Introduction
Contemporary research into psychotic illness is both elaborating and altering long-standing concepts of its psychopathology, pathobiology and trajectory, both before and after initiating treatment (Kahn et al. Reference Kahn, Sommer, Murray, Meyer-Lindenberg, Weinberger, Cannon, O’Donovan, Correll, Kane, van Os and Insel2015; Owen et al. Reference Owen, Sawa and Mortensen2016). We have recently proposed (Zhen & Waddington, Reference Zhen and Waddington2018) that this reflects two types of processes: evolving insights increase our understanding incrementally along established lines with which we feel comfortable, typically in the context of conventional diagnostic categories such as schizophrenia; in contrast, emerging insights act heuristically by deviating from established lines to challenge fundamentally our understanding and can engender discomfiture.
On a background that extends to the earliest concepts of psychotic illness (Ram et al. Reference Ram, Bromet, Eaton, Pato and Schwartz1992; Million et al. Reference Million, Grossman and Meagher2004), one important vehicle for elaboration and alteration has been the emergence of a new generation of systematic studies of first-episode psychosis that seek to define, prospectively, the nature and course of illnesses such as schizophrenia from what was conceptualised conventionally as ‘onset’, through the early phases of treatment, to long-term care and outcome. Following initial appreciation by 1992 that this vehicle had ‘taken to the road’ (Keshavan & Schooler, Reference Keshavan and Schooler1992; Kirch et al. Reference Kirch, Lieberman and Matthews1992), by 2005 no less than 25 such studies had begun their journey. Among these, the Cavan-Monaghan First Episode Psychosis Study (CAMFEPS; Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005), a collaboration between Cavan-Monaghan Mental Health Services (CMMHS) and the Royal College of Surgeons in Ireland (RCSI), adopted what was then an idiosyncratic and heuristic approach: it sought to ascertain and evaluate ‘all’ incident cases of psychosis within the counties of Cavan and Monaghan, on an epidemiological basis, with diagnosis being just one of multiple post-inception assessments rather than any criterion for entry; furthermore, the study was embedded within CMMHS during a period of progressive evolution into a home-based model of care with substantive primary care liaison, and subsequently into a new service model for the early detection and treatment of psychosis.
Why did CAMFEPS adopt this then idiosyncratic approach? One of the most challenging contemporary concepts, sustained and elaborated by increasing evidence, is that psychotic psychopathology, underlying developmental pathobiology and associated risk genes are not only disrespectful to conventional diagnostic boundaries but also indicate the arbitrary nature of our diagnostic categories, both cross-sectionally and prospectively. As clinical practice evolves (see our companion article, Russell et al. Reference Russell, Nkire, Kingston and Waddington2019), it is likely to continue to involve the establishment of psychotic diagnoses, as enshrined in DSM-5 and ICD-11; these provide a guide to treatment, generate much useful data that can inform on the above challenges and have pragmatic value such as for reimbursements from health insurers. Yet they are under increasing scrutiny. This is exemplified by current research emphasis on the increasingly blurred and porous schizophrenia-bipolar interface (Cross-Disorder Group of the Psychiatric Genomics Consortium, 2013a, 2013b; Owen Reference Cardno and Owen2014; Pearlson, Reference Pearlson2015; Birur et al. Reference Birur, Kraguljac, Shelton and Lahti2017) that is at variance with and challenging to classical nosology (the ‘Kraepelinian dichotomy’; Craddock & Owen, Reference Craddock and Owen2010).
Further concepts and increasing evidence are equally challenging: that psychotic ideation and associated psychopathology can be present in young persons across the general population and, according to nature and intensity, may be early harbingers of risk for clinical psychosis (Fusar-Poli et al. Reference Fusar-Poli, Borgwardt, Bechdolf, Addington, Riecher-Rössler and Schultze-Lutter2013; van Os & Reininghaus, Reference van Os and Reininghaus2016); and that early intervention for features associated with clinical high risk/at-risk mental state and at the first psychotic episode may, respectively, ameliorate the emergence of diagnostic psychotic symptoms and improve long-term outcome (Clarke et al. Reference Clarke, McDonough, Doyle and Waddington2016; Millan et al. Reference Millan, Andrieux, Bartzokis, Cadenhead, Dazzan and Fusar-Poli2016). Additionally, this evolving evidence base continues to influence how service models for both early intervention and subsequent longer-term care for psychotic illness should be conceptualised and applied; should conventional models of care be modified to incorporate such new insights; or do these insights suggest more radical restructuring of service provision within an alternative concept of mental health and dysfunction (McGorry et al. Reference McGorry, Hartmann, Spooner and Nelson2018)? Here we review several of these challenges using findings from CAMFEPS to illustrate their complexity in a real-world setting.
