found to be associated both with the transformation of the experience of reality associated with psychosis (Ritunnano 2022) and with a propensity to violence (Gilligan 2003). Action that appears on initial consideration to be erratic (including, in occasional cases, violent action) may be motivated by the anticipated attenuation of dissonant emotions. Although these dissonant emotions may have been triggered by the experience of reality being disrupted in the context of developing psychosis, the assessor should also consider whether there is a pre-existing sensitivity to these emotions. If, for instance, there has been early life trauma that incited an enduring sensitivity to negative social emotions (especially shame and humiliation), then the sudden and subjectively unexpected and inexplicable appearance of these same emotions during the emergence of psychosis may foster an urge to act to dispel associated feelings of subordination. One way this may be achieved is by dominating others, sometimes with the use of force. Thus, what to an onlooker appears to have been a random outburst directed at an unrelated bystander may have been, from a subjective perspective, an act of desperation to forestall an anticipated state of being overwhelmed by a dysregulated state of mind provoked by a set of emotions that both is a reaction to a dramatic step change in the nature of reality and brings to mind remembered representations of early life trauma.

Social context

Many phenomenological accounts of psychosis include descriptions of changes to the quality of one's existence as a social agent. This includes both the relational quality of specific experiences and a more generalised change in the parts of subjectivity that are ordinarily dependent on real or imagined others. With regard to an example of the former, it would limit the assessor's understanding of a patient's behavioural reactions to the experience of being the subject of surveillance if they only took the assessment up to the point that allows determination of whether or not the experience is consistent with the definition of persecutory delusions. Establishing an assessment context in which the patient feels able to share the emotion-infused nature of this experience will produce a richer understanding of any resulting action. For example, in the event that it has an intensely violating quality (Ratcliffe 2012), then an awareness of this should motivate the clinician to be interested in the ways in which the patient tends to respond to such violation, and whether and how the violation was relevant to any acts of violence when experiencing surveillance.

A more generalised disruption of the sense of connectedness to others may be a corollary of the changes to the basic sense of self (ipseity) described earlier. The capacity for relatedness may be impaired via various mechanisms, such as perceived dissimilarity from others, disruption to corporeal contributions to reciprocity, and the incoherence of mental states (of others in addition to the patient's own) (Nelson 2009). The resulting disconnectedness and lack of sense of belonging in society has the potential to diminish the level of commitment to the moral obligations prevailing in the society in which the patient lives. Given that such social contract dynamics have an inhibitory effect on antisocial urges, then their diminution increases the likelihood of such urges being acted upon. As well as explicit paranoid suspicion of representatives of authority (e.g. police, members of the government and mental health professionals), less authority (or trust) may be invested in such representatives in the first place owing to an indiscriminate detachment from others. Further, understanding how connected the patient is to those around them (and the nature and cause of any disconnectedness) is an important component of relational security in an in-patient forensic psychiatric context.

An individual with psychosis is also at increased risk of actual social isolation, which may not be apparent to them as their attention is consumed by their aberrant experiences. Loss of employment or absence from school or university is often the first indicator of a drop in functioning during or following the prodromal period and can lead to reduced contact with friends and family (Lee 2011). This withdrawal from society reinforces the lack of inherent desire to adhere to social norms as described above and exacerbates feelings of isolation. This separation from their ‘old self’ may have an impact on the normal inhibitory processes associated with regulating behaviour and the need to be perceived as a morally responsible agent by society (Adams 2020).

Psychotic symptoms

Delusions

As already noted, broadening the scope of the assessment to more comprehensively and accurately capture the nature of the patient's experiences should not be a reason to demote the contribution to violent behaviour of experiences that are co-terminus with definitions of psychotic symptoms. However, our contention is that bounding the assessment in accordance with narrow symptom definition boundaries is likely to diminish the quality of the explanatory formulation. An enhanced insight into the relationship between delusions and violence is likely to be achieved if, as well as understanding that delusions are firmly held false beliefs (American Psychiatric Association 2013), the assessor also recognises that delusions are ‘strongly individualised and inherently complex phenomena’ (Ritunnano 2022).

A clinician open to this wider conceptualisation may be helped by drawing on phenomenological frameworks of delusions. A recent systematic review and qualitative evidence synthesis identified a tripartite model comprising (a) radical rearrangement of the lived world dominated by emotions, (b) doubting, losing and finding oneself within delusional realities and (c) searching for meaning, belonging and coherence (Ritunnano 2022). In the case of a patient labouring under a persecutory delusion who attacks a person whom they believe poses a serious threat to them, the nature and interpretation of the belief is likely to be a critical part of the explanation. However, if, as Ritunnano et al's framework suggests can occur, the delusion sits within a wider emotion-infused transformation of reality, then it is likely that such broader changes to the patient's experience would also be an important influence on their actions.

