Establishing a clinical service to prevent psychosis: What, how and when? Systematic review

The first rate-limiting step to successfully translate prevention of psychosis in to clinical practice is to establish specialised Clinical High Risk for Psychosis (CHR-P) services. This study systematises the knowledge regarding CHR-P services and provides guidelines for translational implementation. We conducted a PRISMA/MOOSE-compliant (PROSPERO-CRD42020163640) systematic review of Web of Science to identify studies until 4/05/2020 reporting on CHR-P service configuration, outreach strategy and referrals, service user characteristics, interventions, and outcomes. Fifty-six studies (1998–2020) were included, encompassing 51 distinct CHR-P services across 15 countries and a catchment area of 17,252,666 people. Most services (80.4%) consisted of integrated multidisciplinary teams taking care of CHR-P and other patients. Outreach encompassed active (up to 97.6%) or passive (up to 63.4%) approaches: referrals came mostly (90%) from healthcare agencies. CHR-P individuals were more frequently males (57.2%). Most (70.6%) services accepted individuals aged 12–35 years, typically assessed with the CAARMS/SIPS (83.7%). Baseline comorbid mental conditions were reported in two-third (69.5%) of cases, and unemployment in one third (36.6%). Most services provided up to 2-years (72.4%), of clinical monitoring (100%), psychoeducation (81.1%), psychosocial support (73%), family interventions (73%), individual (67.6%) and group (18.9%) psychotherapy, physical health interventions (37.8%), antipsychotics (87.1%), antidepressants (74.2%), anxiolytics (51.6%), and mood stabilisers (38.7%). Outcomes were more frequently ascertained clinically (93.0%) and included: persistence of symptoms/comorbidities (67.4%), transition to psychosis (53.5%), and functional status (48.8%). We provide ten practical recommendations for implementation of CHR-P services. Health service knowledge summarised by the current study will facilitate translational efforts for implementation of CHR-P services worldwide.


Introduction
The clinical high risk for psychosis (CHR-P) paradigm 1 represents one of the most established preventive approaches in clinical psychiatry 2 . It originated in Australia around 25 years ago 3 and since then, it has progressively gained importance 4 . CHR-P individuals are young and accumulate risk factors for the disorders [5][6][7] , that lead to functional impairments 8 and attenuated psychotic symptoms 9 . Because of these features, these individuals seek help 10 at specialised CHR-P mental health services. The detection 11 , prognostic assessment 12 and preventive treatment [13][14][15][16] in CHR-P individuals 15 have the potential to maximize the benefits of early interventions in psychosis 17,18 . A recent evidence-based summary by the European College of Neuropsychopharmacology Network for the Prevention of Mental Disorders and Mental Health Promotion 19 indicated that the first rate-limiting step to prevent psychosis is to establish specialised CHR-P services 20 . Accordingly, several CHR-P services have been implemented worldwide, as recently mapped by the International Early Psychosis Association (IEPA 21 : https://iepa.org.au/list-a-service).
Despite these progresses, health service research in this field has been fragmented to the point that the characteristics ("what") of a CHR-P service per se are poorly defined. As CHR-P services expand globally 21 , it becomes essential to synthetize the core CHR-P health service features that have been implemented in real-world scenarios. While a CHR-P clinic can be broadly defined as a "multidisciplinary community mental health service that provides treatment and support to people at high risk of developing psychosis" (page 16 from NHS England 22 ), this definition remains elusive. Similarly, there is no clear guidance on "how" to integrate different service components. The three main models for delivering CHR-P services include the "stand-alone", "hub and spoke", and "integrated" models 22 . While the standalone model works independently from other more generic community mental health teams, in the "hub and spoke" model, dedicated team workers ("spokes") are based within more generic community teams to route patients needing more intensive services to the central "hub" 23 . In an integrated model, the CHR-P service is completely integrated into the community mental health care. In addition, these models can be combined within broad mental health services enhancing transitional primary care platforms across adolescents and young adults 24 . The additional limitation of knowledge is that the timing ("when") for preventive approaches, which is reflected by CHR-P entry age criteria is uncertain. While this has been typically set for young people aged 8-40 4 years, more recent lifespaninclusive approaches for those under the age of 25 (0-25 years) 25 models have been piloted.
While previous systematic reviews have addressed these issues for services taking care of patients with a first episode of psychosis 26,27 , CHR-P research has remained mostly "academic" and did not systematically address real-world service characteristics such as: service configuration, outreach strategy and referrals, service user characteristics, interventions, and outcomes. The current systematic review summarizes, for the first time, evidence on these domains to inform the real-world implementation (i.e., what, how, and when) of CHR-P clinical services worldwide.