Outline of CAMFEPS
In 1991, the Health Research Board funded the Schizophrenia Research Unit, a collaborative endeavour involving inter alia RCSI/CMMHS and St. John of God Hospitaller Ministries/Cluain Mhuire Community Mental Health Services. On the basis of studies carried out by the Unit (Waddington & Larkin, Reference Waddington and Larkin1993) in 1995, the [then] Theodore & Varda Stanley Foundation [subsequently the Stanley Medical Research Institute] awarded philanthropic funding to initiate and sustain two parallel first-episode psychosis studies: one (CAMFEPS) based in two contiguous rural counties in the Northeast of Ireland having a total population of approximately 109,000; the demographic and socioeconomic characteristics of these counties have been described previously in detail (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005); the other based in a more urban setting proximal to Dublin (St. John of God Hospitaller Ministries/Cluain Mhuire Community Mental Health Services, Co. Dublin; Browne et al. Reference Browne, Clarke, Gervin, Waddington, Larkin and O’Callaghan2000). These two studies operated in parallel for 4 years, after which the St. John of God Hospitaller Ministries/Cluain Mhuire Community Mental Health Services study began a process of evolution into the Dublin and East Treatment and Early Care Team (DETECT) early intervention service (Renwick et al. Reference Renwick, Gavin, McGlade, Lacey, Goggins, Jackson, Turner, Foley, McWilliams, Behan, Lawlor, Cullen and O’Callaghan2008), while CAMFEPS continued for a total of 15 years, during which it evolved into the Carepath for Overcoming Psychosis Early (COPE) early intervention service (Nkire et al. Reference Nkire, Sardinha, Nwosu, McDonough, De Coteau, Duffy, Waddington and Russell2015), as outlined below (see Early intervention) and elaborated in a companion article (Russell et al. Reference Russell, Nkire, Kingston and Waddington2019). The structure of CAMFEPS involved a Clinical Research Fellow/Registrar embedded within CMMHS and having both a research role and sessional service commitment; the service model involved two community mental health teams, a specialist service for the elderly and a community rehabilitation team, with primary care liaison and use of home-based treatment as an alternative to hospital admission being central to the delivery of health services in this model (McCauley et al. Reference McCauley, Rooney, Clarke, Carey and Owens2003; Nwachukwu et al. Reference Nwachukwu, Nkire and Russell2014; see Russell et al. Reference Russell, Nkire, Kingston and Waddington2019).
While the operation of CAMFEPS in terms of case identification and assessment has been described previously in detail (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013), essential points include (a) case ascertainment via all routes to care [i.e. public (CMMHS: home-based, outpatient and inpatient), private (St. John of God Hospital, Co. Dublin; St. Patrick’s University Hospital, Dublin) and forensic (Central Mental Hospital, Dublin)]; inception throughout the adult lifespan (age 16 and above, i.e. no arbitrary upper age cut-off); inception of all 12 DSM-IV psychotic diagnoses [schizophrenia; schizophreniform disorder; schizoaffective disorder; delusional disorder; brief psychotic disorder; bipolar disorder; major depressive disorder, with psychotic features; substance-induced psychotic disorder; psychotic disorder due to a general medical condition; substance-induced mood disorder, with manic features; mood disorder due to a general medical condition, with manic features; psychotic disorder not otherwise specified].
Epidemiology
Classical research and literature on the epidemiology of psychotic illness derives primarily from studies of schizophrenia and indicates a disorder having the following profile: an incidence that can vary across countries and cultures, has its onset primarily in young adulthood, and occurs earlier in men than in women. However, while some of these features have been generally sustained by contemporary studies, they now appear to be simplistic and incomplete, particularly in relation to diversity in case ascertainment, social milieu and ‘strictness’ of diagnostic criteria, over-representation of data on schizophrenia vis-à-vis other diagnoses in which psychosis can occur, and application of arbitrary upper age cut-offs (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; McGrath Reference McGrath2005; Kirkbride et al. Reference Kirkbride, Errazuriz, Croudace, Morgan, Jackson, Boydell, Murray and Jones2012).
In CAMFEPS, on ascertaining cases via all routes to care, dispensing with diagnostic restriction and arbitrary upper age cut-off and applying internationally recognised DSM-IV diagnostic criteria, a more complex and nuanced epidemiological landscape is apparent. Over the 15-year period (1995–2010), CAMFEPS incepted 432 cases: 92 major depressive disorder with psychotic features (MDDP); 89 bipolar disorder (BD); 81 schizophrenia (SZ); 25 schizoaffective disorder (SA); 25 delusional disorder; 25 substance-induced psychotic disorder; 23 brief psychotic disorder; 20 schizophreniform disorder; 13 psychotic disorder due to a general medical condition; 8 substance-induced mood disorder with manic features; 4 mood disorder due to a general medical condition, with manic features; and 25 psychotic disorder not otherwise specified. There were also two cases of simple deteriorative disorder, a DSM-IV Appendix category characterised by all the hallmarks of schizophrenia in terms of negative symptoms and functional decline but without sufficiently prominent positive symptoms to satisfy criteria for SZ; this condition overlaps with attenuated psychosis syndrome, a DSM-5 Condition for Further Study. The overall incidence of any psychotic illness was 34.1/100,000 of population age over 15 years (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, Russell, O’Callaghan and Waddington2013; Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013; Nkire et al. in preparation).
