• Home
  • >Articles
  • >How can an understanding of Anxiety assist in the treatment of Psychosis

How can an understanding of Anxiety assist in the treatment of Psychosis


CBT has now moved on to consider the area of psychosis. Anxiety related models have proved a useful tool for developing theory and guiding practice. Early stress vulnerability models are considered before going on to consider the relevance of Clark’s panic model to hallucinatory and delusional experience. The role of trauma is considered with particular regard to childhood sexual abuse and early life trauma. New developments in the area of Metacognition and GAD are considered. Psychosis is considered to be a normative process with difficulties arising from the misattribution of intrusions (delusions or hallucinations) in a catastrophizing or unhelpful fashion; use of safety behaviour complicates the picture. Implications from the normalization of psychosis are considered with regards to treatment. Discussion is made of the weaknesses involved with the model regarding its effectiveness and the research base. A positive outlook is cautiously assumed for the rightful utilization of a formulation driven approach of anxiety related models adapted for psychosis.


Cognitive Behavioural Therapy (CBT) has been applied to a wide range of mental health conditions and a growing body of research (Beck, 1993) has supported its value. CBT has now moved onto a consideration of more complex concerns, e.g. personality disordered conditions. Traditionally psychosis was seen as a contraindication for CBT, though Beck (1952) presented an early case. However, increasingly over the last fifteen years a number of researchers have developed CBT for psychotic patients, either for specific symptoms (e.g. Bentall et al, 1994; Chadwick and Birchwood, 1994); as a way of increasing patient’s coping skills (see Tarrier et al, 1993) or as part of a normalizing approach (e.g. Kingdon and Turkington, 1996). All the main approaches are similar in their basics; each takes a similar approach in their dealings with Axis 1 emotional disorders and with psychotic disorders - i.e. they concentrate upon cognitive and information processing and other factors influencing cognitive appraisal such as early experience.

In many respects the treatment of psychosis remains a developing area and will, continue to develop our knowledge and understanding of psychotic disorders and a host of traditional syndromes etc. More recently the emphasis has changed from a concentration on ‘syndromes to cure’ to a consideration of discreet problem areas which affect an individual – such as hallucinations, delusional belief, and paranoia (see Chadwick et al, 1996) and also that individualized CBT might effectively treat the core symptoms of psychosis, rather than just helping individuals cope better with their difficulties (Morrison et al, 2004).

In this essay we shall consider what a CBT understanding of the anxiety disorders has to offer for an understanding of the origins, maintenance and treatment of the positive symptoms of psychosis. We shall consider anxiety models with regard to both auditory hallucinations and to delusional inference. Positive symptoms are conceptualized as intrusions into awareness (e.g. hallucinations) and culturally unacceptable interpretations of intrusions (e.g. delusions). Hallucinations and delusions are common elements of psychosis. For the sake of clarity initially we shall consider each separately (i.e. while conducting an analysis of Clark’s, 1996 model of panic as applied to psychosis); however, we shall combine both elements when considering other models of anxiety and when considering how to work psychologically with delusions and hallucinations. We shall consider a number of related areas such as the relationship of metacognitive theories on worry and the relationship between psychosis and trauma. As psychosis is a complex and at times contradictory area, and perhaps hinges upon our understanding of self (Chadwick, 2006) it would seem there is room for several areas of separate but complimentary development.
Stress vulnerability models.

We start by briefly reviewing early work, which though not stemming strictly from the cognitive literature perhaps set the scene for later developments in CBT. Life stress and anxiety has been implicated in the origins of psychosis for many years. The medical sociologist George Brown (1968) found that sufferers with a recent relapse or first episode of schizophrenia had often experienced a ‘significant life event’ within three weeks of the commencement of psychotic experiences when contrasted to a control group of non-psychotic individuals. Paykel and Cooper (1992) suggested that experiencing a stressful life event (such as moving house, losing a job or a domestic crisis) approximately doubled the likelihood of developing psychosis over the subsequent six months. Zubin and Spring (1977) proposed a stress vulnerability model of psychosis: each individual has varying biological, psychological and social strengths and vulnerabilities for coping with stress and anxiety. A person with a low vulnerability to anxiety can withstand a large amount of stress; however, at times of extreme stress (e.g. solitary confinement) they may experience psychotic phenomena. Other persons may have a higher vulnerability to stress due to a genetic loading or individual psychological factors, if they experience early traumatic anxiety provoking incidents such as a maternal bereavement (hence a psychological vulnerability to later stress) and a significant stressful life experience in early adulthood then psychotic experience may follow. This model is obviously simplistic. However it does unite different approaches to psychosis. Vulnerability is not a judgmental term but an attempt to understand the variables involved. Nuechterlein and Dawson (1984) suggest a tentative model of psychosis in which certain characteristics of individuals serve as vulnerability factors for psychosis: information-processing deficits, autonomic reaction dysfunction, social competence and coping limitations are seen as vulnerability factors. Stressors in the form of discreet life events as well as general social stress are seen to interact with individual vulnerability factors to produce vicious circles, which eventually lead to psychotic experiences.

