Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Front Psychiatry . 2026 Mar 6;17:1711659. doi: 10.3389/fpsyt.2026.1711659 Search in PMC Search in PubMed View in NLM Catalog Add to search PERMEPSY: a multicentre, randomized, double-blind proof-of-concept trial of personalized metacognitive training for adults with psychosis — a study protocol Maria Lamarca Maria Lamarca 1 Research Unit, Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Barcelona, Spain 2 Grup Etiopathogenesis and Treatment of Severe Mental Disorders (MERITT), Fundació Sant Joan de Déu, Institut de Recerca Sant Joan de Déu, Esplugues de Llobregat, Barcelona, Spain 3 Consorcio de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Salud Carlos III, Madrid, Spain 4 Departament de Psicologia Clínica i de la Salut, Facultat de Psicologia, Universitat Autònoma de Barcelona, Bellaterra, Cerdanyola del Vallès, Barcelona, Spain Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing Find articles by Maria Lamarca 1, 2, 3, 4, *, † , Claudia Requejo Claudia Requejo 1 Research Unit, Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Barcelona, Spain 2 Grup Etiopathogenesis and Treatment of Severe Mental Disorders (MERITT), Fundació Sant Joan de Déu, Institut de Recerca Sant Joan de Déu, Esplugues de Llobregat, Barcelona, Spain 5 Facultad de Psicología, Universitat de Barcelona, Barcelona, Spain Investigation, Writing – original draft, Writing – review & editing Find articles by Claudia Requejo 1, 2, 5, † , Adrianna Aleksandrowicz Adrianna Aleksandrowicz 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland Investigation, Methodology, Writing – review & editing Find articles by Adrianna Aleksandrowicz 6, † , Adrien Goncalves Adrien Goncalves 7 University of Strasbourg, Inserm, Strasbourg, France Investigation, Writing – review & editing Find articles by Adrien Goncalves 7 , Martyna Kreżołek Martyna Kreżołek 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland Data curation, Writing – review & editing Find articles by Martyna Kreżołek 6 , Hanna Gelner Hanna Gelner 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland Investigation, Methodology, Writing – review & editing Find articles by Hanna Gelner 6 , Justyna Piwińska Justyna Piwińska 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland Investigation, Methodology, Writing – review & editing Find articles by Justyna Piwińska 6 , Rabea Fischer Rabea Fischer 8 Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Methodology, Writing – review & editing Find articles by Rabea Fischer 8 , Merle Schlechte Merle Schlechte 8 Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Data curation, Investigation, Writing – review & editing Find articles by Merle Schlechte 8 , Alvaro Cavieres Alvaro Cavieres 9 Departamento de Psiquiatría, Escuela de Medicina, Facultad de Medicina, Universidad de Valparaíso, Valparaíso, Chile 10 Unidad de Trastornos Psicóticos. Hospital Del Salvador de Valparaíso, Valparaíso, Chile Writing – review & editing Find articles by Alvaro Cavieres 9, 10 , Vanessa Acuña Vanessa Acuña 9 Departamento de Psiquiatría, Escuela de Medicina, Facultad de Medicina, Universidad de Valparaíso, Valparaíso, Chile 10 Unidad de Trastornos Psicóticos. Hospital Del Salvador de Valparaíso, Valparaíso, Chile Methodology, Supervision, Writing – review & editing Find articles by Vanessa Acuña 9, 10 , Fabrice Berna Fabrice Berna 5 Facultad de Psicología, Universitat de Barcelona, Barcelona, Spain 7 University of Strasbourg, Inserm, Strasbourg, France Methodology, Supervision, Writing – review & editing Find articles by Fabrice Berna 5, 7 , Steffen Moritz Steffen Moritz 8 Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Methodology, Supervision, Writing – review & editing Find articles by Steffen Moritz 8 , Caroline König Caroline König 11 Soft Computing Research Group at Intelligent Data Science and Artificial Intelligence Research Center Universitat Politècnica de Catalunya, Barcelona, Spain Formal analysis, Methodology, Resources, Software, Supervision, Validation, Writing – review & editing Find articles by Caroline König 11 , Łukasz Gawęda Łukasz Gawęda 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland Methodology, Supervision, Writing – review & editing Find articles by Łukasz Gawęda 6, ‡ , Susana Ochoa Susana Ochoa 1 Research Unit, Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Barcelona, Spain 2 Grup Etiopathogenesis and Treatment of