Implementation of early psychosis services in Latin America: A scoping review

The evidence of the effectiveness and cost‐effectiveness of early intervention for psychosis (EIP) services has motivated their implementation worldwide. However, complex interventions of such EIP services require local adaptations to successfully match population needs and cultural differences. Latin America is a heterogenous region where EIP services are progressively being adopted. Our aim is to map such initiatives in the region with a focus on implementation outcomes.


| Early intervention services worldwide
Early intervention for psychosis (EIP) services have consistently been shown to improve outcomes for people with psychosis in terms of symptoms, relapses, employment, suicide attempts and quality of life (Correll et al., 2018;Csillag et al., 2016). Furthermore, from an economic perspective, although a more specialized service might incur in higher costs, current evidence shows that EIP services reduce some costly interventions (Randall et al., 2015) and might be cost-effective across different countries (Aceituno, Vera, Prina, & McCrone, 2019).
The implementation of EIP around the globe is rather heterogenous, regardless of country income and type of health system. For instance, a recent review of EIP services in Europe found high disparities in the region, with countries such as the United Kingdom, Norway, and Denmark providing nationwide programmes, while in France and Germany, EIP services seldom exist (McDaid, Park, Iemmi, Adelaja, & Knapp, 2016). Similarly, in Asia, EIP services have been successfully implemented in Singapore, Hong Kong, Korea, and Japan, developing a regional network of collaboration (Lee, 2013). However, the implementation of EIP services remains at local level in other Asian countries, supported mostly by research endeavours (Lee, 2013;Rangaswamy, Mangala, Mohan, Joseph, & John, 2012).
In low-and-middle-income countries (LMICs), the implementation of these services faces greater problems, such as scarce resources, weak infrastructure, absence of mental health policies, lack of healthcare workforce, and stigma (Saxena, Thornicroft, Knapp, & Whiteford, 2007).
Despite this, the evidence behind EIP services has motivated researchers and early adopters to design ways of implementing such initiatives in less resourced settings. Different proposals have been suggested (see for example Chisholm et al (2008) and Farooq (2013) and more recently, the World Psychiatric Association (WPA) has set up an expert panel to develop guidelines and recommendations for EIP in LMICs, with expected results by mid-2020 (Singh, Javed, & WPA Expert International Advisory Panel for Early Intervention in Psychosis, 2020).
Finally, successful implementations require evidence that goes beyond clinical effectiveness. The emerging field of Implementation Science has expanded the scope of health service research by measuring outcomes at the implementation stage. For instance, Proctor et al (Proctor et al., 2011) propose a taxonomy of these outcomes for mental health, where the acceptability, adoption, appropriateness, feasibility, fidelity, penetration and sustainability of interventions are considered as relevant as symptoms reduction (Peters, Tran, Adam,,, & World Health Organization, 2013;Proctor et al., 2011). This information is highly valuable because it allows planners and policy makers to implement the necessary adjustments that a particular setting might require. Also, it helps to understand the process by which an intervention is incorporated and to foresee eventual challenges other adopters might have encountered. However, thus far, the Early Intervention research agenda has focused mostly on fidelity measurements (Addington et al., 2018), which although relevant, neglect other important aspect of the implementation process.

| Latin American context
Latin America is a vast region comprising approximately 632 million people and 20 countries (United Nations, 2018). This region shares a colonial history and some cultural similarities, such as Latin-rooted language and Christian religion. In addition, several recognized risk factors of psychosis, such as violence, inequality, substance misuse and political and economic instability are especially prevalent in Latin America (Kohn et al., 2005;Pan American Health Organization, 2013).
Such disparities are reproduced at health systems level, which mirrors the state of EIP services development. For instance, in 2011, a literature review identified five EIP initiatives in only two Latin American countries (Brazil and Mexico) (Brietzke, Araripe Neto, Dias, Mansur, & Bressan, 2011). All these programmes were developed alongside research centres or universities, and they had not yet been scaled-up to the national level. However, more recently, new services have been implemented in the region, new evidence has been produced and a collaborative network for the study of psychosis has been created to move the field forward in the region (N. . Hence, we aimed to review the state of implementation of EIP services in Latin America. We sought to map the evidence of such implementation, looking at the characteristics of the services currently taking place, with emphasis on implementation outcomes measures, such as acceptability, feasibility, fidelity, effectiveness, cost-effectiveness and sustainability.