Search strategy and selection criteria
A multistep systematic literature search strategy was used to identify relevant articles by two independent researchers (GSP, AE). First, the Web of Science database (Clarivate Analytics) was searched, incorporating the Web of Science Core Collection, BIOSIS Citation Index, KCI-Korean Journal Database, MEDLINE, Russian Science Citation Index, and SciELO Citation Index as well as Cochrane Central Register of Reviews, and Ovid/PsychINFO databases, as well as the OpenGrey database (for grey literature) from inception until 4th May 2020, with no restrictions on language. The following search terms were applied: ("risk" OR "prodrom*" OR "ultra-high risk" OR "clinical high risk" OR "attenuat*" OR "high risk" OR "genetic high risk" OR "risk syndrome" OR "at risk mental state" OR "at-risk mental state" OR "ARMS" OR "risk of progression" OR "schizophrenia" OR "schizoaffective disorder" OR "schizophreniform disorder") AND ("psychosis") AND ("prevention" OR "intervention" OR "early intervention" OR "referral" OR "assessment" OR "service" OR "clinical service" OR "psychiatric service" OR "implementation" OR "care pathways"). The references of the articles identified in previous reviews and relevant commentaries and the references from the included studies were manually searched to identify additional relevant records. Abstracts were screened, and potential full texts were assessed against inclusion and exclusion criteria.
The inclusion criteria were a) being an original study published in international databases or in the grey literature, b) describing clinical services for individuals in a CHR-P state as defined according to established instruments: Comprehensive Assessment of At-Risk Mental States (CAARMS 3 ), Structured Interview for Psychosisrisk Syndromes (SIPS 29,30 ), Bonn Scale for the Assessment of Basic Symptoms (BSABS 31 ), Basel Screening Instrument for Psychosis (BSIP 32 ), Schizophrenia Proneness Instrument 33 -Adult (SPI-A) and Child and Youth (SPI-CY) version -, Positive and Negative Syndrome Scale (PANSS 34 ), Scale for the Assessment of Negative Symptoms (SANS 35 ), Brief Psychiatric Rating Scale (BPRS 36 ) and Early Recognition Inventory (ERIraos 37 ), c) providing information on any of the following: service configuration, outreach strategy and referrals, service user characteristics, interventions and outcomes, d) providing relevant information without any restrictions on language, sex, age, or ethnicity. The exclusion criteria were a) nonoriginal studies such as abstracts, conference proceedings, study protocols, reviews, guidelines, b) studies with a primary research focus (e.g., research networks) and lacking description of CHR-P clinical services, c) studies describing clinical services for conditions other than the CHR-P or services without a CHR-P component, d) national or regional survey studies with aggregate data and lacking a service-specific description.

Descriptive measures and data extraction
Independent researchers (GSP, AE, MC) extracted data from the included studies; discrepancies were resolved through consensus, consulting a senior researcher (PFP). The variables included were those necessary to describe "what, how and when" to implement CHR-P services. These variables were grouped according to health service domains previously established (beyond general data such as first author, year of publication, name of the CHR-P service, country) 38 : (i) service configuration: continent, service set-up date, population in the catchment area, type of service, professionals involved, (ii) outreach strategy and referrals: outreach activities-measured using an adapted version of the Longitudinal Youth-At-Risk Study (LYRIKS) study 39 classification-, referral sources (iii) service user characteristics: sociodemographic characteristics, CHR-P assessment, CHR-P subgroups (defined as in previous studies) 40 , minimum and maximum age inclusion criteria and service use age range, comorbidities and employment (iv) interventions: type of intervention (non-pharmacological vs. psychopharmacological) and duration of service provision and (v) outcomes: type of outcomes monitored and outcome instruments. Furthermore, we reported quality assessment (see below).

Data analysis Systematic review
All the studies were systematically summarized in tables reporting on various health service domains: service configuration, outreach strategy, and referrals (Table 1), service user characteristics (Table 2), interventions and outcomes (Table 3). We complement this with descriptive analysis of common operational and clinical challenges. An online tool (https://www.maptive.com) was used to create a graphical representation of the geographical distribution of the CHR-P services included in the review.