Severe Mental Disorders (MERITT), Fundació Sant Joan de Déu, Institut de Recerca Sant Joan de Déu, Esplugues de Llobregat, Barcelona, Spain 3 Consorcio de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Salud Carlos III, Madrid, Spain Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing Find articles by Susana Ochoa 1, 2, 3, *, ‡ Author information Article notes Copyright and License information 1 Research Unit, Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Barcelona, Spain 2 Grup Etiopathogenesis and Treatment of Severe Mental Disorders (MERITT), Fundació Sant Joan de Déu, Institut de Recerca Sant Joan de Déu, Esplugues de Llobregat, Barcelona, Spain 3 Consorcio de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Salud Carlos III, Madrid, Spain 4 Departament de Psicologia Clínica i de la Salut, Facultat de Psicologia, Universitat Autònoma de Barcelona, Bellaterra, Cerdanyola del Vallès, Barcelona, Spain 5 Facultad de Psicología, Universitat de Barcelona, Barcelona, Spain 6 II Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland 7 University of Strasbourg, Inserm, Strasbourg, France 8 Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany 9 Departamento de Psiquiatría, Escuela de Medicina, Facultad de Medicina, Universidad de Valparaíso, Valparaíso, Chile 10 Unidad de Trastornos Psicóticos. Hospital Del Salvador de Valparaíso, Valparaíso, Chile 11 Soft Computing Research Group at Intelligent Data Science and Artificial Intelligence Research Center Universitat Politècnica de Catalunya, Barcelona, Spain * Correspondence: Maria Lamarca, [email protected] ; Susana Ochoa, [email protected] †These authors share first authorship ‡These authors share senior authorship Roles Maria Lamarca : Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing Claudia Requejo : Investigation, Writing – original draft, Writing – review & editing Adrianna Aleksandrowicz : Investigation, Methodology, Writing – review & editing Adrien Goncalves : Investigation, Writing – review & editing Martyna Kreżołek : Data curation, Writing – review & editing Hanna Gelner : Investigation, Methodology, Writing – review & editing Justyna Piwińska : Investigation, Methodology, Writing – review & editing Rabea Fischer : Methodology, Writing – review & editing Merle Schlechte : Data curation, Investigation, Writing – review & editing Alvaro Cavieres : Writing – review & editing Vanessa Acuña : Methodology, Supervision, Writing – review & editing Fabrice Berna : Methodology, Supervision, Writing – review & editing Steffen Moritz : Methodology, Supervision, Writing – review & editing Caroline König : Formal analysis, Methodology, Resources, Software, Supervision, Validation, Writing – review & editing Łukasz Gawęda : Methodology, Supervision, Writing – review & editing Susana Ochoa : Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing Received 2025 Sep 23; Accepted 2026 Jan 8; Revised 2025 Dec 26; Collection date 2026. Copyright © 2026 Lamarca, Requejo, Aleksandrowicz, Goncalves, Kreżołek, Gelner, Piwińska, Fischer, Schlechte, Cavieres, Acuña, Berna, Moritz, König, Gawęda and Ochoa. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. PMC Copyright notice PMCID: PMC13003757 PMID: 41868841 Abstract Background While psychological interventions are effective at improving symptoms of psychosis, accessible, cost- and time-efficient treatments remain limited. Personalized medicine has emerged as a promising approach, tailoring interventions to individual needs. Metacognitive Training (MCT), with its established efficacy and adaptable format, is well-suited for personalization. The PERMEPSY project (Towards a Personalized Medicine Approach to Psychological Treatment for Psychosis) aims to deliver tailored MCT intervention for individuals with psychosis. Methods PERMEPSY is an international study funded by ERAPerMed (JTC2022) involving five clinical partners (Spain, Chile, France, Germany, Poland) and one technological partner (Spain). The project involves a proof-of-concept clinical trial recruiting 51 participants from each center for a total of 255 adult participants with psychosis in a prospective study (Registration: NCT06603922 , 19-09-2024). The trial will test the efficacy of a Machine Learning (ML)-derived platform at predicting clinical and functional outcomes from baseline scores and compare a personalized MCT (P-MCT) to a classical MCT based on the platform’s predictions. Aims PERMEPSY