The nature of auditory hallucinations

As many as 60% of patients with a diagnosis of schizophrenia experience auditory hallucinations (Slade and Bentall, 1988). Hallucinations are defined as sensory perceptions that have a compelling sense of reality but which occur without external stimulation. It has been shown that auditory hallucinations are accompanied by subvocalization (Inouye and Shimizu, 1970), brain imaging techniques show speech centers of the brain are implicated in auditory hallucinations; however, there remains disagreement whether hallucinations result from a neuro-cognitive deficit or follow from a cognitive bias in normal psychological process (Morrison et al, 1995). The finding that verbal tasks which block subvocalization also decrease the occurrence of auditory hallucinations perhaps adds credence to the cognitive bias rational (Morrison, 1998). Bentall (1990) argues that cognitive bias may be influenced by ‘top down processes’ - i.e. individual’s beliefs and expectations about what kinds of event are likely to occur, and anxiety-reduction (as a negative reinforcement) may assist the misclassification of certain types of internally derived events (e.g. negative thoughts about the self) as externally generated. Bentall’s account would explain why cultural difference in the experience of hallucinations is found, i.e. cultural expectations about how what kind of event is real are socially influenced. Morrison et al (1995) outline an account that proposes that metacognitive beliefs inconsistent with intrusive thoughts lead to their external attribution as auditory hallucinations and this misattribution is maintained by reducing cognitive dissonance. The appraisal of the experience may elicit behavioural, emotional and physiological reactions, which maintain the entire process – e.g. a vicious circle.

A cognitive model of auditory hallucinations based on Clark’s model of panic.
Haddock et al (1996) notes that early approaches to psychosis tended to fit into three main categories: those which involve distraction techniques for psychotic phenomena, those which involve focusing the patient directly onto the phenomena and those which involve anxiety reduction as a target for intervention. The suggestion is that psychosis research has followed a similar path to anxiety disorder in that guided relaxation used to be the mainstay of treatment. Haddock cautions, the beneficial effects of distraction methods applied to psychosis appear to be restricted to the time the technique is employed in the session and do not generalize to real life situations. Potentially we can see a basic similarity to both anxiety disorders and psychosis - with both the emphasis has shifted to directly tackling the problematic symptoms. As with anxiety disorders, in psychosis distraction techniques, and even relaxation methods, may serve the role of inappropriate safety behaviors, which ultimately maintain the psychotic distress. (see Salkovskis, 1991).

Clark’s (1986) model of panic states that panic attacks result from a tendency to catastrophically misinterpret normal body sensations (perhaps in response to an anxiety provoking situation). The sensations are often perceived as indicative of impending doom or death. Clark hypothesizes the tendency to make catastrophic interpretations is maintained in two ways; firstly, selective attention or hypervigilance to individually supposed signs of threat, e.g. monitoring the body for physiological signs of danger. Secondly, the use of safety behaviours which prevent the individual from disconfirming the perceived threat, Salkovskis’s (1991) account of the interaction between threat cognitions and safety seeking behavior is important here.

In panic it has been assumed that normal body sensations are misinterpreted as abnormal events indicative of danger. Internally generated voices can be seen, according to Morrison’s (1998) account, as a relatively normal psychological experience and can be seen as lying on a continuum with normal experience.

Clarke’s model of panic disorder (1986)

Internal / external trigger

Perceived threat

Misinterpretation Apprehension

Body Sensations

According to Kingdon and Turkington (1993) internal voices can occur as a result of sleep deprivation, sensory deprivation, meditation, alcohol or drug withdrawal and at times of increased stress/anxiety. The concept that auditory hallucinations are normal (e.g. Morrison, 1998) is gradually being accepted. Morrison suggests that just as the catastrophic misinterpretation of internal (bodily) experience is crucial to Clark’s panic model so the catastrophic misinterpretation of the normative experience of an inner voice is crucial to psychotic experience. Kingdon and Turkington, (1993) suggest that the meaning attributed to hallucinatory experience is also important, e.g. “that’s the devil talking to me” versus “my brain’s working overtime, I must be tired”. Morrison (1995) suggests that the interpretation of hallucinatory experience may mediate the emotional, physiological and behavioural response to the hallucination in a similar manner to catastrophic misinterpretation in panic; or that the misinterpretation of the hallucination as threatening the physical or psychological safety of the individual is crucial in determining the stress and disability associated with the psychotic experience. Morrison argues that negative appraisal of the uncontrollability and perceived dangers of hallucinatory experience are particularly likely to result in perceived threat to physical or psychological safety; also, increased physiological arousal and negative emotional experiences such as worry and anxiety are in themselves likely to lead to increased and maintained hallucinatory experiences – thus the vicious circle perpetuates.