An unexpected finding on applying such methodology in this ‘real-world’ setting was the breadth of diagnoses under which psychotic psychopathology was manifest, with the incidence of psychosis under the three most populous diagnoses being indistinguishable between SZ, BD and MDDP. Yet each of these diagnostic categories showed a distinct profile: SZ was threefold more common and was first diagnosed at a younger mean age in men than in women; BD was indistinguishably common and first diagnosed at an indistinguishable mean age in men and women; MDDP was indistinguishably common and first diagnosed at an indistinguishable mean age in men and women but that mean age was substantially older, by some 20 years, than for SZ and BD. This later finding illustrates the impact of arbitrary upper age cut-offs in studies of psychotic illness through exclusion of many MDDP cases and this artefact applies also to a smaller, but not inconsequential, number of SZ and BD cases. Critically, these distinct epidemiological ‘signatures’ in CAMFEPS should not be misinterpreted as validating these diagnostic categories; when ascertained appropriately these differences are quantitative rather than qualitative, such that each diagnosis of SZ, BD and MDDP (and also SA, schizophreniform disorder, delusional disorder and psychosis not otherwise specified) can and does occur in either sex, with the onset of psychosis occurring at any age over the adult lifespan, from the teens through to the ninth or, occasionally, the tenth decade (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, Russell, O’Callaghan and Waddington2013; Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013; Nkire et al. in preparation).
These findings from CAMFEPS have contributed to a recent international collaborative analysis of SZ across 43 methodologically diverse datasets amounting to 133,693 cases, which elaborates differential susceptibility to SZ for men and women with age (van der Werf et al. Reference Van der Werf, Hanssen, Kohler, Verkaaik, Verhey, van Winkel, van Os and Allardyce2014). Additionally, CAMFEPS shows that while these relationships differ quantitatively in BD and MDDP vis-à-vis SZ, they do not constitute qualitative discriminators between them; rather, these, together with other psychotic diagnoses encountered less commonly, are in epidemiological continuity with each other in the absence of any points of rarity (Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019).
Diagnostic instability
An important aspect of psychotic illness is the stability or otherwise of such diagnoses on a long-term basis; in particular, do initial diagnoses remain stable, diverge or converge to a smaller number of more highly populated categories that might be considered more fundamental diagnostic nodes? In a 6-year follow-up of the first 202 cases incepted into CAMFEPS, similar robust ascertainment methods allowed diagnostic reassessments for 196 (97%) cases with quantification of prospective and retrospective consistency. While the 12 initial psychotic diagnoses were characterised by numerous transitions, these were diverse with only limited convergence towards a smaller number of more stable diagnostic nodes, most commonly SZ. Notably, for 85% of cases having an initial diagnosis of brief psychotic disorder this was the harbinger of long-term evolution to a serious psychotic illness of diagnostic diversity (SZ, SA, delusional disorder, BD, MDDP), as were the two cases of simple deteriorative disorder. Furthermore, 15% of cases having an initial diagnosis of MDDP were deceased at 6 years due to diverse causes, a concern that may not simply reflect cases of MDDP being a mean of some 20 years older at inception than cases of SZ and BD. Enigmatically, 31% of cases having an initial diagnosis of psychotic disorder not otherwise specified continued over 6 years to defy any other DSM-IV diagnosis (Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, Russell, O’Callaghan and Waddington2013).
These findings from CAMFEPS have contributed to a recent international meta-analysis of diagnostic stability across 42 methodologically diverse first-episode psychosis datasets amounting to 14,484 cases, which provides estimates of diagnostic stabilities and transitions across diverse periods of follow-up for a circumscribed range of DSM-IV psychotic diagnoses, among which BD and MDDP are compacted into a single category of ‘affective spectrum psychosis’ (Fusar-Poli et al. Reference Fusar-Poli, Cappucciati, Rutigliano, Heslin, Stahl, Brittenden, Caverzasi, McGuire and Carpenter2016). Across all 12 DSM-IV psychotic diagnoses, CAMFEPS reveals a diversity of stabilities in, and transitions between, these diagnoses over a 6-year period that indicates some ‘fluidity’ in and thus longitudinal disrespect to such categorisation; in particular, a first episode of brief psychotic disorder or MDDP may be an indicator for vigorous and sustained interventions (Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, Russell, O’Callaghan and Waddington2013; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019).
Clinical characteristics
That CAMFEPS methodology reveals MDDP to be just as common as SZ and BD at a first psychotic episode indicates that recent emphasis on the SZ-BD interface (see Introduction) may need to be broadened yet further. Indeed, we have argued that MDDP is an archetype for the challenge of confluence between psychotic and affective domains of psychopathology, even more so than BD or the long-standing conundrum of SA (Waddington & Buckley, Reference Waddington and Buckley2013; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019). Current theory emphasises dimensional rather than solely diagnostically based perspectives of psychotic illness (van Os & Kapur, Reference van Os and Kapur2009; Barch et al. Reference Barch, Bustillo, Gaebel, Gur, Heckers, Malaspina, Owen, Schultz, Tandon, Tsuang, van Os and Carpenter2013; Owen, Reference Owen2014; Owen et al. Reference Owen, Sawa and Mortensen2016; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019) in a manner complementary to the above epidemiological evidence. On this basis, CAMFEPS has undertaken systematic, heuristic comparisons between SZ, BD and MDDP at the first episode across the clinical domains of psychopathology, neuropsychology, neurology, premorbid adjustment and quality of life, the interim results of which can be summarised as follows (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013):
Psychopathology
SZ, BD and MDDP evidenced indistinguishable severity of positive symptoms, with negative symptom severity being highest in SZ, slightly less so in MDDP and less so in BD; assessment of negative symptom severity in MDDP may be confounded with symptoms of depression.
