Improving treatments for psychotic disorders: beyond cognitive behaviour therapy for psychosis

Abstract

More effective treatments for people with psychotic disorders are urgently required. Here, we make three suggestions for progress: 1. Targeting the disorders’ core phenomenological features (‘phenomenological phenotype’), 2. Addressing social disconnection, isolation and loneliness, and 3. Leveraging ‘hot’ cognitions and using symptom capture approaches that combine psychotherapy with advances in technology and neuroscience.

Current treatments for schizophrenia and other psychotic disorders have limited efficacy. Treatment guidelines for these disorders recommend the use of a second-generation antipsychotic medication, psychosocial support, and psychological treatment (i.e. cognitive-behaviour therapy for psychosis, CBT-P; Galletly et al., 2016; Group, E. P. G. W., 2010; Group, I. E. P. A. W., 2005; Lehman et al., 2004; NICE, 2014). However, studies have repeatedly shown that at least a third of patients, possibly up to two thirds of patients (Samara et al., 2019), will show an inadequate response to antipsychotic medication (Andreasen et al., 2005; Kane et al., 1988; Lehman et al., 2004). Psychological treatments do not fare much better (Jauhar et al., 2019; McKenna et al., 2019). A 2014 meta-analysis found end of treatment effect sizes that were uniformly in the small range (0.13–0.33; Jauhar et al., 2014). A more recent network meta-analysis of various psychological interventions to reduce positive symptoms in schizophrenia again found pooled effect sizes for CBT in the small range (0.29–0.30; Bighelli et al., 2018). A 2018 Cochrane Review concluded that there is no clear and convincing advantage for CBT on any measure apart from leaving the study early over other, and sometimes much less sophisticated and expensive, psychosocial therapies for people with schizophrenia (Jones et al., 2018).

Given the limited benefit of the existing recommended “gold standard” treatments for people with psychotic disorders, it is clear that more effective treatments are urgently required for this patient population (Lopez et al., 2006; Neil et al., 2014). In this piece, we make three suggestions for over-coming the current impasse in treatment: 1. Targeting the core phenomenological features (“phenomenological phenotype” (Stanghellini & Rossi, 2014)) of the conditions in question, 2. Focusing on social disconnection, isolation and loneliness, and 3. Leveraging “hot” cognitions and using “symptom capture” approaches that combine CBT-P with advances in technology and neuroscience.

Although we recognise that one of the first reports of CBT was its use for psychosis (Beck, 1952), much of the development of this therapy has been geared towards the treatment of depression and anxiety. Based on reasonable success in these clinical domains, it was imported and transposed into the area of psychosis treatment (Fowler et al., 1995; Kingdon & Turkington, 1994; Tai & Turkington, 2009). However, this was essentially a pragmatic move and influenced by the growing field at the time of cognitive models of psychosis, rather than a therapy that was developed in response to characteristic phenomenological features of psychotic disorders and possible associated pathogenic mechanisms. We have recently argued that close examination of a psychiatric disorder’s phenomenological phenotype is required in order to direct research into mechanisms of disorder onset and integrate levels of analysis (Nelson et al., 2019). The same could be argued with regard to treatment. That is, rather than mechanically applying whatever happens to be the dominant psychotherapeutic approach (in this case, CBT), treatment should be selected to match phenomenology. A substantial body of theoretical and empirical research has now accumulated indicating that a core phenomenological feature of schizophrenia spectrum conditions is a self-disorder marked by structural distortions of subjectivity and consciousness (Nelson et al., 2014; Parnas & Zandersen, 2018). This basic self-disturbance is associated with a range of mutually implicating anomalous subjective experiences such as diminished sense of presence or inner nucleus, hyperreflexivity, diminished sense of ownership and agency of experience and action, dissociation and derealisation, and “common sense” disturbances.