For panic attack disorder we have noted the importance of hypervigilance and safety behavior in maintaining the condition. Possibly psychotic individuals who use hypervigilance and safety behaviours such as shouting back at the voices or, other distraction methods which in turn raise anxiety (e.g. self harm, alcohol etc), may prevent the individual from naturalistically disconfirming the interpretation of the hallucination, such safety behaviours may maintain the negative impact of the hallucinations. Salkovskis (1991) notes the difference between coping strategies and avoidance strategies; it seems likely that the choice between each strategy is cognitively mediated. Some coping strategies reinforce the individual’s perceived inability to cope with their experience, e.g. shouting “leave me alone” in the face of a critical inner commentary is likely to reduce belief in self efficacy or empowerment; a coping strategy such as a cognitive reappraisal of the relevance and power relationship of the ‘voice’ may lead to a qualitatively different outcome.

Morrison’s (1998) cognitive model of auditory hallucinations

(Internal or external e.g. stress, sleep deprivation, isolation)

Auditory hallucinations

Mood and physiology Safety behaviours

Misinterpretation of hallucinatory experiences

Delusional belief related to a cognitive understanding of anxiety
Delusions are erroneous beliefs involving a mis-interpretation of perceptions or experiences (see, Chadwick et al, 1996). The traditional view is delusions represent unchallengeable beliefs out of keeping with the evidence available and outside the patient’s social and culturally accepted mores. However, delusions are notoriously difficult to define succinctly and cognitive theorists refute the idea that delusional belief is unalterable (see, Kingdon and Turkington, 1996). Delusions represent a misattribution of an internal or external event or in the Beckian A-B-C model a ‘B’ or belief, as opposed to an ‘A’ or activating event as in the case of hallucinations. An example of a delusion would be “aliens are taking me over” as a misattributed response to a TV show. Delusions of reference, of control and persecutory delusions are all common. Several different accounting theories of origin or meaning have been proposed but it is beyond the scope of this essay to address the differences in detail. Nevertheless, there is much evidence that delusional beliefs are formed and maintained in a similar manner to normal beliefs (Gareth and Hemley, 1994) but that patients experiencing delusional belief show an over-confidence in their judgment and reach their judgments more rapidly than ordinarily - one may note here a number of cognitive and attentional biases. Biases may reflect over activity of the cognitive systems designed to protect against low self-esteem (see for example Bentall, 1996) however, there has been mixed results in studies looking at self-esteem and paranoid belief and as such the case remains unproven. Potentially the importance of low self-esteem in psychosis is a clinically relevant heuristic, which has an impact for some patients but is largely irrelevant for others.

In a similar fashion to hallucinatory experience, delusions are more likely to occur at times of greater stress or bodily or cognitive disturbance. Morrison (2004) sees a continuum with the anxiety disorders; the difference is in the cultural unacceptability of the appraisals. With panic disorder an individual may attribute an increase in heart rate as evidence of an impending heart attack; a delusional attribution could be belief in an alien ray that interferes with the heart, or, that the patient is being poisoned by his family etc. As with panic and anxiety models and in common with hallucinatory experience we should highlight the importance of metacognitive belief and selective attentional bias in the maintenance of delusional symptoms. Morrison (2001) argues that delusions may result from information that is accurately perceived but is misinterpreted due to faulty self and social knowledge. Metacognition may play an important role in delusions with specific beliefs leading the person to engage in counterproductive attempts at control that may maintain the delusional belief. Also, metacognitive beliefs may lead to strategies in order to reduce internal inconsistencies. Potentially internal distress caused by anxiety may lead people with vague delusional ideas to use a confirmatory evidence gathering style and hasty decision making, the reduction in anxiety could lead to the reinforcement of such a strategy. Freeman and Garety (1999) have shown a number of similarities between Generalized Anxiety Disorder (GAD) and delusions, indicating that intolerance of uncertainty may have a role in the formation and maintenance of delusions, we shall turn our attention to GAD later on.

Complex trauma and Post Traumatic Stress in relation to Psychosis
There is growing awareness of the relationship between trauma and psychosis. (See, Shaner and Eth, 1989) and associations between traumatic life experiences and the development of psychosis. There is a high incidence of trauma in general and physical and sexual abuse in particular (Read and Argyle, 1999). It is suggested that some patients develop psychosis as a reaction to the trauma, Romme and Escher (1989) discovered that 70% of hallucinators developed voices after a traumatic event and suggested that hearing voices may be part of a coping process; though they are less specific on what this coping process actually involves. Honig et al (1998) found that for most patients the onset of voices was preceded by either a traumatic event or an event that reactivated a traumatic memory; it was considered that the resulting disability incurred by hearing voices was associated with the previous trauma and abuse.