seeks to (1) develop and test the predictive power of an algorithm that could support decision-making, and (2) ascertain whether P-MCT is more effective than MCT at improving key symptoms and cognitive impairments associated to psychosis. Results A harmonized retrospective database enabled the development of a predictive ML algorithm, integrated into an innovative platform. This platform provides clinicians with the information needed to deliver P-MCT. Predictions include changes in positive symptoms (e.g., delusions), insight, self-esteem, and treatment adherence. Discussion By integrating diverse data types and innovative technology, PERMEPSY addresses the need for personalized, effective treatment in psychosis, aiming to reduce individual and systemic burdens while supporting clinicians in their decision-making. Keywords: clinical trial, metacognitive training, PERMEPSY, personalized medicine, psychological interventions, psychosis, schizophrenia 1. Introduction Schizophrenia and other psychotic disorders are severe mental health disorders affecting approximately 2% of the global population ( 1 ), with a 12-month prevalence of 4.03 per 1000 persons ( 2 ) and an incidence of all psychotic disorders of 26.6 per 100.000 person-years ( 3 ), leading to significant levels of disability and generating a considerable burden for patients, caregivers, and healthcare systems. In Europe, the average annual direct cost per patient is estimated at 5.800€ ( 4 ). Although pharmacological treatment with antipsychotics is widely used, it does not always improve functional outcomes, with discontinuation rates ranging from 50% to 75% among patients ( 5 ). In this context, psychological interventions have demonstrated to be beneficial, even in the absence of pharmacological treatment, with 40% of patients experiencing significant symptom improvement as a result of psychological interventions ( 6 ). Metacognitive Training (MCT) is a low-threshold, manualized intervention available free of charge in 40 languages (available www.uke.de/mct ). Based on the principles of Cognitive Behavioral Therapy (CBT), MCT is an effective psychological intervention for psychosis that has been incorporated into the latest World Health Organization (WHO) treatment guidelines ( 7 ), and is recommended by the German Psychiatric Association (DGPPN) ( 8 ), the German Psychological Society (DGP) ( 9 ), and the Australian and New Zealand College of Psychiatrists ( 10 ).Recent quantitative syntheses have provided a detailed picture of the specific clinical targets most responsive to MCT in schizophrenia. A comprehensive meta-review ( 11 ) reported medium to large effects on delusions ( g = 0.64), positive symptoms ( g = 0.47), and total psychotic symptoms ( g = 0.39), with smaller but significant effects on hallucinations ( g = 0.27) and negative symptoms ( g = 0.23). Additional meta-analyses have shown improvements in general functioning ( g = 0.41), self-esteem ( g = 0.17), and cognitive biases ( g = 0.16) ( 12 ), as well as global social cognition ( d = 0.28) and theory of mind ( d = 0.27) ( 13 ). MCT also improves insight ( g = 0.35), although no significant effects have been found for global neurocognition ( 14 ), aligning with the intervention’s theoretical focus on metacognitive processes rather than fundamental cognitive capacities. These findings are consistent with a recent graded evidence review ( 15 ), which assigned a WFSBP-1 level of evidence to MCT for reducing positive symptoms, delusions, and total psychotic symptoms; and WFSBP-3 for hallucinations, negative symptoms, general functioning, theory of mind, jumping to conclusions, and insight. Together, this literature provides a detailed and clinically relevant framework of expected treatment targets. However, despite its effectiveness, there are currently no predictive models capable of assessing whether an individual is likely to respond well to MCT or which MCT modules would be most suitable for improving a patient’s prognosis. Multiple studies ( 16 – 18 ) have identified cognitive biases, symptom severity, self-esteem, age of onset, and sex as key moderators of treatment efficacy. Similarly, gender differences in response to MCT have been established, with another study finding that women with psychosis respond differently and more favorably to MCT than men ( 19 , 20 ). Biomarkers are considered essential for identifying patient subgroups and enabling