Introduction
The at-risk mental state for psychosis and its conceptual cousins, the clinical and ultra-high-risk states (Yung et al., Reference Yung, McGorry, McFarlane, Jackson, Patton and Rakkar1996, Reference Yung, Phillips, Yuen, Francey, McFarlane, Hallgren and McGorry2003; Cornblatt et al., Reference Cornblatt, Lencz and Obuchowski2002) and the Diagnostic and Statistical Manual of Mental Disorders-fifth edition (DSM-5) psychosis risk syndrome proposal (American Psychiatric Association, 2013) are defined largely by the presence of attenuated or fragmentary psychotic experiences. As proto-diagnoses currently considered only within an appendix into DSM-5, these risk states have rightly directed clinical psychiatry and mental health care towards early intervention, and have changed the way we conceptualise the emergence of psychiatric disorders. In fact, they have raised expectations; psychiatric diagnoses may no longer be inevitable and life-long destinations; caught early enough, psychotic disorders may be averted (McGorry and Mei, Reference McGorry and Mei2018).
Systematic reviews and evidence synthesis including classic studies in the field show that someone with an at-risk mental state has a 30–40% chance of meeting criteria for an International Classification of Diseases (ICD)/DSM first-episode psychotic disorder over the following years (Fusar-Poli et al., Reference Fusar-Poli, Bonoldi, Yung, Borgwardt, Kempton, Valmaggia, Barale, Caverzasi and McGuire2012), almost the toss of a coin. However, data from recent studies of briefer mental state abnormalities suggest a different story. Trials of preventative talking therapies (Morrison et al., Reference Morrison, French, Stewart, Birchwood, Fowler, Gumley, Jones, Bentall, Lewis, Murray, Patterson, Brunet, Conroy, Parker, Reilly, Byrne, Davies and Dunn2012), specific neuroprotective medications (McGorry et al., Reference McGorry, Nelson, Markulev, Yuen, Schäfer, Mossaheb, Schlögelhofer, Smesny, Hickie, Berger, Chen, de Haan, Nieman, Nordentoft, Riecher-Rössler, Verma, Thompson, Yung and Amminger2017) or observations of referrals from primary care (Perez et al., Reference Perez, Jin, Russo, Stochl, Painter, Shelley, Jackson, Crane, Graffy, Croudace, Byford and Jones2015) suggest that the short- to medium-term transition rates are much lower: less than 10%, whereas diagnoses of anxiety and depression are common (Hui et al., Reference Hui, Morcillo, Russo, Stochl, Shelley, Painter, Jones and Perez2013; Lin et al., Reference Lin, Wood, Nelson, Beavan, McGorry and Yung2015). In addition, only a few people who develop the first-episode psychosis are detected in at-risk mental state clinics (Ajnakina et al., Reference Ajnakina, Morgan, Gayer-Anderson, Oduola, Bourque, Bramley, Williamson, MacCabe, Dazzan, Murray and David2017).
This low chance of full-syndrome psychotic illness is great news for young people with the so-called at-risk mental state, especially if those looking after them begin helping with their trans-diagnostic presentation (van Os and Guloksuz, Reference van Os and Guloksuz2017), rather than focusing only on the prospect of a full-syndrome psychotic disorder that they almost certainly will not develop. The apparent paradox in the literature is the familiar one arising from the predictive value of a diagnostic or predictive test, here the identification of an at-risk mental state. Its usefulness for predicting full-syndrome psychosis depends on the prevalence of that target disorder; a test that performs well in the rarefied context of a specialist service accepting referrals whom other experts feel may be developing psychosis (such as in the studies that originally fuelled the at-risk concept) will have disappointing characteristics nearer the general population or even in primary care where the ability to predict would be valuable but where psychotic disorder is relatively rare. Diagnostic magic is simply more powerful in the hands of specialist wizards than generalist gatekeepers (Mathers and Hodgkin, Reference Mathers and Hodgkin1989).
Furthermore, there is usually much more to a first-episode, full-syndrome psychotic disorder than the psychotic phenomena that define it; depression and anxiety almost always lurk close by and may be central. We need to give careful attention to all the symptoms and other problems that people with an at-risk mental state bring into the consultation. Many are dealing with moderate to severe depression, anxiety, substance misuse, recent trauma or vivid echoes of adverse childhood experiences. More often than not these are as much part of their presentation as are the fragments of psychosis in the at-risk state; the psychopathology is often kaleidoscopic, changing day by day and evolving week by week.