The link between childhood sexual abuse and psychosis is prominent (See, Reid, 1997); patients who report sexual abuse are more likely to report positive psychotic symptoms. Reid believes that traumatic sexual abuse related psychosis may constitute a subset of schizophrenia, versus a more endogenous driven route. Berembaum (1999) noted that reported childhood mistreatment was associated with unusual perception and beliefs. Some authors have suggested a link between the content of abusive experiences and the contents of the psychosis, i.e. the patient may ‘hear’ the abuser commenting upon them – this suggests a causal link. Morrison et al (2004) suggests that both psychosis and Post Traumatic Stress Disorder (PTSD) may form a continuum of responses to trauma, for at least some people. Potentially it is the cognitive, metacognitive, and behavioural factors that determine which disorder follows. There is a high incidence of PTSD in individuals with a diagnosis primarily of psychosis and people with a diagnosis of PTSD may go on to develop psychotic symptoms.

Mueser et al (2002) have suggested a model in which PTSD may mediate the negative effects of trauma for the individual. They suggest PTSD influences psychosis both directly through effects of specific symptoms including avoidance, overarousal and re-experiencing the trauma; and indirectly through the effects of the consequences of PTSD such as re-traumatization, substance misuse and inter-personal difficulties. Once again we see a continuum between an anxiety disorder (PTSD) and psychosis. Morrison (2005) argues that both PTSD and psychosis are linked regarding intrusions; it is the subsequent interpretation of these intrusions that partly defines the diagnosis or type of disorder.

After a traumatic incident(s), which leads to PTSD, the individual may negatively change his/her beliefs about the world, self and self-blame. Despite the high rate of trauma experience within psychosis, Morrison (2004), notes that few studies have looked into the link between trauma, dissociation and psychosis; however, those patients with high levels of dissociative experience in their psychosis have more severe trauma histories. Morrison proposes that it is likely that traumatic experiences will contribute to the development of faulty self and social knowledge and therefore will influence the interpretation of the intrusions, which may explain the link between trauma and the development of psychotic symptoms – e.g. a childhood of sexual abuse may lead an individual not to trust others which would make delusional interpretation of another event more likely. In an attempt to explain the link between trauma and psychosis Allen et al (1997) suggest that dissociative symptoms, which are associated with trauma related incidents, may place an individual at risk of having a psychotic experience. They propose that disassociation prevents reality testing and an associative link with external reality; disassociation may also encourage confusion, disorganization and disorientation etc. In effect Allen suggests the individual may believe and experience that they are locked into an internalized private hell where once again they are victims of abuse from a perpetrator. Misattribution of these experiences will increase distress in a similar manner to the earlier outline of Clark’s panic model and Morrison’s (2001) model of psychosis.

Obsessional Compulsive Disorder and Ruminative thinking

It may be useful to stress the similarity between intrusive thoughts, which are the mainstay of OCD and obsessional disorders, and the intrusions, which sometimes lead to hallucinations or delusions in psychosis. Berman (2001) notes that frequently there is a difficulty in differentiating between OCD and psychotic symptoms, for example both OCD and delusional patients may develop excessive and extreme belief structures. When OCD patients lose insight their symptoms could be described as psychotic, for psychotic symptoms, which become repetitive and intrusive the correlation could be described as OCD-like. The prevalence of OCD symptoms in patients with schizophrenia is conjectured at 20-50% (Berman, 2001) and is considerably higher than symptoms of OCD in the general population; this may lead us to speculate their is more than just simple comorbidity. Kingdon and Turkington (2005) suggest that the failure to recognize hallucinations as one’s own thoughts defines the difference between obsessions and hallucinations, although in practice they lie on a continuum.

The Role of metacognition applied to Generalized Anxiety Disorder and psychosis
An advance in the cognitive analysis of the anxiety disorders is the development of the role of ‘metacognition’; metacognition refers to an individual’s way of thinking about the role of his/her thinking. Wells and Mathews (see Wells, 1997) have proposed a self-regulatory executive function model of emotional disorders (S-REF). The S-REF model suggests that vulnerability to psychological difficulty is associated with: increased self focused attention, attentional bias, increased ruminative processing and activation of maladaptive beliefs. According to the S-REF model cognitive attentional experiences such as biased information processing and self directed executive processes (i.e. self management procedures), which are modified by the individual’s belief system mediate cognitive intrusions. The S-REF model comprises three interacting levels: a level of automatic processing, a level of voluntary self-directed attentional processing and thirdly a level of stored operations available to the individual.

