Partly amplifying previous observations by Ajnakina, David, and Murray (Reference Ajnakina, David and Murray2018), Moritz, Gaweda, Heinz, and Gallinat (Reference Moritz, Gaweda, Heinz and Gallinat2019) systematically press central trigger points in contemporary early detection conceptual landscape. They highlight four reasons, including decreasing transition rates, why early detection centers for psychosis should be renamed and their treatment targets reconsidered. Transition to psychosis could be a more faceted outcome than usually thought and it is crucial to discuss potential limits of the concept itself and how these limits may have influenced the judgment on the early detection paradigm. Indeed, in the wake of the almost quarter of a century since the conceptualization of ultra-high-risk (UHR) states for psychosis (Yung & McGorry, Reference Yung and McGorry1996), empirical evidences are constantly re-assessed through meta-analytic lens, which converge on three key-points: (1) pre-test risk enrichment (Fusar-Poli et al., Reference Fusar-Poli, Schultze-Lutter, Cappucciati, Rutigliano, Bonoldi, Stahl and McGuire2016a), (2) stratification of risk among UHR subgroups (Fusar-Poli et al., Reference Fusar-Poli, Cappucciati, Borgwardt, Woods, Addington, Nelson and McGuire2016b), and (3) progressive decline of transition rates to psychosis, aka ‘dilution effect’ (Fusar-Poli et al., Reference Fusar-Poli, Cappucciati, Borgwardt, Woods, Addington, Nelson and McGuire2016b; Hartmann et al., Reference Hartmann, Yuen, McGorry, Yung, Lin, Wood and Nelson2016; Simon, Umbricht, Lang, & Borgwardt, Reference Simon, Umbricht, Lang and Borgwardt2014; Yung et al., Reference Yung, Yuen, Berger, Francey, Hung, Nelson and McGorry2007).
The dilution effect and the Janus-faced nature of transition
Whereas pre-test risk enrichment and within-UHR gradient are rather intuitive phenomena, the dilution effect remains rather obscure in its genesis and multi-causality. Several concurrent and non-mutually exclusive factors have been hypothesized: the decrease of duration of symptoms prior to first clinical contact (Yung et al., Reference Yung, Yuen, Berger, Francey, Hung, Nelson and McGorry2007), the possible preventive role of focused interventions [i.e. psychological therapy or antipsychotic (AP) medication] (Nelson et al., Reference Nelson, Yuen, Lin, Wood, McGorry, Hartmann and Yung2016; van der Gaag et al., Reference van der Gaag, Smit, Bechdolf, French, Linszen, Yung and Cuijpers2013) as well as the different clinical intake of recent UHR cohorts in comparison with earlier cohorts (Hartmann et al., Reference Hartmann, Yuen, McGorry, Yung, Lin, Wood and Nelson2016). However, none of these factors, although all empirically plausible and partly substantiated, is explanatory enough or satisfactory at a conceptual level. A better understanding of the dilution effect is mandatory for the field of early detection/intervention, since it would impact its evidence-basis as well as its strategic societal goals. In this perspective the mere criterial approach to define transition to psychosis has been criticized (van Os & Guloksuz, Reference van Os and Guloksuz2017) and a more radical and widespread aspect (i.e. the classical ‘elephant in the room’) could have been often overlooked, namely the prescription of antipsychotics (AP) in UHR samples (Raballo, Poletti, & Carpenter, Reference Raballo, Poletti and Carpenter2019).
In the UHR model, in addition to the criterial transition (based on rating scales), a functional equivalent of transition has been explicitly mentioned as the threshold at which AP treatment would be commenced in common clinical practice (Yung et al., Reference Yung, Phillips, Yuen, Francey, McFarlane, Hallgrem and McGorry2003). Albeit apparently subjective and arbitrary, such threshold is based on the real-world, collegial decision making of the treating staff and reflects a global apprehension of the severity of a clinical status requiring AP medication. Clearly, this indicates the end-point of the UHR state and signals ‘the threshold for onset of a psychotic episode’ (Yung et al., Reference Yung, Yuen, McGorry, Phillips, Kelly, Dell'Olio and Buckby2005).
Just a little bit like Janus Bifrons, conversion to psychosis could have two complementary faces, one looking to escalating positive symptoms (i.e. the criterial psychometric transition), and the other to the complexity of the global clinical trajectory (i.e. the functional equivalent indexed by the therapeutic need of AP medication).
Lost in transition: the glaring evidence of a clinical–conceptual scotoma
The concept of functional transition is crucial considering that AP need is rather frequent in the clinical management of UHR subjects. Of the 33 studies included in a recent meta-analysis addressing psychosis risk stratification (Fusar-Poli et al., Reference Fusar-Poli, Cappucciati, Borgwardt, Woods, Addington, Nelson and McGuire2016b), 24 studies (72.7%) reported exposure to AP during follow-up, yet without justifying it on the basis of worsened clinical severity or considering it a functional equivalent of conversion to psychosis; among these studies, at least seven (above 20%) included in the UHR sample subjects already on AP at the baseline assessment (i.e. in ostensible contradiction with the original UHR definition).