A therapy that targets these features may be most relevant to patients and be more effective than traditional CBT-P given its engagement of core phenomenological features. We advocate the development and testing of such an approach. Although the precise nature of this therapy remains to be clarified, core components have already been identified. These include:

• An emphasis on the overall framework of altered experience rather than on specific contents of cognition (i.e. sometimes referred to as a focus on form/structure over content (Sass, 2019)). An example would be privileging the recognition and discussion of a “background” delusional mood over challenging specific paranoid thoughts.

• The value of implicit and nonverbal intersubjective experience rather than overemphasising the relevance and need for direct interpersonal interaction (Sass, 2019). As recently suggested (Nischk & Rusch, 2019), this may be an element of the Soteria approach to acute psychosis treatment that contributes to its effectiveness and appeal to patients, i.e. a community-based social milieu that emphasises personal relationships (“being-with”) and meaningful shared activities (“doing-with”) over direct verbal discussion about mental state and symptoms.

• Developing and supporting a more robust sense of self-presence. It has been suggested else-where (Nelson et al., 2009; Skodlar & Henriksen, 2019) that helpful strategies to achieve this include body-oriented strategies (e.g. physical activities, mindfulness, yoga, movement and music therapies (Geretsegger et al., 2017; Mehta et al., 2016; Mittal et al., 2017; Priebe et al., 2016)) and immersive activity (e.g. simulation, “flow” states (Csikszentmihalyi, 1990, 1996)). These may cultivate a “dereflexion”, i.e. a reduction in excessive analysis, reflection and anxious awareness, and stronger sense of embodiment (Skodlar & Henriksen, 2019). Interoceptive processes, which have been found to be disturbed in schizophrenia, may play a crucial role in sense of presence (Allen and Tsakiris, 2018; Seth, 2016; Seth et al., 2011). Physical exercise may be a means of increasing implicit interoceptive and proprioceptive awareness. It is also noteworthy that a key feature of Virtual Reality (VR) paradigms is to maximise a sense of presence or immersion in the virtual environment. If this could be maximised in the virtual environment, then features that contribute to this effect could be identified together with the patient so that they could possibly be drawn on and utilised in normal self-world experience. The “transfer effect” from VR to real environments is critical to achieve a sustained impact on disturbances of self-presence in psychotic disorders. In a recent feasibility study using VR to train social attention, we found medium to strong effect sizes for symptom reduction following 10 training sessions (Adery et al., 2018). Interestingly, we also found that emotional embodiment improved following training, highlighting the potential role of VR in addressing bodily self-disturbances.

Secondly, we argue that it is also essential to address the self-other/self-world relationship in order to improve functional outcomes and quality of life. Indeed, self-experience and interpersonal/social experience are intimately intertwined (Nelson et al., 2009; Ratcliffe, 2017; Sass et al., 2018). The phenomenology of psychotic conditions is marked by perceived social disconnection, isolation, and/or loneliness (Tremeau et al., 2016) and the DSM-V describes avolition (which can be associated with asociality, “an apparent lack in social interactions”) as a core negative symptom of schizophrenia (Association, A. P, 2013). Notably, objective social isolation and disconnection have been found to significantly impact the quality of life of those affected (Green et al., 2018). In addition, social isolation has been shown to induce schizophrenia-like behaviour in animal models (Li et al., 2017). Perceived and objective social isolation (and related constructs of social defeat (Selten et al., 2013) and marginalisation (Morgan et al., 2010)) are not only risk factors for schizophrenia, but they are also core symptoms for those already affected by the illness and predict overall outcome. Thus, social isolation and loneliness are essential targets for treatment. Two approaches to reducing social isolation are improving social skills and tailored online social media. With regard to the former, VR is again an excellent candidate as it allows for controllability of the environment and repeated practice, which are key ingredients for learning. Several studies have already shown the efficacy of VR for improving social skills in schizophrenia (e.g. Adery et al., 2018; Rus-Calafell et al., 2014; Tsang & Man, 2013). Specifically designed online social media has also been found to reduce social isolation and increase meaningful social connections in patients with psychosis (Alvarez-Jimenez et al., 2014, 2018).