Some beliefs are metacognitive in nature and are linked to the interpretation, selection and execution of particular thinking processes. Wells suggests that such metacognitive beliefs include: beliefs about the thought processes (e.g. “I am unable to concentrate”), the advantages and disadvantages of certain types of thinking (e.g. “My worrying could drive me mad”) and beliefs about the actual content of the thought (e.g.” It is wrong to think about the possibility of death”). Wells argues that for these patients it is both the appraisal and the response to the cognitive process, which distinguishes them from non-patients as opposed to the content of the cognition itself, a similar point to the one made with psychotic intrusions.
To concentrate a little more on psychosis: Stimuli initially undergo automatic processing; this may generate intrusions, which then activate the S-REF. At this stage the S-REF appraises the intrusions from lower level processing and initiates action aimed at reducing discrepancies between cognitive goals (e.g. to reduce cognitive dissonance or to keep oneself safe by being on guard) and the current state of the system. Apart from influencing the immediate focus of attention it affects the sensitivity of the processing system to particular types of information in the lower system. So the S-REF model offers a way of considering the influence of self-knowledge on the distress caused due to S-REF activity. The S-REF model suggests that there is a particular type of attentional ‘disorder’, which underlines all psychological disturbances, and self-focused attention draws attention to this; also, metacognitive belief is considered an important aspect of self-knowledge.

Morrison (2004) suggests that Wells S-REF model potentially makes an important contribution to the understanding of the maintenance of psychotic disturbance. As attentional and metacognitive processes may maintain anxiety; arguably the same processes, i.e. selective attention, metacognitive beliefs (about the controllability of thoughts, about the harmfulness of worry etc) and perhaps cultural beliefs about the content and nature of reality and the role of safety behaviours - all may maintain psychotic process. Wells and Mathews (1994) have incorporated elements of the S-REF into several anxiety models including an update to Clark’s panic model (See, Wells, 1997) and a metacognitive model of OCD; however, it may be useful to consider Wells (1995) model of General Anxiety Disorder (GAD), which in large part derives from the S-REF model.

Worrying is recognised as a major feature of GAD similar to distressing psychosis; for both worry is found difficult to control but for GAD problematic worry is a key feature of the disorder. GAD sufferers report anxiety, restlessness and feeling on edge, tension and sleep disturbance, these difficulties impact on general functioning and cause considerable distress. It is interesting to note that the symptom list (e.g. sleep disturbance, increased distress) is similar to the list of precipitating factors for psychosis proposed earlier by (Kingdon and Turkington (1993).

Wells notes that worrying is both a normal phenomena and an activity which occurs in the emotional disorders, using Morrison’s normalizing analysis it seems to play a role in psychosis too. GAD and ‘normal’ worriers differ little in the content of their worries (See, Craske et al, 1989) but GAD individuals differ in the perceived control they have over worry and are less easy to self-reassure or behaviorally self test when compared to a non-clinical population. Wells suggests ‘worry’ can be seen as a form of conceptual problem solving activity - although it doesn’t work too well. Patients with GAD report periods of chronic worrying or ruminating on a variety of topics, which is believed to be uncontrollable and distressing. Wells’ model’s central feature takes into account the patients appraisal of worrying. Wells distinguishes between two types of worry, type 1 and type 2. Type 1 constitutes the normal content of worrying (external daily events and non-cognitive internal events such as body sensations). Type 2 focuses on worry about worry e.g. “all this worrying will drive me insane”, Wells suggests that GAD is associated with a large amount of Type 2 worry, Morrison’s (2004) analysis assumes that psychotic sufferers have similar Type 2 worries in their concern over their misattributed hallucinations and delusions and in their concerns over their failed attempt to deal with the symptom (e.g. attempts to ignore the inner voice or intrusive thought/delusion). The model also assumes that GAD (and presumably psychosis sufferers) worriers use worry as a coping strategy – i.e. ruminating has benefits, one can think through a problem to avoid a catastrophe. Continuous worrying can be seen as an internal safety behavior, “I’m always prepared for trouble”, “be vigilant in case the devil tells me to harm myself again”. Unfortunately this strategy is often unsuccessful: worry increases attention to threat related information and generates a number of potential future catastrophes – all of which can be worried over in their own right, thus we are back into a vicious circle. Also, the initial belief, which the worry was supposed to resolve, remains unchanged. Once Type 2 worry has been substantively established over time a number of additional factors escalate and maintain the condition, i.e. behavioural responses, thought control attempts and emotional symptoms. Some individuals use worry to block more distressing thoughts; a type of cognitive-emotional avoidance, which potentially leads to a failure to emotionally process information such as the actual nature of a hallucination or delusion. So, for example, it is possible that similar to anxiety disorders, delusions may be associated with an attentional style characterized by self focused attention, attentional bias and activation of dysfunctional metacognitive beliefs.

