The Clinical High Risk for Psychosis (CHR-P) paradigm is the predominant approach for primary indicated prevention of psychotic disorders (i.e. interventions that target high-risk individuals who are exhibiting subthreshold but detectable symptoms of psychosis).1 The success of preventive care through the CHR-P paradigm is contingent on both effective early detection strategies and the availability of effective and acceptable preventive interventions. Despite considerable progress in CHR-P research over the past two decades, there remain several challenges and corresponding opportunities for the further advancement of primary indicated prevention.
(1) Early detection as a rate-limiting step
Early detection of at-risk individuals is the first rate-limiting step for effective preventive care, and is especially important given that longer durations of untreated illness in prodromal and full-threshold psychosis are associated with significantly poorer prognosis.2 Whilst the large majority (meta-analytic estimate of 78.3 %) of first episode psychosis (FEP) patients experience a prodrome prior to FEP onset,3 long-standing CHR-P services can only identify 5–12 % of individuals that present with FEP in secondary mental health care services.4 Similarly inefficient detection is observed in FEP services, which detect a small proportion (37.3 %) of FEP cases in the general population. This partially stems from the dependence on gold-standard CHR-P assessments (e.g., the Comprehensive Assessment of At-Risk Mental States [CAARMS]) that are limited by their temporal, financial, and training needs, especially in settings with fewer clinical resources. Several developments are underway to overcome this. Shorter versions of gold-standard, validated CHR-P screening instruments (e.g., mini-CAARMS5) are being developed that show promise for implementing routine, large-scale screening with drastically reduced clinician burden. Similarly, automated digital general-population screening through self-report scales (e.g., the Prodromal Questionnaire) may facilitate the assessment of hard-to-reach and low-prevalence populations, and has been shown to be feasible in a multi-site study.6 Other promising approaches are clinical prediction models including the individualised transdiagnostic risk calculator that has been extensively validated and refined,7 or the Psychosis Polyrisk Score that integrates a range of environmental and genetic risk and protective factors to detect at-risk individuals.8 However, despite their prognostic accuracy and potential clinical utility for at-risk detection, few clinical prediction models in psychiatry have been considered for real-world implementation. To overcome this, future research ought to thoroughly consider ethical guidance for and barriers to implementation and to embed public and patient involvement into the lifecycle of clinical prediction models.9 The accessibility and harmonisation of large real-world datasets (from biobanks and electronic health records [EHRs]) will also be paramount for the continued development of such advanced detection tools.
(2) Need for novel preventive interventions
A recent meta-analysis of randomised controlled trials (RCTs) for preventive interventions in CHR-P10 found limited evidence for their effectiveness in preventing transition compared to treatment-as-usual. This may be a result of clinical improvements in “needs-based interventions” over the past decade, indicating that treatment-as-usual itself may be suitable preventive care - and indeed, CHR-P services offer a broad range of psychological, social, and pharmacological interventions. Future trials ought to compare these different approaches more directly to identify their key active therapeutic components. Nonetheless, there is a need for new preventive interventions, especially those which are less intensive and non-stigmatising given the prognostic uncertainty observed in CHR-P individuals. Antipsychotics are not indicated for preventing psychosis in CHR-P individuals due to their adverse effects, and are not even effective for all individuals with full-threshold psychosis. One promising approach is cannabidiol (CBD) which is non-intoxicating, has a good safety profile, and has exhibited antipsychotic properties in individuals with full-threshold psychosis. Unlike current antipsychotics, CBD acts on the endocannabinoid system; this novel treatment may therefore be used as monotherapy or as a complementary adjunct to existing antipsychotics. Short-term treatment in CHR-P individuals appears to be well-tolerated and effective in reducing symptom severity,11 and a larger-scale, multinational Phase III trial (STEP12) will soon be starting to explore its long-term clinical utility in individuals across the psychosis spectrum. The AMP-SCZ13 program is also accelerating the development of effective treatments for schizophrenia such as the drug MT1988, a novel treatment that targets the cognitive impairment associated with psychosis through modulating nicotinic receptors. Following a successful Phase I trial demonstrating good safety and tolerability of MT1988 in healthy controls, a Phase II placebo-controlled RCT with 150 CHR-P participants is now underway to assess its impact on cognitive, positive, and negative symptoms as well as its safety and tolerability.14
(3) Transdiagnostic nature of prodromal features