Neuropsychology
SZ, BD and MDDP were indistinguishable on screening for current general cognitive function and on the estimation of premorbid intellectual functioning prior to the onset of psychotic symptoms; SZ and MDDP evidenced indistinguishable severity of executive dysfunction, with BD appearing somewhat less impaired.
Neurology
SZ and MDDP evidenced indistinguishable prominence of neurological soft signs (minor neurological signs that indicate non-localising cerebral dysfunction), with BD appearing somewhat less impaired; each of SZ, BD and MDDP showed indistinguishably low levels of extrapyramidal movement disorder and abnormal involuntary movements, as expected given no or only brief exposure to antipsychotic drugs when evaluated at the first psychotic episode.
Premorbid adjustment
SZ, BD and MDDP evidenced only subtle differences in impairment of premorbid adjustment, which varied slightly between the age ranges of up to 11 years versus 12–15 years.
Quality of life
SZ and MDDP evidenced indistinguishable impairment in quality of life, with BD appearing less impaired.
Overview
In CAMFEPS, interim findings in SZ, BD and MDDP indicate very similar clinical profiles, with any differences being quantitative rather than qualitative in the presence of considerable overlap and no points of rarity between them (Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013). While every diagnosis of SZ indicates, by definition, an intrinsically psychotic disorder, independent of severity, this is less clear for BD across all severities of illness in relation to presence or absence of the DSM-IV specifier for BD ‘severe, with psychotic features’ (Pearlson, Reference Pearlson2015; Anderson et al. Reference Anderson, Marder, Reise, Savitz, Salvadore, Fu, Li, Turkoz, Han and Bilder2018; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019). To clarify this issue, CAMFEPS has investigated BD cases divided according to the presence or absence of this DSM-IV specifier and compared these subgroups across the same clinical domains as above: at the first manic episode, BD patients with the specifier ‘severe with psychotic features’ (the majority of cases) showed, by definition, greater severity, but overlapping extents, of positive symptoms than BD patients without this specifier; furthermore, these two subgroups were indistinguishable in terms of negative symptoms, neuropsychology, neurology, premorbid adjustment and quality of life (Owoeye et al. Reference Owoeye, Kingston, Scully, Baldwin, Browne, Kinsella, Russell, O’Callaghan and Waddington2013). The same question can be asked for major depressive disorder (MDD) in relation to presence or absence of the DSM-IV specifier ‘severe with psychotic features’. While CAMFEPS, like essentially all first-episode studies, does not ascertain and incept cases of MDD in the absence of psychosis, review of the available evidence suggests a similar conclusion (Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019).
Gene–environment interface
Current theory posits in SZ a gene–environment interaction process that may take many forms, from the heterogeneity of pathobiological processes, through subsequently converging pathways, to a more homogeneous, down-stream pathobiological process (Tost & Meyer-Lindenberg, Reference Tost and Meyer-Lindenberg2012; Howes et al. Reference Howes, McCutcheon, Owen and Murray2017). Among many attendant challenges are how to specify the relevant process(es) and if/how they might generalise to psychotic illness occurring under differing diagnoses (European Network of National Networks Studying Gene-Environment Interactions in Schizophrenia, 2014).
Classical research and literature on the origins of psychotic illness derive primarily from studies of SZ in twins; studies on concordance for SZ among monozygotic versus dizygotic twin pairs constitute the naturalistic experimental paradigm par excellence for partitioning the relative roles of genetic versus common/unique (non-shared) environmental factors in the origin(s) of the disorder. In the most recent and most extensive study, involving all pairs in the Danish Nationwide Twin Register, comparison of monozygotic versus dizygotic concordance rates indicates that additive genetic effects account for 78.9% and unique environmental effects for 21.1% of variance in liability for SZ, in elaboration of the classical literature (Hilker et al. Reference Hilker, Helenius, Fagerlund, Skytthe, Christensen, Werge, Nordentoft and Glenthoj2017). While the same approach indicates generally similar partitioning of variance in liability for genetic versus unique environmental effects in BD (Craddock & Sklar, Reference Craddock and Sklar2013), there are only limited data in SA (Cardno & Owen, Reference Cardno and Owen2014) and even less so in MDDP (Domschke, Reference Domschke2013).
Molecular genetics
The impact of genome-wide association studies (GWAS) on our understanding of SZ, BD and an increasing number of other neuropsychiatric disorders continues to be profound (Sullivan et al. Reference Sullivan, Agrawal, Bulik, Andreassen, Borglum, Breen, Cichon, Edenberg, Faraone, Gelernter, Mathews, Nievergelt, Smoller and O’Donovan2018). CAMFEPS and other centres in Ireland have contributed data, via a national dataset assembled and curated by colleagues at Trinity College Dublin, to several recent global meta-analyses of the molecular genetics of psychotic illness, under the auspices of the Psychiatric Genomics Consortium and the Brainstorm Consortium.