Wells (1995) cognitive model of GAD


Positive meta-beliefs activated
(Strategy selection)

Type 1 Worry

Negative meta-beliefs activated

Type 2 Worry

Behaviour Thought Control Emotion

A continuum between Psychosis and the Anxiety Disorders

Bentall (1996) notes that early psychological research into schizophrenia concentrated upon the assumed neuro-cognitive deficits of the disorder - thus more nearly purveying psychosis as a biologically qualitatively different experience to that of non-psychotic experience. Morrison (e.g. 2004) offers an approach in the tradition of Aaron Beck and his general approach to the emotional disorders of depression and anxiety. It is assumed that the difference between people with anxiety related or psychotic difficulties are mainly related to the interpretations they form about internal and external events. As opposed to psychosis being some un-understandable primarily bio-physiological process this approach both normalizes the experience of psychosis and suggests the tried and tested cognitive methods of working with anxiety can be adapted to work with psychosis; Morrison maintains that psychotic experiences are on a continuum with anxiety disorders - with psychosis the misinterpretation of intrusive thoughts or physiological symptoms is culturally unacceptable, with anxiety disorders the misinterpretations, appraisals and catastrophizing are seen as culturally more acceptable. Similar to the anxiety disorders the role of safety behaviours is important in the maintenance of psychotic conditions.
Morrison (2001) Cognitive model of psychosis

Intrusions into awareness
(cognitive, body state, emotional or external information)

Culturally unacceptable
Interpretation of intrusion

Beliefs about self, world, others
(procedural and declarative beliefs)

Life experiences

Cognitive and behavioural responses Mood and physiology
including negative symptoms (and negative symptoms)
(including safety behaviours, selective
attention and thought control strategies)

Cognitive therapy of psychosis

A consequence of Morrison’s model is that a hallucinating patient needs to have a catastrophic misinterpretation in order for the hallucination to be a problem – ordinarily a benign internal voice (recognized as such) would not cause distress or need therapy. In the same way that Clark’s model would not advocate treating heart palpitations directly, with psychosis a therapist would not work directly on the voices but would aim to limit distress and disability and the negative consequences associated with the misattribution of the hallucination. As Chadwick (1996) notes, weakening the beliefs associated with the hallucinations weakens the power and the distress of the internal voice. Ultimately the hallucination needs to be accepted and integrated into the individual’s concept of self.

Clinical intervention would involve working with the patient’s perceived difficulties and an individualized formulation. A (perhaps) simplified A-B-C-D framework may be used to introduce the patient to the cognitive framework and the link between internal/external events beliefs/bias/attributions, physiological and emotional consequences and the role of safety behaviours in the perpetuation of the problem. It may be important to eventually confront rather than distract from the hallucination, this would allow for appropriate use of behavioural experiments to test and potentially disconfirm previously held interpretations of hallucinations and review the evidence for and against possible alternative explanations. The normalizing approach of Kingdon and Turkington (1996) is consistent with this rationale.

Based on Wells S-REF model it may be possible to engage the patient into a consideration of the role of metacognitive processing. A therapist could attempt to examine the beliefs and values underlying the use of various cognitive strategies. Recently a delusional patient had a conversation with the present author about the role of paranoid worrying – eventually the conclusion reached was that paranoid worrying both prevented the patient from attending to activities in her life which were important but anxiety provoking in themselves; also, the thought of being victimized while anxiety provoking also helped the client to feel important and was thus a metacognitive activity to bolster low self-esteem. Here, what the patient attended to, and how she paid attention effected how she felt and ultimately what she came to believe.

In general a cognitive therapy of psychosis should aim to be normalizing. The patient can begin to understand that the difficulties they are experiencing are understandable by others. Morrison (2004) takes the approach that either the concerns of the patient are real or else they believe them to be, either way this explains the distress the individual feels. The importance of testing out beliefs by behavioural testing is also to be stressed.

Weaknesses of using a CBT approach to psychosis

Throughout this essay we have been fairly positive to Morrison’s and others approach to a consideration of anxiety models as applied to psychosis. As psychosis remains a relatively new area the importance of adapting a successful and tested model has proved its worth. However, we need to note that often study effect sizes are small and not all results are significant (See, Kingdon and Turkington, 2005). Kingdon also notes that CBT works for some people some of the time. Positive symptoms are easier to work with than negative symptoms although some evidence is accumulating on the role of CBT approaches with negative symptoms the outlook remains poor at present.