Likewise, the impact on behaviour of a sense of disempowerment or disrupted meaning-making processes (characteristic of the second and third themes of this framework respectively), if present, should be considered when formulating why a violent act occurred. Fuchs’ framing of delusions as an interpersonal phenomenon (Fuchs 2020) highlights the need not only to consider the role of the belief content in explaining the violent urge, but also how perspectival inflexibility and accompanying egocentricity may mean that ordinary inhibitory processes that depend on implicit perspective taking may be attenuated.

Hallucinations

The DSM-5 defines hallucinations as a ‘perception-like experience with the clarity and impact of a true perception but without the external stimulation of the relevant sensory organ’ (American Psychiatric Association 2013). The prominence of such narrow definitions of hallucinations in diagnostic systems may give the impression that they are more important in understanding concurrent behaviour than related aberrant experiences that happen to extend beyond, or fall outside, the definitional limits. To take the example of voice-hearing, the quality of the experience may be quite varied even within the same individual at different times. The phenomenological literature confirms our clinical experience that patients often find it difficult to stand apart from the phenomenon in a way that allows them to articulate its quality in words (Parnas 2023). Whether the experience is closer to a thought-like or a voice-like experience becomes difficult to discern and may change over time. The extent to which the patient's account conforms to specific elements of an auditory hallucination definition may be influenced by the extent to which they take on board the vocabulary offered by the clinician's questions (Parnas 2023).

As with delusions, allowing the patient to reflect more fully on the experience (than would usually be tolerated in a symptom-focused assessment) is not a reason to ignore the part played in influencing action of the explicit meaning of the alien words or thoughts. In patients with a diagnosis of a psychotic mental illness, the meaning often takes the form of instructions to act or of criticism or abuse directed at the patient (McCarthy-Jones 2012). Other aspects of the experience can also affect meaning. For many patients, voices can be threatening and even exert control over them, influencing their emotional responses and causing significant disruption to their daily functioning (Upthegrove 2016). Patients liken auditory verbal hallucinations to overhearing or participating in an actual conversation and often the voices are considered to be powerful or authoritative (Upthegrove 2016) with a distinct expressive tone (Woods 2015). They can be associated with intense fear and distress, particularly when derogatory in nature (Bendall 2006), and it has been suggested that internally driven events may be more distressing than external events, because of the personalised aspect of these experiences (Berry 2013). Deciding to carry out an instruction may be for different reasons, such as because the patient experiences the voice as having some power or control over them, or because they anticipate that a tragedy will befall them or others if they do not (sometimes because this is conveyed explicitly by the voice). Some patients report that they ‘give in’ to voices because they have learned that as a result, they gain temporary relief from them (Leudar 1997). Although in clinical settings the focus can be on whether or not the patient is ‘responding to’ the voice, even if the violent act is not a direct response, voice-like experiences may feature indirectly in the causal chain to action. Account should also be taken of the effect of incessant voices on more general tolerance to frustration. Simple day-to-day tasks become increasingly hard to undertake owing to the constant and interfering nature of auditory verbal hallucinations (Upthegrove 2016). Voices that interrupt conversations have a negative effect on relationships because they give rise to difficulties in understanding what others are saying (Woods 2015).

Conclusions

Although diagnostic symptoms helpfully highlight key experiences that are commonly shared by patients with psychosis, the evidence for a correlation between specific symptoms and violence remains unclear. A more individualised understanding of the nature of the relationship between psychosis and disordered behaviour is likely to be gained by widening the assessment frame to take into account the full extent of the change in subjectivity. The phenomenological approach is an established way of reaching a fuller approximation of the patient's subjectivity. The approach does not dismiss the role of irregular experiences that correspond to symptom definitions. However, the clinician is encouraged not to believe that these experiences necessarily stand apart from the rest of subjectivity or that other aspects of subjectivity are less relevant to understanding behaviour. We have endeavoured to demonstrate the strengths of this approach by examining some phenomenological analyses of psychosis through the lens of subjective processes proximal to violent behaviour. This examination indicates that in the rare cases of psychosis where violence occurs, when assessing the influence of ‘psychotic symptoms’ on the violent behaviour, the resulting formulation is likely to benefit from the clinician extending their focus beyond just the explicit in-the-moment meaning conveyed by the experience (e.g. the nature of the delusional belief or the content of the auditory verbal hallucination) to the more implicit (i.e. emotionally based) meanings. To reach the closest approximation to the patient's subjectivity, the clinician should strive to elucidate the unique meaning to the patient of their altered state of consciousness. Contextualising this extended assessment within a developmental perspective can add further depth to understanding, since the implicit meaning may be linked to mental representations formed in early life. Accepting, as the phenomenological approach does, that it is not just about ‘symptoms’ should also promote interest in broader changes in the way the world is experienced. Such changes, which are often manifest in the way the self and the self/world interface are experienced, may influence processes that facilitate or inhibit aggressive urges and thus the propensity for violence.