Quality assessment
We adapted the mixed Methods Appraisal Tool (MMAT) 41,42 questions for non-randomized clinical studies due to the heterogeneity expected in the included studies to assess the quality of the included studies (eMethods 1), considering the content and characteristics of the studies according to our inclusion criteria.

Database
The literature search yielded 12,130 citations, which were screened for eligibility. Two hundred and twentyone full-text articles were evaluated for eligibility, and 165 were excluded. In total, 49 studies reporting information on individual CHR-P services (eTable 2), and seven multisite studies (eTable 3) were selected (PRISMA, Fig. 1). All CHR-P services (100%) used validated assessment instruments and no studies were excluded for this reason. The final pool of 56 included studies were published between the years 1998 and 2020. The total sample  [43][44][45], and the Swiss Early Psychosis Project (SWEPP 46 ). Two additional multisite studies report on five centres operating under the Italian Departments 47 and six CHR-P services in Canada 48 .

Service configuration
The CHR-P services were located mostly in Europe (58.8%), followed by North America (25.5%), Australia (7.8%), Asia (5.9%), and South America (2.0%; Fig. 2 and Table 1). The first program to be implemented was the Personal Assessment and Crisis Evaluation (PACE) clinic in 1994, in Melbourne 49 , and the most recent one the City & Hackney At-Risk Mental State Service (HEADS UP) in Table 1 continued

Outreach strategy and referrals
Outreach activities and audiences were highly variable (Table 1). Within active strategies, workshops for referral sources were the most frequent (97.6%), often targeting healthcare professionals (85.4%), educational professionals (48.8%), or community organisations (34.1%). Services also approached NGOs and community services Includes two services that enrolled "adolescents and adults" without further specification. b Number of services providing data for the service user characteristics as % CHR-P individuals.
c Diagnosis according to DSM or ICD criteria stablished using either structured interviews or clinical interviews. APS attenuated psychosis symptoms, BLIPS brief limited intermittent psychotic symptoms, BSABS Bonn scale for the assessment of basic symptoms, BSIP Basel screening instrument for psychosis, CAARMS comprehensive assessment of atrisk mental states, ERIraos-CL early recognition inventory retrospective assessment of symptoms checklist, GRD genetic risk and deterioration, SIPS structured interview for psychosis-risk syndromes, SPI-A schizophrenia proneness instrument (adults version), SPI-CY schizophrenia proneness instrument (child and youth version).  [50][51][52] , posters 53 , articles in professional journals and local newspapers 51,54 , presentations in scientific conferences 46 , newsletters 51 , and promotional videos 43,55 .
Most CHR-P services received young people with a putative risk of psychosis from health-related organizations, including both outpatient or community mental health services (90.0%) and general healthcare services (75.0%). Education organisations are also frequent referral sources (65.0%), followed by self (60.0%), family or relatives (60.0%), inpatient mental health services (42.5%), and accident and emergency departments (22.5%). Other referral sources were reported in less than 20% of CHR-P services.

Service user characteristics
The total sample size of service users was of 5637 CHR-P individuals, ranging from 4 56 to 467 57 individuals: most of them were males (% of CHR-P females = 42.8, see Table 2). CHR-P status was most frequently assessed using the Comprehensive Assessment of At-Risk Mental States (CAARMS) (48.6%), followed by the Structured Interview for Psychosis-risk Syndromes (SIPS) (35.1%) and the Early Recognition Inventory retrospective assessment of symptoms checklist (ERIraos-CL) (16.2%; Table 2). The Basel screening instrument for psychosis (BSIP) and basic symptoms instruments were infrequently used (<6% of services). Most services provided treatment starting in adolescence, from the ages 12 to 17 (83.4%). The most frequent minimum inclusion age range was 13-15 years (31.3%); only two services reported the inclusion of children from the age of 8 and 6 46,48,58 . Most services accepted users until 30-35 years (49.0%) or 24-29 years (28.6%). A few services (6.1%) accepted service-users of 65 years or older 56,59 . The most frequent age range (70.6% of services) was 12 and 35 years.
82.6% of the CHR-P individuals fulfilled APS criteria, 10.7% fulfilled BLIPS criteria and 8.5% fulfilled GRD criteria (not mutually exclusive). Baseline comorbid mental disorders were reported in 69.5% CHR-P individuals. Mood disorders were the most common: depressive disorders (42.3%), bipolar disorders (15.5%), and persistent depressive disorder (6.7%). Anxiety disorders were also frequent (24.1%), including social phobia (5.9%) and obsessive-compulsive disorder (OCD) (5.1%). Adjustment disorder appeared in 11.6% of CHR-P individuals. Comorbid personality disorders were present in 15.5% of CHR-P individuals, particularly schizotypal personality disorder (11.0%). Substance use disorders were present in 12.4% of CHR-P individuals. Past history of suicide attempts was present in 10.5% of CHR-P subjects. Unemployment rate (i.e., neither work nor study) was observed in 36.6% in CHR-P individuals in clinical services.
Outcomes were most commonly evaluated with standard clinical interviews with the service users (93.0%), and more infrequently with psychometric instruments (37.2%). In the latter case, the CAARMS (23.2%) and SIPS (11.6%) were more frequently employed. About 16.3% of CHR-P services evaluated outcomes via electronic health records.