Clinically the present author has noted that paradoxically CBT works best with those psychotic patients who least need intervention, patients who are hardest to help unfortunately remain hard to help. The model has little to say about why some symptoms are so florid for some patients other than saying a ‘misattribution’ is at work. To be fair the model’s authors do not propose a full answer and do not verge onto bio-medical matters. Nevertheless, continuously blaming a misattribution style can seem tautologous:- If its psychotic then it’s a misattribution, if it’s a misattribution then its psychosis. Or, . misattribution = psychosis = misattribution.
For some patients the model seems a little pat, as if telling someone in distress that they have simply made an interpretive mistake. Yet, despite psycho-education, behavioural testing and prolonged discussion an individual can remain distressed by their symptoms – knowing that voices are internally self-generated does not help all patients.

The model is somewhat hard to test empirically, i.e. which components are most useful to whom and when? and when would a clinician use one components of an anxiety model as opposed to a component of another. In general, it is hard to analyze which components of a clinical model are stand-alone useful - maybe social validation or therapeutic rapport are more crucial than an understanding of metacognitive process or the role of catastrophization etc, at this stage of research certainty remains impossible.

At times aspects of the anxiety model can read more like a description of a process than a model per se, this perhaps is particularly so with the trauma related model. In general it could be considered that at present the model (in total) represents an important development and a useful clinical heuristic – a general understanding of axis 1 disorders can be applied and at times used to understand and work with a range of psychotic difficulties. Finally, however, theory, therapy and evidence are still evolving.


Schizophrenia according to Kingdon and Turkington (1991) can be sub-divided into five groups: sensitivity disorder, catatonia, anxiety psychosis, traumatic psychosis and drug induced psychosis - it is accepted this is a useful clinical tool rather than a new classificatory system. Nevertheless, their system has implications for what intervention to use for which patient. It would seem that an anxiety based approach might be useful to some with psychosis but not to all.

CBT anxiety models are numerous and cover a lot of ground, in this essay we have covered only a few and have not considered for instance the area of social anxiety, health anxiety or even OCD in any depth; neither have we considered the role of depression in the psychogenesis and maintenance of psychosis, or the importance of personality factors. Potentially many Axis 1 models could be useful in considering an approach to an individual’s psychosis. This would seem to stress the importance of a thorough assessment of the individual and the importance of defining an approach which is individually formulation driven rather than shoehorning an individual into a favorite model of pathology.


Allen, J.G. Coyne, L. and Console, P. (1997) Dissociative detachment
relates to psychotic symptoms and personality decompensation.
Comprehensive Psychiatry. 38:327–334.

Beck, A. (1952) Successful outpatient psychotherapy of a chronic schizophrenic with a delusion based on borrowed guilt. A casebook of Cognitive Therapy for Psychosis (2002). Edited by Morrison, A.P. Brunner-Routledge.

Beck, A. (1993) Cognitive therapy: Past, present, and future. Journal of Consulting and Clinical
Psychology, 61, 194-198.

Bentall, R.P. (1990) The syndromes and symptoms of psychosis: Or, why you cant play twenty questions with the concept of schizophrenia and hope to win. In R.P. Bentall (Ed) Reconstructing Schizophrenia. Routledge.

Bentall, R.P. (1994) Haddock, G. and Slade P.D. Psychological treatment for auditory hallucinations: From theory to therapy. Behaviour Therapy 25, 51-60.

Bentall, R.P. (1996) From cognitive studies of psychosis to cognitive behavior therapy for psychotic symptoms. Cognitive-Behavioural Interventions with Psychotic Disorders. Routledge.

Berembaum, H. (1999) Peculiarity and childhood maltreatment. Psychiatry. 62, 21-35.

Berman, I. (2001). Obsessive-Compulsive Symptoms in Schizophrenia. Psychiatric Times.

Brown, G.W. and Birley, J.L.T. (1968) Crisis and Life change at the onset of schizophrenia. Journal of health and social behaviour 9, 203-24.

Chadwick, P. and Birchwood, M. (1994) Challenging the omnipotence of voices. British Journal of Psychiatry 164, 190-201.

Chadwick, P. Birchwood, M. Trower, P. (1996) Cognitive therapy for delusions voices and paranoia. Wiley Series in Clinical Psychology.

Chadwick, P. (2006) Person Based Cognitive Therapy for Distressing Psychosis. Wiley series in Clinical Psychology.

Clark, D.M. (1986) A cognitive approach to panic. Behaviour Research and Therapy. 24, 461-

Freeman,D. and Garety, P. (1999) Worry,worry processes and dimensions of delusions: an exploratory investigation of a role for anxiety processes in the maintenance of delusional distress. Behavioural and Cognitive Psychotherapy. 27, 47–62.

Garety, P.A. and Hemley, D.R. (1994) Delusions. London. Psychology press. In, Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004). Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Haddock, G. Bentall, RP and Slade, P.D. (1996) Psychological treatment of auditory hallucinations: Focusing or distraction? Cognitive-Behavioural Interventions with Psychotic Disorders. Routledge.