Studies involving thousands of young people from the general population have demonstrated that psychotic experiences, depressive and anxiety symptoms are best seen as manifestations of a unitary, latent continuum of common mental distress, on which psychotic experiences measure severity (Stochl et al., Reference Stochl, Khandaker, Lewis, Perez, Goodyer, Zammit, Sullivan, Croudace and Jones2015). Therefore, psychotic experiences may act not only as markers for unlikely future psychotic illness or for rare, affective psychotic disorders to be found in current diagnostic classifications but also for more severe species of common, but disabling mental disorders that are hiding in plain sight beyond the eye of the DSM and ICD (Murray and Jones, Reference Murray and Jones2012). Few people with at-risk mental states will develop a psychotic syndrome but many may suffer from persistent mental ill-health that should be a target for intervention. Therefore, we must not assume that the at-risk mental state is some kind of chrysalis that will metamorphose into a psychosis butterfly. In the specialist clinic such a transition may be the result of a clinical judgement increasing a rating-scale score by a single point, far from a dramatic metamorphosis. We wonder whether the at-risk mental state for psychosis may represent a hitherto unrecognised species trapped in an elaborate disciplinary web of knowledge that the field of psychiatrists and researchers has woven over recent decades.
Alfred Lord Tennyson's ‘The Lady of Shallot’ [Tennyson, 1833 (Reference Tennyson1857)] is a lyric ballad poem concerning a cursed Lady imprisoned in a tower on the island of Shalott, near Camelot. Forbidden even a single glance from her window onto the real world, she sits each day happily weaving a charming web that represents the fragmentary reflections of real-life that she glimpses through a small mirror in her bedroom-cum-prison cell. Is our own medical speciality similarly fixated? Are we weaving more and more threads of information on an established disciplinary web whilst, simultaneously, oblivious to being, ourselves, engulfed by the web as it imposes a restricted view of what we see as an at-risk mental state for psychosis?
To look beyond the current notion of at-risk mental state, we must face the conundrum of breaking a familiar web that psychiatrists and psychologists may be weaving. Closer investigation of our metaphor may be useful.
The web
Representing a modern evolution of prodromal psychotic phenomena described in older versions of current diagnostic manuals (American Psychiatric Association, 1987), the at-risk mental state concept and its operationalisation in clinical high-risk criteria provided a new way to cluster and research the early phases of psychotic disorders first, in people who clearly met diagnostic criteria and later, in people who had many but not all of their requirements. Thus, the at-risk mental state epistemology and associated research methodology have been built upon two interlinked paradigms in psychiatry: a categorical nosology and the prediction paradigm. The notion of being ‘at-risk’ for psychosis implied a heightened likelihood of developing categorical disorders included in diagnostic manuals.
Humans have a yen to classify. Just as sensory perception makes sense of complex, multi-domain stimuli to construct our everyday world (and is prone to illusion), psychiatrists and other mental scientists try to reduce the complex patterns of mental phenomena by grouping them into classes with common characteristics. The reasons behind psychiatric classification are diverse and not so different from other medical specialities. In fact, the most basic function of any medical epistemology is to create distinctions (Sokal, Reference Sokal1974; Berrios, Reference Berrios1999). For many years, psychiatry was in need of robust scientific foundations to avoid further ideological contaminations and divisions. The introduction of categorical diagnostic systems in psychiatry was an innovation that provided clear diagnostic criteria, preventing whimsical changes whilst raising the profile of psychiatry as a medical science. Classifications offered a common language and improved the validity and reliability of our diagnostic constructs. The diagnostic categories had to be valid, reliable, stable and specific (Kendler, Reference Kendler1990). This reductionist approach, mostly based on contemplative medicine, inevitably exposes clinicians to uncertainties around clinical presentations that do not fit into any category, and, therefore, are inexistent from a nosological perspective.