Quality of the included studies
Study quality scores ranged from 1 to 5. The overall mean quality score for included studies reporting on individual services was 3.8 (moderately high quality) on the MMAT scale, with a SD of 1.3 (eTable 2).

Discussion
To our knowledge, this is the first systematic review to comprehensively summarize the evidence from real-world implementation of CHR-P clinical services. This review encompasses 56 studies describing a total of 51 services for CHR-P individuals.
Consistent with recent surveys of CHR-P services 21,47,78 , there is great diversity in how clinical services have been implemented in real-world scenarios, across all aspects of service delivery: 1) service configuration, 2) outreach strategy and referrals, 3) service user characteristics, 4) interventions and 5) outcomes. We discuss these points, while also mentioning common challenges. The evidence summarised will then be used to operationalise ten empirical recommendations for overcoming these challenges and facilitating the real-world implementation of CHR-P services (Table 4).
In terms of service configuration, several CHR-P clinical services have been implemented across-at least-15 countries (Fig. 2), covering a catchment area of over 17 M people. Following a period of rapid expansion (2000-2009), new CHR-P services continue to emerge 38,76 . At present, CHR-P services spread across most continents 21 , although they are mostly established in high-income countries. While most CHR-P services are configured as integrated services (80.4%), standalone models of care (19.6%) seem to be associated with high levels of service efficiency 27 . For example, CHR-P standalone services had dropout rates in the range of 12-19.2% 50,77,79-82 compared to 25.4% in integrated services 52 . One possible explanation is that in integrated models of care, healthcare resources are typically diverted towards more severe service users (e.g., first-episode vs. CHR-P patients) 38 . In line with this notion, the actual caseload of CHR-P individuals was minimal (n = 4 out of 239 clients) in some integrated services 56 , and more severe patients had more frequent contacts with these services 73 . Another issue is that standalone services may be physically located outside general psychiatric services, which is preferable to reduce stigmatisation risks 77,81 . Table 4 Ten simple recommendations for real-world implementation of CHR-P service. Service users and their relatives were generally more satisfied with standalone CHR-P services, particularly with clinical contact being outside traditional mental health settings 67 . Conversely, family disengagement was the most significant barrier (71.4%) in integrated services 60 . Likewise, primary care clinicians favoured standalone models of care because of the superior accessibility of the services 67 . Standalone services are more costly to set up in the first year but deliver highest economic savings in the longer term 69 , mainly associated with the improved outcome of the disorder 68 . These considerations are of relevance given that poor financial support and lack of adequate infrastructures are frequently cited barriers for the establishment of standalone CHR-P services outside mental healthcare 21 . Future health service research is expected to consolidate these speculations, as well as to test the efficiency of innovative models of care. For example, although there were no hub and spoke services, this organization design, which arranges service delivery assets into a network, may be particularly promising in this field and fit well with the youth friendly mental health reform which is undergoing in several countries 83 . Based on this evidence we recommend to preferably implement standalone services (Table 4). This review also indicates that the CHR-P clinical services are essentially multidisciplinary, reflecting the complexity of the psychopathological assessment and case formulation 84 . Based on the most frequent professionals involved in CHR-P services, we recommend a minimum team encompassing psychiatrists, clinical psychologists, or counsellors, case managers and nurses (Table 4). Because multidisciplinary work requires adequate articulation and training of staff, a core associated recommendation is to ensure adequate training 85 . National surveys have found lack of specialised training in evidence-based interventions to cause dismal across staff 78,86 . Ensuring proper training is particularly challenging for non-academic services, with less resources and limited organizational support 48 .