Honeg, A. Romme, M. Ensink, B. Escher, S. Pennings, M. and DeVries, M. (1998) Auditory hallucinations: A comparison between patients and non-patients. Journal of nervous and mental disease. 186, 646-651. In: Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Inouye, T. and Shimizu, A. (1970) The electromyographic study of verbal hallucination. Journal of Nervous and Mental Disease. 151, 415-422. In: Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Kingdon, and Turkington, D. (1991) The use of cognitive behavioural therapy with a normalizing rationale in schizophrenia Journal of Mental and Nervous Disease. 179, 207-211.

Kingdon, D. and Turkington, D. (1996) Using a normalizing rationale in the treatment of schizophrenic patients. Cognitive-Behavioural Interventions with Psychotic Disorders. Routledge.

Kingdon, D. and Turkington, D (2005) Cognitive therapy of schizophrenia. The Guilford Press.

Morrison, A.P. Haddock, G. and Tarrier, N. (1995) Intrusive thoughts and auditory hallucinations: A cognitive approach. Behavioural and Cognitive Psychotherapy. 23, 265-280.

Morrison, A.P. (1998) A cognitive analysis of the maintenance of auditory hallucinations: Are voices to schizophrenia what bodily sensations are to panic? Behavioural and Cognitive Psychotherapy. 26, 289-302.

Morrison, A. P. (2001) The interpretation of intrusions in psychosis: an integrative cognitive approach to hallucinations and delusions. Behavioural and Cognitive Psychotherapy. 29, 257–276.

Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Mueser, K.T. Rosenberg, S.D. Goodman, L.A. and Trumbetta, S.L. (2002) Trauma, PTSD and the course of severe mental illness: an interactive model. Schizophrenia Research. 53,123–143.

Nuechterlein, K. and Dawson, M. E. (1984) A Heuristic Vulnerability-Stress model of Schizophrenia. Schizophrenia Bulletin. 10,300-12.

Paykel, E.S. and Cooper, Z. (1992) Life events and social stress. Quoted in: Gelder, M., Gath, D., Mayou, R. and Cowen P. (1995). In: Oxford Textbook of psychiatry. Third edition. (1996) Oxford university press. P 217.

Read, J. and Argyle, N. (1999) Hallucinations, delusions, and thought
disorder among adult psychiatric inpatients with a history
of child abuse. Psychiatric Services. 50, 1467–1472.

Reid, (1997) Read J. Child abuse and psychosis: a literature review and
implications for professional practice. Professional Psychology: Research and Practice. 28, 448-456. In: Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Romme, M. and Escher, A. (1989) Hearing Voices Schizophrenia Bulletin. 15, 209 – 216.

Salkovskis, P.M. (1991) The importance of behaviour in the maintenance of anxiety and panic: A cognitive account. Behavioural Psychotherapy. 19, 6-19.

Shaner, A. and Eth, S. (1989) Can schizophrenia cause post-traumatic
stress disorder? American Journal Psychotherapy. 4,588–597. in: Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Slade, P.D. and Bentall, R.P. (1988) Sensory Deception: A scientific analysis of hallucination. London: Croom-Helm. In: Morrison, A.P. Renton, J.C. Dunn, H. Williams, S. and Bentall, R.P. (2004) Cognitive therapy for psychosis: A formulation-based approach. Routledge.

Tarrier, N. Beckett, R. Harwood, S. Baker, A. Yusopoff, L. and Ugarteburu, I. (1993) A trial of two cognitive behavioural methods of treating drug resistant residual psychotic symptoms in schizophrenia: 1 outcome. British journal of Psychiatry 162, 524-32.

Wells, A (1997) Cognitive Therapy of Anxiety Disorders. Wiley and sons.

Wells, A. and Matthews, G. (1996) Modeling cognition in emotional disorder. The S-REF model. Behaviour Research and Therapy. 34, 881-888. in: Wells, A (1997). Cognitive Therapy of Anxiety Disorders. Wiley and sons.

Zubin, J. and Spring, B. (1977) Vulnerability – A new view of Schizophrenia. Journal of Abnormal Psychology 86, 2, 103-126.

Counselling Directory is not responsible for the articles published by members. The views expressed are those of the member who wrote the article.

Share this article with a friend

Written by Ray Maloney

I offer people Psychotherapy and Counselling and also practice as an NHS professional and psychotherapist within a Community Mental Health Team I take referrals from GP’s, companies, solicitors, concerned family members  - and of course from the individual. At all times a strict policy of confidentiality is maintained.
Please feel free to email or phone for a free, no obligation… Read more

Written by Ray Maloney

Find a therapist dealing with anxiety

All therapists are verified professionals.

Real Stories

More stories

Related Articles

More articles