CHR-P individuals have a heterogenous clinical presentation, experiencing a wide variety of comorbid mental disorders at baseline, highlighting the importance of taking a transdiagnostic approach to assessment. However, the CHR-P state importantly has specificity for the emergence of psychotic disorders only and not other nonpsychotic disorders.15 Instead, different psychometric instruments are required to enable the early detection and prevention of other mental disorders more comprehensively, such as the Semi-structured Interview for Bipolar At-Risk States (SIBARS)16 which shows promising reliability and validity in detecting individuals at clinical high-risk for bipolar disorders. Such individual psychometric assessments are limited in their ability to improve detection efforts due to their temporal and financial costs; joint transdiagnostic detection has distinct advantages for scaling up preventive care. Given the extensive overlap in the dynamic progression of prodromal features across severe mental disorders, a clinical prediction model has recently been developed and validated for the transdiagnostic detection of psychotic disorders or bipolar disorder in real-world clinical care.17 The model used only routinely collected, EHR-derived data derived without the inclusion of more costly or invasive (e.g., genetic, neurobiological, blood-based) predictors, and demonstrated excellent discrimination performance (C-index = 0.80) and potential clinical utility. Consequently, this clinical prediction model shows promise as an initial screening tool that can be integrated into clinical workflows to effectively identify individuals who require additional in-depth assessment and potential intervention.
(4) Universal vs. primary indicated prevention
Some clinicians and researchers have suggested that universal, population-level prevention approaches should be prioritised over primary indicated prevention through the CHR-P paradigm.18 Their main criticisms are that: (1) the rate of transition in CHR-P individuals is supposedly low at 19 % in two years; (2) there is little evidence of interventions that prevent transition in CHR-P individuals; (3) CHR-P services only detect a small proportion of future FEP cases; and (4) indicated prevention targets few individuals and is less effective than a universal approach (the so-called “prevention paradox”). These concerns are unwarranted given that, firstly, the two-year risk of psychosis in CHR-P is almost 50 times higher than in socio-demographically matched individuals,1 indicative of a very notable risk in this population. Further, the majority of CHR-P individuals either continue to experience non-psychotic comorbidities or do not remit from the CHR-P state – so the CHR-P paradigm remains prognostically important for poor outcomes in this population. Secondly, as aforementioned, the comparable effectiveness of preventive interventions and treatment-as-usual does not mean that the former are ineffective. Alternatively, this highlights the need to consider the effectiveness of “control” needs-based interventions, to conduct larger trials of preventive interventions (for sufficient power to reduce confidence intervals and detect effects), and to continue research into promising avenues. Thirdly, inefficient detection does not mean that CHR-P research should be dismissed; instead, as almost 80 % of FEP cases can be detected by a CHR-P state, the development of novel detection methods (that are underway) should be prioritised to overcome this challenge in early intervention. Fourth, indicated prevention is designed to target few individuals with subtler, but prognostically important, symptoms – and this approach is common in other branches of medicine. Meta-analyses have also estimated that reducing individuals experiencing the CHR-P state by 81 % would reduce the incidence of schizophrenia-spectrum disorders by 10 %.19 Further, despite calls to move to universal approaches, no such approach has been found to be effective in psychosis. CHR-P services also already implement extensive public health initiatives to promote good mental health, in addition to their primary indicated prevention approach.20 It is evident that both the targeted approach of the CHR-P paradigm and universal approaches (e.g., reducing environmental risk factors such as cannabis use; youth engagement strategies) ought not to be framed as oppositional, but instead be pursued together and integrated to maximise the impact of these preventive provisions.1
ConclusionsIn conclusion, despite challenges in early detection and preventive approaches, primary indicated prevention has distinct promise in providing effective care for individuals at-need. Global, longitudinal studies (e.g., PSYSCAN21 and AMP-SCZ13) capturing a wide range of multimodal data are also now underway. These will bring novel insights into the multifactorial mechanistic processes underlying psychosis progression, thereby informing early detection strategies and highlighting potential treatment targets for preventive care.
Ethical considerationsNone.
FundingYL is supported by the UK Medical Research Council (MR/W006820/1). PFP received financial support from the NIHR Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London. PFP is also supported by the Italian Ministry of University and Research, National Recovery and Resilience Plan, project MNESYS. The views expressed here are those of the authors and not necessarily those of the NHS, the NIHR, or the UK Department of Health and Social Care.
None.