In terms of outreach strategy, school, mental health, and physical health practitioners were the core targets of community outreach 43 . We have confirmed high heterogeneity across two main strategies: active (up to 97.6%) and passive (up to 63.4%) outreach. The first strategy involved active efforts to organise workshops more frequently targeting healthcare professionals (85.4%), or service promotion activities in the community (up to 34.1% of CHR-P services) and implementing general public awareness campaigns (36.6%). The second strategy involved passive approaches such as a dedicated online site (63.4%), or printed and other media materials (53.7%). This heterogeneity is likely to reflect diverse culturally sensitive approaches across CHR-P services that led to variable pathways to care. In terms of referrals, most CHR-P services received young people with a putative risk of psychosis from health-related organisations such as mental health services (90.0%) and general healthcare services (75.0%). Implementing an outreach to promote referrals of CHR-P individuals is challenging. In the lack of clear guidance, there is high risk of inefficient use of resources (e.g., staff) and inappropriate referrals that eventually do not meet CHR-P criteria. For example, some CHR-P services reported a high number of inappropriate referrals following intense media campaigns, switching to more focused outreach strategies 49,51,61,79,[87][88][89][90] . At times of financial constraints, the core outreach activities and referral targets summarised in the current study can be used as benchmark to maximise the efficiency of resources when implementing a new CHR-P service. There are also empirical constraints. For example, difficulties in recruiting participants is the most difficult challenge in countries where the CHR-P paradigm is starting to be implemented 76 and in culturally diverse catchment areas 43 . Even in countries with an established CHR-P network like the UK, increasing numbers of referrals following the implementation of new national policies resulted in more dedicated CHR-P services that were needed to manage the referrals 59 . Finally, the type of outreach and referrals determine the accumulation of established risk factors for psychosis 5,7,91 , thus influencing the level of psychosis risk among individuals recruited for undergoing a CHR-P assessment (also termed as pretest risk enrichment) 92,93 . For example, individuals sampled from inpatient units may have accumulated more risk factors for psychosis and therefore present with a higher level of psychosis compared to those sampled from the community. This level of risk enrichment 93,94 , substantially impacts the clinical utility of CHR-P instruments 12 . Accordingly, intense outreach strategies mainly targeting the general population end up diluting the level of pretest psychosis risk 93 , and therefore impeding a clinically meaningful identification of CHR-P individuals 11,95 . In line with recent psychometric guidances 12,20 , we recommend CHR-P outreach to primarily target healthcare agencies to promote referrals from these sources (Table 4). Community outreach and recruitment from the general public should be considered only if adequate risk enrichment strategies can be implemented (for a detailed review see ref. 11 ). For example, prescreening approaches can increase pretest risk enrichment among referrals 21 and was employed by some services 47 .
In terms of service users characteristics, we confirmed that males were relatively more represented than females, in line with the epidemiological gender distribution of psychosis risk 6 . Currently, the vast majority (83.7%) of CHR-P services employ the CAARMS or the SIPS, while basic symptoms instruments failed to enter clinical practice at large scale. This suggests that the harmonisation of these two instruments could deliver a widely used gold standard assessment measure for clinical practice. A rapid response to referrals 62 and flexibility with time and setting of assessments 67 have been found to improve engagement with CHR-P services.
Age intake is a core implicit criterion (along with the help-seeking behaviour) defining the CHR-P state 4,10,20 . The most frequently applied age range (70.6% of services) was of 12-35 years, in line with epidemiological research indicating that the peak of risk is between 15 and 35 years 6 . Empirical research confirms that CHR-P psychometric assessment (e.g., the CAARMS) is valid in young people aged 12 years upwards 64 . This finding also confirms the transitional nature of the CHR-P paradigm that cuts across adults and children and adolescent mental health services 25 . Accordingly, most services provided treatment starting in adolescence (between 13 and 15 years). Conversely, only a few services accepted users beyond 40 years 56,59,96 . The requirement of extending the assessment and care of emerging psychosis in the older people, introduced by national guidelines such as the Access and Waiting Time Standards in the UK 22 is against the evidence that CHR-P instruments are valid up to 40 years 4 . Furthermore, it conflicts with recent mental health reforms that are lowering-as opposed to increasing-the age threshold for preventive approaches to those aged from 0 to 25 years 25 . Based on these findings we recommend that CHR-P services ascertain the at-risk status through the CAARMS or SIPS in both adolescents and young adults (Table 4). This review also indicated that presentation to CHR-P services was associated with frequent comorbid mental health conditions (in particular mood and anxiety disorders 97,98 ) in two-thirds (69.5%) of the individuals, coupled with past history of suicide attempts in about one in ten (10.5%) and unemployment in about