| METHODS
We conducted a scoping review of the literature about the implementation of EIP services in Latin America, following the Preferred Reporting Items for Systematic review and Meta-Analysis extension for Scoping Reviews (PRISMA-ScR) guidelines (Tricco et al., 2018). A scoping review protocol was designed before starting the search based on the approach suggested by Arksey and O'Malley (Arksey & O'Malley, 2005) and the Joanna Briggs Institute (The Joanna Briggs Institute, 2015). The PRISMA-ScR checklist and the protocol are available in the supplementary materials.

| Eligibility criteria
According to current recommendations (Munn et al., 2018;Tricco et al., 2018), inclusion criteria of scoping reviews should be based on the type of participants under study, the concept to be examined and the context where the review is taking place. In our review, we applied the following inclusion criteria: • Participants: Studies describing services for people in their early stages of psychosis, including first-episode psychosis (FEP) and clinical high-risk of psychosis (CHR-P) regardless of age, gender, location or comorbidities. We included studies using standardized methods to define psychosis according to the Diagnostic and Statistical Manual of Mental Disorders or the International Classification of Disease criteria. We did not discard studies including affective psychoses.
• Concept: We selected studies addressing or describing EIP services.
This included any type of intervention, programme or initiative specifically developed for people in their early stages of psychosis.
• Context: we included studies conducted in a Latin American country, according to the definition given by the United Nations Economic Commission for Latin America and the Caribbean. Studies including Latino population but conducted outside the region were excluded.
Anticipating a low number of studies, we kept our review as inclusive as possible; therefore, we included any type of study design, including qualitative analyses, case studies, observational and analytical studies.

| Search strategy
In order to make our search as comprehensive as possible, we included terms related to psychosis and Latin America without language or time restrictions. We used terms such as "psychosis," "psychotic disorders," "schizophrenia," "at-risk mental state," "prodrome" and every Latin American country. See Supplementary Materials for a full list of search terms.

| Selection process
Two reviewers (DA and ED) screened titles and abstract independently and compared them against inclusion criteria. Disagreements were solved by consensus or discussion with a third reviewer (AGV) if consensus was not achieved.

| Data extraction and management
We developed a data extraction based on the Standards for Reporting Implementation Studies Checklist (Pinnock et al., 2017) and the framework developed by Proctor et al. (Proctor et al., 2011).
Consequently, for each included study we extracted general characteristics, such as study type, country and participants, as well as methods and implementation outcomes, such as acceptability, feasibility, fidelity measures, effectiveness, cost-effectiveness and sustainability. A detailed taxonomy can be found in the Supplementary Materials.
The data extraction was conducted by one of the team member (DA) and verified by another author (CM).

| Synthesis of results
We conducted a narrative synthesis of the results, considering the diverse methodologies, settings, and outcomes measured. Besides, our aim was mapping EIP services in the region rather than obtaining a summary statistic.