Among the studies to which CAMFEPS contributed was the then-largest GWAS of SZ undertaken, which identified 108 independent genetic loci among 128 loci associated with risk for the disorder among 36,989 cases and 113,075 controls (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). However, this finding, though of enduring import, should no longer be considered in isolation. CAMFEPS has also contributed to a more recent GWAS meta-analysis of 265,218 patients having one of 25 neuropsychiatric disorders and 784,643 control participants, which found that psychiatric disorders share an unexpected degree of common genetic risk: for example, genes associated with risk for SZ are also associated, to varying extents, with risk for BD, MDD, autism spectrum disorder, attention-deficit/hyperactivity disorder, obsessive-compulsive disorder and anorexia nervosa; in contrast, neurological disorders such as epilepsy, stroke, Parkinson’s disease, migraine and multiple sclerosis appear more genetically distinct (Brainstorm Consortium, 2018).
More specifically, CAMFEPS has also contributed to a recent GWAS meta-analysis that identified 32 independent genetic loci among 114 loci associated with risk for a single diagnostic phenotype obtained by combining 33,426 cases of SZ with 20,129 cases of BD in comparison with 54,065 controls (Bipolar Disorder and Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2018); this study also identified polygenic risk scores for SZ that are associated with psychotic symptoms in BD and polygenic risk scores in BD that are associated with manic symptoms in SZ, together with a small number of loci that differentiated between these diagnostic phenotypes.
These GWAS studies elaborate the clinical challenges to diagnostic orthodoxy in, and the alternative dimensional construct for, psychotic illness identified above (see Epidemiology, Diagnostic instability and Clinical characteristics) by indicating SZ and BD to be neither independent nor the same but, rather, sharing particular symptom dimensions that can be reflected in genetic architecture (Bipolar Disorder and Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2018). However, they throw less light on putative mechanisms of gene–environment interactions.
Social milieu
Though genetics clearly holds’ ‘centre stage’ in relation to the origins of psychotic illness, the very small extent of risk associated with each genetic locus (other than extremely rare copy number variations; Sullivan et al. Reference Sullivan, Agrawal, Bulik, Andreassen, Borglum, Breen, Cichon, Edenberg, Faraone, Gelernter, Mathews, Nievergelt, Smoller and O’Donovan2018) and robust evidence for unique environmental effects (see Gene–environment interface) requires that non-genetic factors continue to receive appropriate investigation. The epidemiology of psychotic illness, deriving primarily from studies in SZ, now encompasses a wide range of characteristics and exposures that involve all phases of the illness, from the intrauterine environment, through infancy, childhood and ‘onset’, to adulthood (McGrath, Reference McGrath2005; Brown, Reference Brown2011; Kirkbride et al. Reference Kirkbride, Errazuriz, Croudace, Morgan, Jackson, Boydell, Murray and Jones2012; Owen et al. Reference Owen, Sawa and Mortensen2016); these exposures may act independent of genetic risk, or, more likely, via gene–environment interactions (European Network of National Networks Studying Gene-Environment Interactions in Schizophrenia, 2014). The epidemiology of BD has also received extensive study (Bortolato et al. Reference Bortolato, Kohler, Evangelou, Leon-Caballero, Solmi, Stubbs, Belbasis, Pacchiarotti, Kessing, Berk, Vieta and Carvalho2017; Vieta et al. Reference Vieta, Berk, Scxhulze, Carvalho, Suppes, Calabrese, Gao, Miskowiak and Grande2018), though only occasionally from a systematically trans-diagnostic perspective (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005; Kirkbride et al. Reference Kirkbride, Errazuriz, Croudace, Morgan, Jackson, Boydell, Murray and Jones2012), while the epidemiology of SA and particularly of MDDP has received considerably less attention (Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019).
Among environmental factors associated with increased incidence of SZ, one currently attracting renewed attention is that of neighbourhood-level social factors, particularly greater social deprivation and lower social capital-cohesion/fragmentation (O’Donoghue et al. Reference O’Donoghue, Roche and Lane2016; Radua et al. Reference Radua, Ramella-Cravaro, Ioannidis, Reichenberg, Phiphopthatsanee, Amir, Yenn Thoo, Oliver, Davies, Morgan, McGuire, Murray and Fusar-Poli2018). These factors are closely associated with urbanicity and ethnicity/in-migration, each of which has been related consistently with risk for SZ (Kelly et al. Reference Kelly, O’Callaghan, Waddington, Feeney, Browne, Scully, Clarke, Quinn, McTigue, Morgan, Kinsella and Larkin2010; Castillejos et al. Reference Castillejos, Martin-Perez and Moreno-Kustner2018; Radua et al. Reference Radua, Ramella-Cravaro, Ioannidis, Reichenberg, Phiphopthatsanee, Amir, Yenn Thoo, Oliver, Davies, Morgan, McGuire, Murray and Fusar-Poli2018). As the region in which CAMFEPS is embedded is characterised by no major urban centres or ethnic diversity, with the vast majority of cases being white Irish, it constitutes an important rural resource for investigating social deprivation and lower social capital-cohesion/fragmentation at a neighbourhood level in the absence of urbanicity or associated issues relating to ethnicity/in-migration. Furthermore, the majority of studies to date have not been able to disentangle reliably whether these socioeconomic factors are most important in relation to the neighbourhood where the case was born [in accordance with a developmental model for psychosis (Waddington et al. Reference Waddington, Hennessy, O’Tuathaigh, Owoeye, Russell, Brown and Harrison2012; Weinberger, Reference Weinberger2017)] versus where he/she experienced the ‘onset’ of their psychotic illness [as a proximal ‘trigger’, perhaps interacting with developmental risk (Davis et al. Reference Davis, Eyre, Jacka, Dodd, Dean, McEwen, Debnath, McGrath, Maes, Amminger, McGorry, Pantelis and Berk2016)]; CAMFEPS has been able to address this issue also.