one third (36.6%) of cases. We further observed regional heterogeneity in clinical presentation: substance misuse was more prevalent in Western services 38,51,54,60,65 , while non-existent in Japan 52 . These findings recommend that CHR-P services should adopt a broader "transdiagnostic" approach" [99][100][101] , which is cutting across several psychopathological dimensions ( Table  4), given that psychosis onset can occur from preceding mood dysregulation 102 or substance abuse. This recommendation is also relevant for current operationalisations of at-risk syndromes, which require formulating a differential diagnosis between psychosis risk and other psychopathological dimensions such as the SIPS or the DSM-5-APS 2 . Although psychotic experiences are frequent in the general population 103,104 , clinical attenuated psychotic symptoms are infrequent and not normally distributed. Only 0.3% of the general young population meet DSM-5-APS criteria 2,105 .
In terms of interventions, most services (72.4%) provided care for 2 years or less (see outcomes below), with some exceptions 38,48,52,54,55,62,63,72,74,81,87,106 , encompassing clinical monitoring (100%), psychoeducation (81.1%), psychosocial support (73%), family interventions (73%), CBT-based individual interventions (67.6%), group psychotherapy (18.9%), physical health interventions (37.8%), antipsychotics (87.1%), antidepressants (74.2%), anxiolytics (51.6%), and mood stabilisers (38.7%). It appears that CHR-P clinical services currently provide a wide range of psychosocial and biological interventions to meet the clinical needs of CHR-P service users. Clinical monitoring, case management and targeted case management are essential elements of preventive treatment 22 , based on the principles of social psychiatry and the importance of engaging CHR-P individuals with healthcare services 107 . These often included psychoeducation and informing patients about their risk, as done in other preventive approaches in medicine 84 . Despite current guidelines recommending psychological interventions (such as cognitive behavioural therapy) as first-line treatment, about one-third of CHR-P services did not provide them. Evidence to favour psychotherapy over other types of interventions in this population is currently uncertain 13,15,16,20 . Conversely, antipsychotic treatment, which is discouraged by current treatment guidelines, was frequently considered, although typically at low dosages and only when the symptoms were deteriorating. This is consistent with data from global and national surveys of CHR-P services 21,78 that report frequent use of antipsychotic drugs. The relatively frequent use of anxiolytics, antidepressants, and mood stabilizers-which is not considered by current guidelines-can index the transdiagnostic nature of the CHR-P state with frequent affective and anxiety comorbidities. The variety in provision of treatments likely reflects the high clinical heterogeneity of this population and the lack of clear treatment guidelines stratified on their individual needs. For example, current guidelines are not stratified across CHR-P subgroups. Individuals with brief psychotic episodes may be defined through researchbased operationalisations, such as brief and limited intermittent psychotic symptoms (BLIPS) or standard psychiatric classifications including "Acute and Transient Psychotic Disorder" as per ICD-11 or DSM-5 "Brief Psychotic Disorder". There is diagnostic and prognostic overlap across these definitions of brief psychotic episodes 108,109 . Individuals with brief psychotic episodes have the highest risk of developing psychosis 20 -especially when recurrent or presenting with seriously disorganizing or dangerous features- 108,110 . They also display poor clinical outcomes and do not engage with the recommended cognitive behavioural therapy 40,108 , leaving them with unmet need for care 110 . Stratification across these clinical subgroups has been proposed in recent revisions of the CHR-P paradigm 1,20 and should be considered in future clinical guidelines 21 . Because the uncertainty of current evidence is high, we align with the recent recommendations of the European College of Neuropsychopharmacology Prevention of Mental Disorders and Mental Health Promotion Network 19 to still offer needs-based interventions and psychological interventions, titrating the intervention according to the characteristics and risk profile (i.e., transition risk, symptom severity, and functional impairment) 20 as well as the values and preferences of the CHR-P individuals (Table 4) 20 . For example, it seems important to individualise physical health and lifestyle interventions on the needs presented by each service user 60,76 .