| Literature search
We obtained 3282 unduplicated references from the searches. From these, 48 articles were reviewed in full text and 10 articles were included for analysis according to our inclusion and exclusion criteria.
A PRISMA flowchart is depicted in Figure 1

| Characteristics of included studies
The studies described seven EIP programmes (some articles referred to the same programme) and were conducted in Argentina, Brazil, Chile and Mexico. The study designs consisted of two randomizedcontrolled trials (RCTs) (Valencia, Juarez, Delgado, Díaz, et al., 2017;Valencia, Juarez, & Ortega, 2012), six observational studies (Cabral & Chaves, 2010;Chaves, 2007;Gaspar et al., 2019;Gonzalez-Valderrama et al., 2017;Louzã, Azevedo, Macedo, & Gattaz, 2008;Padilla et al., 2015) and two qualitative analyses (Eisenstadt, Monteiro, Diniz, & Chaves, 2012;Valenzuela, Pastorino, Alvarado, Villalón, & Vanegas, 2012). Sample sizes ranged from 16 to 102 participants with a mean (SD) of 33.7 (34.1) participants. Most of the studies were conducted in specialized settings, mainly research centres or tertiary care, with only two studies carried out in primary care (Padilla et al., 2015;Valenzuela et al., 2012). Additionally, we were able to find two articles describing specialized services for people at high-risk of psychosis (Gaspar et al., 2019;Louzã et al., 2008). These services included formal assessment of prodromal symptoms based on widely recognized instruments (P. Fusar-Poli et al., 2016; Paolo Fusar-Poli, 2017) and a mix of psychosocial interventions and medications to prevent or delay a full-blown psychosis episode, with one case also including a public campaign through newspapers, radio and TV to increase awareness of early psychosis (Louzã et al., 2008).
A list with the main characteristics of the included studies can be found in Table 1.

| Implementation outcomes
In terms of implementation outcomes, the studies were heterogeneous. However, none of the included studies was comprehensive enough to include all of the items suggested by current guidelines (Pinnock et al., 2017; Proctor et al., 2011). All of the programmes F I G U R E 1 PRISMA flowchart were adopted at local level (hospital or health service) and interacted with other elements of the system, which is compatible with evidence of the feasibility and initial penetration of EIP services, according to the Proctor et al taxonomy (Proctor et al., 2011). Nevertheless, only a minority of the programmes described the appropriateness of the intervention. Likewise, the effectiveness of the interventions was reported by 3 out of 10 studies, with favourable results to EIP services. Furthermore, only two programmes (Gaspar et al., 2019;Valencia et al., 2012) measured the fidelity of some of their components and although the interventions described are consistent with other EIP services (Correll et al., 2018), none of the studies reported fidelity measures at service level.
Additionally, we found no evidence about affordability, costs and cost-effectiveness of these programmes. Finally, in terms of sustainability, although some programmes have continued within their hospitals or research centres, none of them have been scaled-up at national level.
A list of the included studies with implementation outcomes is depicted in Table 2
Brazil has a unified health system (Sistema The second programme found was ASAS (Louzã et al., 2008) (Avaliacao e seguimiento de Adolescentes e Adultos Jovenesem Sao Paulo). This intervention consisted of a public campaign through newspapers, radio and television to increase awareness of early symptoms of psychosis, followed by phone screening and an outpatient service for people meeting CHR-P criteria. Reporting was limited to a cross-sectional study describing the intervention and the characteristics of 18 participants meeting CHR-P criteria. The intervention was considered feasible, but the reporting of other implementation outcomes was scarce.

| Chile
The Chilean health system implemented a Health Benefit Plan where people with schizophrenia are prioritized to receive financially protected access to care, including medication and psychosocial interventions from their first episode (Markkula, Alvarado, & Minoletti, 2011).
Although the plan is intended to be nationally implemented, including the public and private sectors, there is evidence of inadequate offer of interventions (Markkula et al., 2011), and people with psychosis do not receive specialized interventions as recommended in EIP guidelines (Addington et al., 2018).
We found three articles describing EIP initiatives in Chile (Gaspar et al., 2019;Gonzalez-Valderrama et al., 2017;Valenzuela et al., 2012). One study (Valenzuela et al., 2012)  Only one report of a stand-alone EIP service was found (Gonzalez-Valderrama et al., 2017). This consisted of a multidisciplinary team working at tertiary level, offering inpatient and outpatient care for FEP patients. Although the programme has been proved to be feasible and adopted by the local system (8 years with an established coordination with other levels of the mental health system), we found no published evidence of fidelity, effectiveness, cost-effectiveness or scale-up to regional or national level.
Furthermore, we found one study describing the only Chilean programme designed for people at prodromal stages of psychosis (Gaspar et al., 2019).