Across the 155 Electoral Divisions constituting the CAMFEPS region, increase in risk for psychosis (SZ + BD + MDDP pooled) was related to increase in neighbourhood deprivation at ‘onset’, subject to some differences of detail between ecological analysis and multilevel modelling; while ecological analysis indicated an additional relationship to increasing neighbourhood social fragmentation, this was not evident in multilevel modelling on incorporating the highly correlated factor of material deprivation and there were no relationships to extent of neighbourhood rurality (Omer et al. Reference Omer, Kirkbride, Pringle, Russell, O’Callaghan and Waddington2014). In contrast, an increase in risk for psychosis was related to a decrease in neighbourhood rurality at birth, in the absence of urbanicity; while there were no relationships to neighbourhood social fragmentation or deprivation at birth, older age at ‘onset’ (as age at first presentation) was associated prominently with lower parental social class (as paternal occupation at birth of the case; Omer et al. Reference Omer, Finnegan, Pringle, Kinsella, Fearon, Russell, O’Callaghan and Waddington2016).
These findings from CAMFEPS indicate effects of neighbourhood- and individual-level socioeconomic factors on risk for psychosis within a rural environment and these relationships differ between neighbourhoods at ‘onset’ and at birth. Thus, such findings are not confined to large urban settings and apply not only across the urban–rural continuum but also across gradations of rurality. One influential theory posits that genetic and environmental factors interact to sensitise the dopaminergic system so that it is vulnerable to acute stressors, leading to progressive dopaminergic dysregulation and the ‘onset’ of psychosis (Howes & Murray, Reference Howes and Murray2014; Howes et al. Reference Howes, McCutcheon, Owen and Murray2017). It appears that these stressors may act diversely from the pre-/perinatal period through to proximity to ‘onset’. Furthermore, a gradient of socioeconomic position appears to influence delay in presentation to mental health services and initiation of treatment.
Developmental pathobiology
While the neurodevelopmental model continues to hold ‘centre stage’ in relation to schizophrenia (Waddington et al. Reference Waddington, Hennessy, O’Tuathaigh, Owoeye, Russell, Brown and Harrison2012; Weinberger, Reference Weinberger2017), controversy endures regarding the extent to which BD might also have developmental origins (Demjaha et al. Reference Demjaha, MacCabe and Murray2012; Murray et al. Reference Murray, Bhavsar, Tripoli and Howes2017; Parellada et al. Reference Parellada, Gomez-Vallejo, Burdeus and Arango2017). While genetic risk for SZ is commonly interpreted in terms of developmental dysregulation that interacts with environmental adversities (Birnbaum & Weinberger, Reference Birnbaum and Weinberger2017; Howes et al. Reference Howes, McCutcheon, Owen and Murray2017), clarification would be facilitated by a ‘hard’ biological index of a developmental abnormality. Anatomical dysmorphologies indicate developmental disruption during early foetal life and are over-represented in SZ (Waddington et al. Reference Waddington, Brown, Schaefer, Goetz, Bresnahan and Susser2008; Xu et al. Reference Xu, Chan and Compton2011). Among these, craniofacial dysmorphologies bear the closest embryological relationship to brain dysmorphogenesis (Marcucio et al. Reference Marcucio, Young, Hu and Hallgrimsson2011, Reference Marcucio, Hallgrimsson and Young2015). Thus, CAMFEPS has contributed to studies in which we have applied 3D laser surface imaging and geometric morphometrics to resolve the topography of craniofacial dysmorphology in SZ (Hennessy et al. Reference Hennessy, Baldwin, Browne, Kinsella and Waddington2007) and show that in BD this appears more similar to than different from dysmorphology evident in SZ (Hennessy et al. Reference Hennessy, Baldwin, Browne, Kinsella and Waddington2010). These findings complement evidence for considerable overlap in genetic risk for SZ and BD (see Molecular genetics) by indicating shared disruptive events operating over early foetal life.
Functional outcome and service models
Though these intricacies of psychopathology, diagnosis, developmental pathobiology, genetics and gene–environment interaction are fundamental to our understanding of psychotic illness, it is functional outcome and the model of care applied to optimise such outcome that are the primary concerns of patients, their families and health service providers.