In terms of outcomes, surprisingly, persistence of symptoms/comorbidities (67.4%) was measured more frequently than transition to psychosis (53.5%), functional status (48.8%), and remission (41.9%). This likely reflects the efforts of CHR-P services to treat comorbid conditions, aiming for improving recovery, functioning, and quality of life 20 . At the same time, other outcomes such as physical health were collected in only about a third (30.2%) of CHR-P services 111 . CHR-P individuals accumulate genetic and environmental risk factors 20 , including cardiometabolic risk factors as decreased physical activity 112 and high rates of substance use 112 , including tobacco 112 , alcohol 112 , and cannabis 113 . Thus, more attention should be paid to recording the physical health of CHR-P individuals in clinical services 114 . Another domain of improvement includes a more frequent monitoring of service users' satisfaction, which is pivotal to higher engagement and decreased drop-out rates. Furthermore CHR-P services should also more extensively monitor healthcare utilisation (e.g., hospital admissions) 115 and broad outcomes such as mortality rates to better characterise the overall burden of this condition 116 . Future research is needed to standardise a core outcome set for CHR-P research and therefore facilitate collaborative efforts. These initiative should also indicate the assessment measures to be employed to monitor outcomes. Currently, clinical outcomes in CHR-P services are most commonly evaluated with standard clinical interviews (93.0%), and psychometric instruments are more infrequently used (37.2% of cases, most frequently CAARMS or SIPS to evaluate transition to psychosis) 20 . In the future, monitoring broad health outcomes in CHR-P services could leverage electronic health records that can provide real-world, real-time valuable clinical information 11,[117][118][119][120] and that are being increasingly implemented in healthcare providers. As noted above, duration of care including clinical monitoring is currently limited to, most frequently (44.8%), 2 years. However, accumulating evidence has clearly indicated that although the risk of psychosis onset peaks within 2 years 121 , it can increase in the longer term at least until 3-4 years 40,122,123 . In addition, non-transitioning CHR-P individuals can continue to experience functional impairment and symptomatology at 6-years 97 . This confirms that a 2-year service provision is insufficient 21 . As such, we recommend clinical monitoring for outcomes to be implemented for at least 3 years (Table 4). Flexible follow-up after this timepoint can help make more efficient use of clinical resources, while tailoring interventions to users' needs 124 . For example, the clinical follow-up can be extended if service users are still symptomatic or present sociooccupational difficulties 55,62 . Finally, CHR-P services should be prepared to collect information and target outcomes other than psychosis such as recovery, physical health outcomes, service users' satisfaction, functioning, and quality of life 20,124 . Harmonisation of core outcome set for CHR-P services is a clinical research priority for the future. Several national and regional networks of CHR-P services started to emerge during the decade of 2010-2019 (e.g., EUGEI, PRONIA, PSYSCAN, NAPLS, PNC, HARMONY, PRONET, and STEP) and may facilitate this enterprise, allowing services to leverage best practices and expertise, increasing lobby capacity and enhancing collaborative efforts 38,44,46 . International clinical research infrastructures have also been developed such as the European College of Neuropsychopharmacology Network for the Prevention of Mental Disorders and Mental Health Promotion (ECNP PMD-MHP) 19 . These initiatives will introduce several innovations in the CHR-P field, encompassing personalised prediction of outcomes and individualised interventions, digital screening for improving detection of psychosis risk and enhancement of transdiagnostic research capability within CHR-P services (e.g., preventive interventions for bipolar risk) 125 .

Limitations
The main limitation of this study is that health service information was scattered across services, and that there are no established standards to measure the core domains. This limited the capacity to quantitatively compare the different services with meta-analyses. Future harmonisation efforts in terms of CHR-P healthcare research would be extremely valuable. The database was nonetheless large and sufficiently powered to analyse different factors including service configuration, outreach strategy and referrals, CHR-P service user characteristics, interventions, and outcomes. Another limitation is the limited knowledge provided about the long-term outcomes. Furthermore, our results are based on data from the literature that has been published. However, some clinical services may be running but not publishing details about service configuration, outreach strategy and referrals, service user characteristics, interventions, and outcomes. At the same time, we hope that our review will stimulate the establishment of a global network of CHR-P services with shared clinical research infrastructures 21 . Finally, a considerable amount of studies were carried out in relatively small samples, with only 45.5% services 39,[50][51][52]59,62,68,69,73,75,80,82,[87][88][89][90]106,[126][127][128] including more than 100 CHR-P individuals.

Conclusions
Health service knowledge summarised by the current study will facilitate translational efforts for implementation of CHR-P services worldwide.