To understand the foundations of categories, we must rewind more than a century to Emil Kraepelin's (1856–1926) categorisation of psychiatric diseases; a landmark that would be endorsed by the DSM-III (American Psychiatric Association, 1980; Shorter, Reference Shorter2015). Kraepelin built taxonomy on the assumption that groups of symptoms were disease-specific (Hoff, Reference Hoff2015). Notably, this view was confronted by contemporaneous German psychiatrists such as Alfred Hoche (1865–1943) or Karl Bonhoeffer (1868–1948). For instance, Bonhoeffer considered that a particular illness may express itself with different psychiatric symptoms and that a single symptom might be the expression of several different psychiatric illnesses. His views countered Kraepelin's assertion that psychiatric phenotypes identify natural disease entities and anticipated the view that these phenotypes are likely to be aetiologically heterogeneous. Bonhoeffer's works provided evidence that psychopathological syndromes for a large variety of somatic and psychological disturbances are limited in number and aetiologically nonspecific (Bonhoeffer, Reference Bonhoeffer1908; Ströhle et al., Reference Ströhle, Wrase, Malach, Gestrich and Heinz2008). Bonhoeffer and others presaged the distinction between disease, in which there seem to be an underlying pathological evidence – and illness, in which a person reports symptomatic distress, but no pathological condition can yet be found (Kleinman, Reference Kleinman1988).
We have had examples of illnesses, mostly characterised by distress for which underlying aetiologies were (or were not) eventually found. Gulf War Illness (GWI) is a recent and pertinent one, apparently affecting many soldiers who had served in the Gulf War campaign at the beginning of the 1990s. Taking an approach to definition along the lines of cause, form, course and outcome that would have been familiar to Kraepelin, numerous hypotheses to explain the GWI were never confirmed; they included exposure to different toxic agents or weapons and also psychiatric explanations. Despite being inconclusive, they offer some interesting parallels with the expression of symptoms in the at-risk psychosis syndrome. For instance, the manifestation of GWI, despite having elements of different psychiatric categories, including post-traumatic stress disorder (PTSD), depression, anxiety and somatic symptoms, rarely met all the diagnostic criteria for any of them. Relevant observations confirmed an increase in formal psychiatric disorders, such as PTSD or depression, but they also showed that this increase was insufficient to explain the entirety of an illness that comprised a variety of a priori disparate psychiatric and physical symptoms (Wessely and Freedman, Reference Wessely and Freedman2006). In fact, conditions such as GWI are extraordinarily difficult to investigate given the difficulty to fit a previously inexistent (or never reported) condition into current diagnostic constructs.
By applying the status of psychosis risk broadly, perhaps we have fortuitously discovered another distinctive condition that, in clinical settings, could be called common mental illness including psychotic experiences (Perez et al., Reference Perez, Russo, Stochl, Clarke, Martin, Jassi, French, Fowler and Jones2018), which does not fit into any diagnostic category and challenges the usefulness of the current psychiatric disciplinary web for many individuals. Common mental illness including psychotic experiences should not be another uselessly stigmatising label, but a distinctive definition for an unrecognised presentation affecting a distressed and disabled group. Indeed, evidence suggests that psychosis, depression and anxiety share causes and mechanisms (Nishida et al., Reference Nishida, Tanii, Nishimura, Kajiki, Inoue, Okada, Sasaki and Okazaki2008; Varghese et al., Reference Varghese, Scott, Welham, Bor, Najman, O'Callaghan, Williams and McGrath2011; Kelleher et al., Reference Kelleher, Keeley, Corcoran, Lynch, Fitzpatrick, Devlin, Molloy, Roddy, Clarke, Harley, Arseneault, Wasserman, Carli, Sarchiapone, Hoven, Wasserman and Cannon2012a; Russo et al., Reference Russo, Stochl, Painter, Dobler, Jackson, Jones and Perez2014); however, current psychiatric diagnostic classifications do not yet acknowledge the presence of psychotic experiences in more common depressive or anxiety disorders. This is in spite of past diagnostic manuals, such as the ICD-9, acknowledging the difficulties clinicians may face in distinguishing between neurosis and psychosis in some presentations (World Health Organisation, 1978).
Ladies of Shalott
The concept of risk is commonly used in other areas of medicine but in psychiatry, it may have gained less traction as it lies on a more speculative knowledge-base (Rothstein, Reference Rothstein2008). Psychosis risk follows a conventional medical science framework that keeps accumulating knowledge upon an assumption (the existence of certain clinical phenotypes that increase the likelihood of developing a psychotic disorder) in order to predict more accurately the conversion to psychosis. For many, such as ourselves, who have embarked in this enterprise the ultimate in predicting such conversion has become the Holy Grail; a scientific success that would bring not only patient benefit through preventative interventions but would also contribute to further alignment with the rest of medical specialities. To suffer the first-episode psychosis is irrefutably devastating and we should aim to prevent it; however, this narrowed vision may neglect the here and now. Psychosis risk is often used as a label for many people that will never develop a DSM/ICD psychotic disorder. They may already be suffering from a condition (wrongly named as ‘at-risk’ for another one) that does not fit into the web, but also deserves scientific attention and intervention.