Considering the threshold at which AP medication would be commenced in common clinical practice as a functional equivalent to threshold for onset of psychosis episode, would substantially change the overall transition rates reported in the literature, since basically all the studies merely report criterial (i.e. psychometric) transitions neglecting the functional ones (Table 1). While this mismatch between criterial and functional equivalents of transition may depend on several, context-dependent factors, it is undeniable that its magnitude cannot be further ignored. For example, just considering the 11 studies in which it was possible to clearly extract the absolute number of both criterial transitions and functional transitions, we found that the amount of functional transitions (n = 350) is 215% larger than one of the criterial transitions (n = 163).
AP, antipsychotics; BS, basic symptoms; CAARMS, Comprehensive Assessment of At Risk Mental States; CBT, cognitive behavioral therapy; CHR, clinical high risk; CT, Criterial transition based on CHR instrument; FT, functional transition based on AP exposure during follow-up; SIPS, Structured Interview for prodromal symptoms; UHR, ultra-high risk.
a % of survival rate of transition calculated on the basis of UHR sample at baseline.
b % of exposure to AP as reported in online eTable2 in Supplementary content of Fusar-Poli et al. (Reference Fusar-Poli, Schultze-Lutter, Cappucciati, Rutigliano, Bonoldi, Stahl and McGuire2016a, Reference Fusar-Poli, Cappucciati, Borgwardt, Woods, Addington, Nelson and McGuire2016b).
c Studies detailing the raw numbers of criterial transitions and AP exposure.
d Transition spread: % functional transitions minus % of criterial transitions.
e Ratio between functional transitions and criterial transitions.
Note: Criterial (psychometric) transition:
Functional equivalent of transition = mental state requiring AP therapy, derived from the reported exposure to AP during the follow-up.
Transition spread = Functional minus Criterial transition, is an index of:
(1) Mismatch between positive symptoms based psychometric assessment and real world clinical severity requiring immediate AP treatment as a collegial decision of the treating staff.
(2) When positive, indicates the amount of undetected psychotic-equivalent mental states.
(3) When negative, indicates the proportion of individuals converting to psychosis not receiving appropriate treatment.
Functional v. criterial transition ratio = proportion between individual exposed to AP medication and individuals reaching psychometric criteria for psychosis, is an alternative index of the amount of undetected psychotic-equivalent mental states relative to the psychometric ones.
This might be due to the fact that dimensional rating scales, when evaluating transition from UHR to psychosis, mainly focus on positive symptoms. In clinical practice, however, the need for AP medication is established through a global clinical evaluation of the ongoing mental state, including overall clinical severity as consensually perceived by the treating staff. Such global evaluation would typically consider not only the level of positive symptoms but also concurrent disorganized, negative and accessory psychopathology as well as subtler features of role functioning and quality of life. This is further corroborated by the fact that, if we focus only on UHR criteria, BLIPS appear at higher risk than APS, while if we include symptom severity, negative symptoms, affective symptoms and psychosocial functioning, no clear differences emerge (at the baseline as well as at the follow-up) between UHR subgroups (McHugh et al., Reference McHugh, McGorry, Yuen, Hickie, Thompson, de Haan and Nelson2018), indicating that clinically-meaningful features of UHR mental state reside outside the positive dimension and need full consideration together with the risk/benefit ratio with AP medication (Raballo et al., Reference Raballo, Poletti and Carpenter2019).
Redeeming the elephant in the room and rethinking the transition paradigm
On the basis of this rather disillusioning photograph (i.e. an average AP exposure that is almost the double of the declared conversion rate: about 40% v. 22%, see Table 1) we could either hypothesize that UHR subjects are unduly over-exposed to off-label AP (although they do not reach the psychometric threshold for psychosis) or – in line with the original PACE criteria – that functional equivalents of transition to psychosis (i.e. a mental state requiring immediate AP medication) are systematically ignored. This widespread clinical and conceptual flaw could be involved in the surface-level phenomenon of dilution effect of criterial transition rate, or – at least – contribute substantially to its magnification. Indeed, when considering both criterial and functional transitions, the magnitude of the overall transition to psychosis almost redoubles and the dilution effect may vary substantially. Even more crucially, mainstream prediction models (typically limited to psychometric transitions and counting UHR undergoing AP treatment as simple non-converters) presumably underestimate natural course transition rates. Therefore, while a re-analysis of available datasets is highly recommendable, a new wave of UHR studies with more transparent and systematic reporting of AP exposure is clearly necessary, with AP continuation without apparent criterial transition being rigorously examined. Finally, the possible underestimation of transition to psychosis in UHR subjects should be considered in the current debate on resource allocation within early detection centers.