A third suggestion is that even if treatments do not target putative core phenomenological features, as suggested above, CBT-P could be enhanced and combined with other elements to maximise its effectiveness. One of the rate-limiting factors for CBT may be that it tends to rely on a considerable degree of abstract self-reflection (e.g. identifying and challenging the triad of negative views of self, others and the future (Beck et al., 2019)) and can, therefore, be divorced from the actual in-vivo experience of symptoms. Such reflection also has the potential to exacerbate the hyperreflexivity that may be central to the pathology itself (Nelson, et al., 2009; Skodlar et al., 2013). The efficacy of CBT may be enhanced if applied during symptom activation (referred to as a “symptom capture” approach (Jardri et al., 2011)), i.e. addressing “hot” versus “cold” cognitions (Roiser & Sahakian, 2013). This approach has been applied with good effect to anxiety disorders and obsessive-compulsive disorder, using exposure and response prevention paradigms. Again, VR paradigms may be useful to achieve this aim (Freeman et al., 2016; Pot-Kolder et al., 2018), as has been conducted to some extent with Avatar therapy (Leff et al., 2013). VR would allow the development and application of flexible and personalised exposure hierarchies.¹

The utility of this approach may be further augmented if combined with bio- or neurofeedback that provides real-time feedback to patients on their biological and/or neural activity (e.g. heart rate variability, electrophysiological patterns) in order for them to modulate this activity during symptom activation. There is proof-of-concept evidence for this treatment approach in psychosis (Bolea, 2010; Cordes et al., 2015; Gruzelier et al., 1999; Nan et al., 2017; Orlov et al., 2018; Ruiz et al., 2013; Surmeli et al., 2012). Again, the transfer effect to everyday life should, of course, be a key treatment goal. Combining treatment components in this way may not only engage multiple treatment targets and possible pathogenic mechanisms (psychological, neurophysiological, neurobiological) simultaneously but is also likely to have a synergistic effect through the interaction of the individual treatment components. For example, real-time bio/neurofeedback provided to patients and therapists in response to VR exercises would feed into the content of CBT, because the patient and therapist would be able to discuss the direct impact the exposure exercises are having on the patient in an immediate, non-retrospective fashion, preserving the benefits of dealing with “hot” cognitive and affective reactions. In turn, CBT may assist the patient in developing strategies for modulating their bio/neural activity in response to the VR exposure paradigms. The neurofeedback component may also increase a sense of self-efficacy, given that it is grounded in patients developing self-directed, endogenous control over their symptoms. In this way, the combined effect of such an integrated intervention (CBT-P, VR, bio- or neurofeedback) may be greater than the sum of the effects of its individual components.

Similarly, blended models of care integrating real-time assessments (ecological momentary assessments, passive sensing (Insel, 2018)) and interventions (EMI (Reininghaus et al., 2016)) with face to face CBT therapy, can yield treatment effects which surpass those of their individual components. For example, real-time CBT strategies can be tailored based on individual response using therapist input and machine learning and adapted to time and context via passive sensing technology (e.g. delivering cognitive strategies in specific locations).

To summarise, there is a need for treatments that foster a more robust basic selfhood, that improve self-other/world functioning, and that integrate learnings from psychological treatment of other conditions (e.g. the benefit of dealing with “hot” cognitions) with advances in technology and neuroscience. Existing mainstream treatments and those suggested here are not mutually exclusive or incompatible with each other. CBT-P remains a useful tool, but other strategies need to be further explored in order to provide more effective types of treatments and to augment the benefits of those already available. Rather than relying exclusively on top-down cognitive approaches, the treatment of psychotic disorders may need to engage multiple targets simultaneously and aim to put the self back into the body and also the person back into society.