While functionality and quality of life are well recognised as critical indices of outcome in all medical conditions and have been widely studied in relation to psychotic illness, this literature derives primarily from studies in SZ, less so in SA and BD, with MDDP receiving little attention (Lally et al. Reference Lally, Ajnakina, Stubbs, Cullinane, Murphy, Gaughran and Murray2017; Santesteban-Echarri et al. Reference Santesteban-Echarri, Paino, Rice, Gonzalez-Blanch, McGorry, Gleeson and Alvarez-Jimenez2017; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019). Therefore, we have systematically compared functional outcome, quality of life and service engagement across these four diagnoses during a 6-year follow-up of the first 202 cases in the CAMFEPS cohort (Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, O’Callaghan and Waddington2018): positive psychotic symptoms were most prominent in SZ and SA, and less prominent in MDDP and BD; negative symptoms, impaired functioning and reduction in objectively determined quality of life were most prominent in SZ, intermediate in SA and less prominent in MDDP and BD; in contrast, subjectively determined quality of life was indistinguishable across diagnoses. Service engagement was lowest for SZ, intermediate for SA and BD, and highest for MDDP. However, these measures showed considerable overlap across diagnoses, with no quantitative points of rarity between them. Indeed, on pooling all cases of SZ, SA, BD and MDDP into a single psychotic composite, an increasingly heuristic practice (see e.g. Molecular genetics), higher general (i.e. less diagnostically specific) psychopathology, together with higher negative but not positive psychopathology, and lower service engagement predicted poor functioning in association with lower educational attainment, never having married and living in unsupported living conditions (Kingston et al. Reference Kingston, Scully, Browne, Baldwin, Kinsella, O’Callaghan and Waddington2018).
These findings relate to studies carried out over a period during which the underlying model of mental health care provision, in which CAMFEPS was embedded, has been progressively revised. This journey, involving primary care liaison and use of home-based treatment as an alternative to hospital admission as central to the delivery of health services (McCauley et al. Reference McCauley, Rooney, Clarke, Carey and Owens2003; Nwachukwu et al. Reference Nwachukwu, Nkire and Russell2014), is described in detail in a companion article; the challenges encountered and opportunities afforded in the course of these innovations are instructive, both locally and nationally, when considered in the context of current professional, public and political debate (Russell et al. Reference Russell, Nkire, Kingston and Waddington2019).
Early intervention
At the ‘core’ of priorities for Ireland in relation to psychotic illness is the National Clinical Programme for Early Intervention in Psychosis, which is driven by assumptions of dual import:
Firstly, that longer duration of untreated psychosis (DUP) at first clinical presentation is associated with poorer long-term outcome, such that reduction in DUP through earlier case identification and initiation of treatment can improve outcome (Millan et al. Reference Millan, Andrieux, Bartzokis, Cadenhead, Dazzan and Fusar-Poli2016). DETECT and CAMFEPS-COPE have recently pointed out that opinion on this issue is diverse, ranging from scepticism, through agnosticism, to proselytism (Clarke et al. Reference Clarke, McDonough, Doyle and Waddington2016), but what is the evidence? The most recent meta-analysis indicates that among available interventions (standalone first-episode psychosis services, standalone clinical high risk services, community interventions, healthcare professional training and multi-focus interventions), there is no summary evidence that they are successful in reducing DUP, though a single, standalone clinical high-risk services study was positive but remains to be replicated (Oliver et al. Reference Oliver, Davies, Crossland, Lim, Gifford, McGuire and Fusar-Poli2018). The CAMFEPS dataset, which contains 15 years of measurement of DUP prior to the introduction of an early intervention service, followed by the COPE dataset, which contains 5 years of measurement of DUP following introduction of the COPE early intervention service, will allow further systematic investigation of this issue. In contrast, the most recent meta-analysis indicates that early intervention services are superior to treatment as usual across multiple indices of outcome (treatment discontinuation, psychiatric hospitalisation, improvement at school or in work, and total, positive and negative symptom severity) over periods up to 24 months (Correll et al. Reference Correll, Galling, Pawar, Krivko, Bonetto and Ruggeri2018). This suggests that early intervention services may have clinical impact through processes other than or additional to a reduction in DUP. Comparisons between the CAMFEPS and COPE datasets will allow further systematic investigation of the relationships between early intervention and outcome vis-à-vis DUP.
Secondly, as psychotic ideation and associated subclinical psychopathology and dysfunction can be present in young persons across the general population (van Os & Reininghaus Reference van Os and Reininghaus2016) and, in some individuals, may be early harbingers of risk for clinical psychosis (i.e. clinical high risk/at-risk mental state; Fusar-Poli et al. Reference Fusar-Poli, Borgwardt, Bechdolf, Addington, Riecher-Rössler and Schultze-Lutter2013), early detection of and interventions for such individuals may reduce and even prevent transition to a psychotic diagnosis (Millan et al. Reference Millan, Andrieux, Bartzokis, Cadenhead, Dazzan and Fusar-Poli2016; Carpenter, Reference Carpenter2018). The most recent meta-analysis indicates that among diverse interventions at this stage [needs-based interventions (supportive psychotherapy, psychosocial assistance, brief family psychoeducation, non-antipsychotic medications), cognitive behavioural therapy, integrated psychological interventions, family-focussed therapy, and pharmacological interventions (currently licensed medications, experimental pharmacotherapies, nutritional supplements, or placebo)] no specific intervention is superior to any other (Davies et al. Reference Davies, Cipriani, Ioannidis, Radua, Stahl, Provenzani, McGuire and Fusar-Poli2018).