| Mexico
The Mexican health system is organized into sub-systems of insurances largely determined by employment status (OECD, 2016 We identified two studies (Valencia et al., 2017(Valencia et al., , 2012 describing one EIP programme from Mexico. These were the only randomized trials found in our review, evaluating the effectiveness of an integrated treatment for FEP people compared with pharmacotherapy alone at the National Institute of Psychiatry in Mexico City.
The programmes were considered acceptable by users and families and included a bespoke fidelity measure to ensure the quality of psychosocial interventions. In both RCTs, participants receiving EIP services had better outcomes in terms of fewer relapses, shorter hospitalizations and lower symptomatology. We did not find information on economic outcomes or implementation at national level.

| Other Latin American countries
One study described an intervention taking place in Argentina (Padilla et al., 2015). This was a cohort study assessing the impact of training primary healthcare workers to reduce the DUP. Although the intervention was successful in terms of the primary outcome, no other implementation outcomes were reported and we are unaware whether the programme has continued or has been scaled-up at a higher level.
Finally, we were also able to find references from other Latin American countries, although they did not meet our inclusion criteria to be described in more detail.

| DISCUSSION
In this review, we have mapped the Latin American efforts to implement services for people in the early stages of psychosis, including all approaches taken. We followed a systematic methodology as recommended for scoping reviews. We also used the Implementation Science framework, in order to thoroughly assess the state of EIP services in the region. We hope this information will be valuable for policy makers, clinicians and service users alike.
Despite the systematic effort to find published and unpublished works reporting EIP initiatives in the region, we were able to identify only 10 studies and just four countries were represented. This review has highlighted the absence of information about cost-effectiveness and affordability of EIP services in Latin America.
Evidence on the cost-effectiveness of EIP in high-income countries may have limited transferability despite the potential value for money of reducing the impact of psychosis. It seems, therefore, paramount to generate local evidence if this model of care is to be expanded in the region as has happened in Europe, Australia and Asia.
Finally, we were not able to find any specific details of the adaptation of EIP to the Latin American population as described in other cultures. For instance, in Asia it was rapidly acknowledged the word "psychosis" might have a negative connotation and EIP programmes modified the term to reduce stigma and improve population reach (Lee, 2013). It has also been highlighted that EIP services should tailor their interventions to meet various ethnic, cultural and religious backgrounds. We did not find any description of such adaptations in our review, which seems essential to secure a culturally sensible implementation.

| Limitations
This review is subject to several limitations. Although we expanded our search to local databases and journals and we did not restrict articles by language, it is possible that relevant non-indexed publications were missed. Hence, we may have overlooked local initiatives that are not reported but already in place. We used other sources and contacted local researchers, but we did not formally search the grey literature.
Similarly, the scarce information on implementation outcomes might reflect reporting problems rather than a lack of attention to them. We used a validated instrument to assess such outcomes and a highly cited framework in mental health. We also triangulated information from several reports when they referred to the same EIP programme. However, it might be the case that researchers were not aware of the importance of implementation outcomes, or simply that it was not the focus of their work. We believe, nonetheless, these outcomes are crucial to inform the adaptation of complex interventions such as EIP programmes.

| CONCLUSIONS
Despite the consistent evidence about effectiveness and costeffectiveness of EIP services from diverse health systems, implementation in Latin America has been slow. The studies in this review are evidence that this model of care is feasible and acceptable to implement in the region, although this has been limited to research centres based in capital cities and at a tertiary healthcare level. It is, therefore, paramount to generate local evidence in terms of cultural appropriateness, fidelity and cost-effectiveness of EIP to successfully scale-up these programmes nationwide.