However, a change of focus is not easy, involving a review of paradigms that are deeply entrenched in our scientific communities. According to Thomas Kuhn (1922–1996), deciding between different paradigms can yield good reasons for favouring one over another, but those reasons cannot be codified into an algorithmic scientific method that would decide the point objectively and conclusively. Thus, science is not irrational, but competing paradigms may be incommensurable: there exists no objective way of assessing their relative merits. This incommensurability precludes the possibility of interpreting scientific development as an approximation to the truth. Scientific revolutions simply result in changes to the ontology. Therefore, a paradigm shift is so much one defying a curse or changing his/her own religion, and the reason to do so is often inscrutable. No matter how much evidence is provided to change the paradigm, it will all come to a time of crisis, a final acceptance that the end of the rainbow cannot be reached. Then, such a change begins to be pursued and endorsed widely, regardless of the outcome; this is the core of any paradigm shift (Kuhn, Reference Kuhn1962; Orman, Reference Orman2016).
We researchers working within established paradigms may become Ladies of Shalott weaving a charmed web, inspired by what we are allowed to see, but also cursed by our peers if we look beyond that. We may seem confident, in control, but, in reality, we are trapped in someone else's web. Notably, Tennyson's Lady of Shalott came to the realisation through a state of crisis when she decided to break through the colourful web she had patiently woven around her (see Fig. 1). Driven by the desire of knowing Lancelot and Camelot in the real world beyond her restricted gaze, she defied the curse and escaped. But what can be done until that sense of crisis and desire for change grips at-risk psychosis researchers, making us, all Ladies of Shalott, confront our curse and defiantly break the web?
The Lady of Shalott: Breaking of the Web. William Holman Hunt's illustration to ‘The Lady of Shalott’ from the Edward Moxon edition of Tennyson's Poems, 1857
“She left the web, she left the loom,
She made three paces thro’ the room,
She saw the water-lily bloom,
She saw the helmet and the plume,
She look'd down to Camelot.”
Breaking the web
There is already a consensus on the high clinical relevance of psychotic experiences in depression and anxiety. This co-occurrence is associated with poor quality of life, bad clinical and functional outcomes, and increased risk of self-harm (Granö et al., Reference Granö, Karjalainen, Souminen and Roine2011; Hutton et al., Reference Hutton, Bowe, Parker and Ford2011; Kelleher et al., Reference Kelleher, Lynch, Harley, Molloy, Roddy, Fitzpatrick and Cannon2012b; Hui et al., Reference Hui, Morcillo, Russo, Stochl, Shelley, Painter, Jones and Perez2013; Fusar-Poli et al., Reference Fusar-Poli, Nelson, Valmaggia, Yung and McGuire2014). People suffering from this combination of symptoms are at least seven times less likely to reach remission of depressive symptoms. Psychotic experiences predict poorer response to standard psychological treatments for depression and anxiety, even in combination with pharmacotherapy (Perlis et al., Reference Perlis, Uher, Ostacher, Goldberg, Trivedi, Rush and Fava2011; Wigman et al., Reference Wigman, van Os, Abidi, Huibers, Roelofs, Arntz, Kelleher and Peeters2014; Perez et al., Reference Perez, Russo, Stochl, Clarke, Martin, Jassi, French, Fowler and Jones2018).
Nonetheless, people with common mental illness including psychotic experiences are served badly by current diagnostic classifications and by the aim to perfect a predictive test for psychotic disorders rather than enhancing interventions for current symptoms (Ajnakina et al., Reference Ajnakina, David and Murray2018; Moritz et al., Reference Moritz, Gawęda, Heinz and Gallinat2019). As stated before, these people have psychotic experiences as part of what is often a kaleidoscopic and disabling mental state. They may be found in secondary mental health services, placed under the conceptual umbrella of the at-risk mental state, but it is also highly likely that given an overall lower severity of their mental health problems, a very high proportion attends primary care settings or remains untreated in their communities (Perez et al., Reference Perez, Jin, Russo, Stochl, Painter, Shelley, Jackson, Crane, Graffy, Croudace, Byford and Jones2015, Reference Perez, Russo, Stochl, Clarke, Martin, Jassi, French, Fowler and Jones2018). Many of these individuals may find themselves in a service no-man's-land when they seek help, with therapies given in the wrong settings to the wrong people.