Experience in these areas in Ireland is limited but evolving under the National Clinical Programme for Early Intervention in Psychosis. Given the challenges to date in implementing this programme, it may be helpful to note that the first two early intervention services have their origins in two substantive and complementary first-episode psychosis studies: the evolution of the St. John of God Hospitaller Ministries/Cluain Mhuire Community Mental Health Services study (Browne et al. Reference Browne, Clarke, Gervin, Waddington, Larkin and O’Callaghan2000) into DETECT (Renwick et al. Reference Renwick, Gavin, McGlade, Lacey, Goggins, Jackson, Turner, Foley, McWilliams, Behan, Lawlor, Cullen and O’Callaghan2008) and the evolution of CAMFEPS (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O’Callaghan and Waddington2005) into COPE (Nkire et al. Reference Nkire, Sardinha, Nwosu, McDonough, De Coteau, Duffy, Waddington and Russell2015). These issues are considered in greater detail in a companion article (Russell et al. Reference Russell, Nkire, Kingston and Waddington2019)
Synthesis
Perhaps the greatest challenge is increasing congruence between studies of psychotic illness across conventional diagnoses and psychotic ideation across the population at large. By way of example, the common manifestation of both psychosis and depression in a given individual can lead to a variety of diagnoses by DSM-IV criteria (the version of DSM on which the great majority of published studies are based, with DSM-5 containing only minor variations): (a) depression in schizophrenia is so widely recognised that its presence may not engender any concern as to a primary diagnosis of Schizophrenia; (b) occasionally, when depressive symptoms in schizophrenia meet criteria for a major depressive episode, an additional diagnosis of Depressive disorder not otherwise specified may be made together with a diagnosis of Schizophrenia; (c) where psychotic symptoms are manifested both concurrently with a major depressive episode and in the absence of prominent mood symptoms, the diagnosis may be SA; (d) where a major depressive episode includes concurrent manifestation of psychotic symptoms but psychosis is not evident in the absence of prominent mood symptoms, the diagnosis may be MDD, severe with psychotic features; (e) the subsequent manifestation of a manic or hypomanic episode may lead to diagnostic revision to bipolar I or II disorder, depressed, severe with psychotic features; (f) where the interplay of psychotic and mood symptoms is unclear, is the subject of contradictory information or does not allow determination of whether it is primary or secondary, the diagnosis may be Psychotic disorder not otherwise specified. To what extent can we be confident that such intricacies and subtleties of diagnosis constitute real and impactful differences, or else reflect mercurial designations within a rich milieu of developmental outcomes at the level of mentation, behaviour and function?
Psychotic illness and associated risk factors across the gene–environment interface are disrespectful to conventional diagnostic categories, with any differences between them being quantitative rather than qualitative, in the presence of substantive overlap and absence of points of rarity between them. Thus, our efforts to impose some underlying order through diagnostic categories constitute, in reality, a series of arbitrary boundaries. Building on previous conceptualisations (Owen, Reference Owen2014; Brainstorm Consortium, 2018; Waddington et al. Reference Waddington, Kingston, Nkire, Russell, van Os, Tamminga, Ivleva and Reininghaus2019), evidence increasingly suggests a continuum of developmental outcomes characterised by variations in pathobiological processes, psychopathological dimensions and functional characteristics. Furthermore, these outcomes, including what we term ‘psychosis’, appear in continuity or intersection with the limits of ‘normal’ human mentation and functioning, with attendant implications for how we conceive and optimally structure provision of mental health services (Arango et al. Reference Arango, Diaz-Caneja, McGorry, Rapoport, Sommer, Vorstman, McDaid, Marion, Serrano-Drozdowskyj, Freedman and Carpenter2018; McGorry et al. Reference McGorry, Hartmann, Spooner and Nelson2018).
Acknowledgements
We thank the Stanley Medical Research Institute for funding the first 10 years of CAMFEPS; Clinical Research Fellows/Registrars Drs. Paul Scully, John Quinn, Maria Morgan, Patrizia Baldwin, David Browne, Tara Kingston, Olabisi Owoeye and Nnamdi Nkire who so capably operated CAMFEPS for 15 years and carried out and authored the studies described in this article; and all our colleagues in Cavan-Monaghan Mental Health Service who supported and fostered CAMFEPS and its transition to COPE. We dedicate this article to the memory of the late Dr. Hanafy A. Youssef, Consultant Psychiatrist, St. Davnet’s Hospital, Monaghan, who, through his dedication to and enthusiasm for mental health research, fostered the initial clinical collaboration between Cavan-Monaghan Mental Health Service and the Royal College of Surgeons in Ireland that led to CAMFEPS.
Conflict of interest
The authors have no conflicts of interest to declare.
Ethical standards
All study protocols relating to the Cavan-Monaghan First Episode Psychosis Study were approved by the Research Ethics Committees of the North Eastern Health Board (and, following restructuring, of the Health Service Executive Dublin North East Area), St. Patrick’s University Hospital, Dublin, St. John of God Hospital, Co. Dublin, and the Central Mental Hospital, Dublin, The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Financial support
The first 10 years of the Cavan-Monaghan First Episode Psychosis Study were supported by the Stanley Medical Research Institute.