Emerging evidence suggest that at least 25% of all the people with common mental disorders treated by the Improving Access to Psychological Therapies (IAPT) programme in England may have this unrecognised condition (Perez et al., Reference Perez, Russo, Stochl, Clarke, Martin, Jassi, French, Fowler and Jones2018). IAPT services sit in primary care and provide access to evidence-based psychological treatments, predominantly within a cognitive-behavioural therapy framework, to over one million people with common mental disorders every year. Also caught up in the web, these services routinely record patient-reported questionnaires for depression and anxiety but do not measure or treat psychotic experiences (NHS England, 2019). If one-in-four people treated in primary care IAPT services suffer from common mental illness including psychotic experiences, the prospects of recovery for a significant population may be affected by the use of standard talking therapies that do not target all relevant symptoms. This all indicates that, perhaps, the best way to start breaking the web is simply trying to improve these people's life. We have got some highly effective, evidence-based interventions for depression and anxiety, and increasingly confident approaches for helping people with psychotic experiences that may ease these phenomena and promote recovery. Effective psychological treatments for common mental illness including psychotic experiences may already exist (Devoe et al., Reference Devoe, Farris, Townes and Addington2019) but, just as the condition is scattered across different chapters of the diagnostic manuals, these treatments are scattered between different clinical settings that are inadequate to assess or treat trans-diagnostic conditions.
The 2016 NHS Access & Waiting Time Standard for people with an at-risk mental state in England (NHS England, 2016) is resulting in more people with psychotic experiences referred to secondary care psychosis services. This is contrary to research on transition rates, especially in people referred from primary care (Hui et al., Reference Hui, Morcillo, Russo, Stochl, Shelley, Painter, Jones and Perez2013; Perez et al., Reference Perez, Jin, Russo, Stochl, Painter, Shelley, Jackson, Crane, Graffy, Croudace, Byford and Jones2015), the limited capacity of these services and their users' needs, many of whom have common mental illness including psychotic experiences. Accommodating them and providing an effective therapy in low-stigma IAPT services will allow specialist early intervention services to concentrate on first-episode psychotic disorders (van Os and Guloksuz, Reference van Os and Guloksuz2017; Perez et al., Reference Perez, Russo, Stochl, Clarke, Martin, Jassi, French, Fowler and Jones2018). If the available treatments were reconfigured and assembled into a practical therapeutic toolbox for use in primary care settings, e.g. IAPT services, where most people with the condition are already seeking help and where they would receive their treatment, prospects for recovery could be greatly improved.
The UK National Institute for Health Research (NIHR) recently funded the TYPPEX programme (National Institute for Health Research, 2018). This national programme primarily aims to provide effective treatment to a vulnerable group in the lowest possible stigma setting. It will develop and test a tailored psychological therapy and training for therapists in IAPT services to detect and treat people with common mental illness including psychotic experiences in their caseloads more effectively. Bold and iconoclastic, it should provide trans-diagnostic evidence beyond the inadequate boundaries of the diagnostic web, and guide further applied health research on treatments and mental health services for this population.
Conclusion
Current psychiatry is built upon accepted paradigms. Proto-diagnoses, such as the at-risk mental state for psychosis, remain loyal to our long-standing categorical nosology and risk and prediction paradigms mostly adopted from other medical specialities. Despite transition rates in people with the at-risk mental state have declined over the years, especially in samples referred from primary care settings, we continue perfecting prediction tools to whittle down populations in order to detect those at the highest risk. This whittling exercise forbids us from perceiving features in a wider horizon where we may have unwittingly discovered another condition that could be called common mental illness including psychotic experiences.
The at-risk mental state, considered but ultimately shelved by the DSM-5, is the exemplar that should make us break free of the entire diagnostic web. However, such a shift of vision would imply defying the spell preventing us from any attempt to think differently. Tennyson's poem ends with the Lady of Shalott breaking the web that she carefully wove over the years, but dying on a boat just before reaching Camelot. She did not manage to see the real world in its entirety, as by breaking a curse she triggered another one. This outcome also reflects the nature of paradigm shifts. No new paradigm is better than an old one; it simply continues enriching our knowledge until gradually turns into another curse that prevents us from finding the truth.
Acknowledgements
JP and PBJ acknowledge salary support from the National Institute for Health Research (NIHR) Programme Grant for Applied Research RP-PG-0616-20003. The views expressed are their own and not necessarily those of the NIHR or the Department of Health and Social Care.
Conflict of interest
JP has consulted for Lundbeck, Otsuka, Lilly, Roche and Ricordati. PBJ has consulted for Lundbeck, Janssen and Ricordati.