personalized approaches in precision medicine in psychiatry ( 21 , 22 ). Considering biological variables, potential inflammatory biomarkers associated with MCT response include C-reactive protein, extensively studied in schizophrenia due to its correlation with symptom severity, and leukocyte counts, which are linked to grey matter volume ( 23 , 24 ). Previous studies support the role of SP1 and SP4 transcription factors as biomarkers that may be modulated by MCT treatment ( 25 – 27 ). Therefore, considering the previous literature biomarkers may become an important moderator to consider in the treatment of psychosis, and enable subgroup analyses that could help determine biological profiles of treatment response. The extensive research and quality of studies researching the effect of MCT provides an ideal context for developing a data-driven approach using an analytical pipeline based on machine learning (ML) based predictive models. Predictive models are used both for diagnosis and prognosis in the personalization of psychiatric healthcare ( 28 , 29 ), and ML serves as a natural enabler of personalized medicine ( 30 – 32 ) when combined with knowledge extraction and explainability strategies ( 33 ). Most researchers and clinicians agree that the treatment of psychotic patients should be personalized and tailored to their specific needs ( 34 ). However, its implementation in clinical practice remains limited and depends on overcoming several challenges One reason is the economic cost of personalization and its adoption by clinicians ( 35 ) among others. Another significant barrier is the limited scientific knowledge regarding the factors that should guide treatment customization, an ongoing challenge in mental health research ( 36 ). The PERMEPSY Project emerges as a response to this need and available data, with the aim of developing a predictive prototype platform to assist clinicians in determining the optimal psychological MCT treatment strategy for each individual. Predictions are based on each individual’s personal characteristics and baseline scores in various clinical, cognitive and functional outcomes. These outcomes were selected both by data-driven insights derived from the ML analysis and an extensive review of the literature by consortium partners ( 11 , 15 , 19 , 37 , 38 ), which took place prior to the development of the prototype platform ( 39 ) to identify key indicators of treatment efficacy in psychosis. This innovative approach leverages ML techniques to analyze a broad spectrum of variables, from environmental factors to biomarkers, thus reflecting the multifactorial nature of mental illnesses. The technological solutions employed in PERMEPSY, such as ML, can facilitate the identification of key variables in predicting treatment response. This may help clinicians focus on both objective and subjective problematic areas, consider the course of illness, and adopt a gender-sensitive approach early on in the treatment stage. The project also integrates methodological advancements, ensuring the personalized treatment for each patient is data-driven and supported by the prototype platform. The overarching goal is to develop an open platform that supports clinicians in treatment decision-making, enabling them to predict MCT response, recommend personalized MCT (P-MCT) treatment, and provide evidence-based information. This will assist clinicians in making informed decisions regarding psychological interventions for psychosis. A clinical trial will be conducted to validate both the platform and the predictive model for personalized MCT. Therefore, the primary objective of this prospective clinical trial is (1) to evaluate the efficacy of P-MCT compared to classical MCT in individuals diagnosed with psychosis, focusing on improvements in positive symptoms (especially delusions), insight, self-esteem, and treatment adherence. Moreover, some specific objectives of this trial include: (2) to validate the ML-based predictive model’s capacity to anticipate MCT treatment response, (3) to evaluate patient satisfaction with P-MCT comparing to classical MCT, integrating both quantitative measures and qualitative insights from focus groups, and (4) to refine the platform’s predictive algorithm through integration of treatment outcome data. Given that P-MCT has not been previously evaluated and that the ML platform is a novel tool, hypotheses 2–4 are considered exploratory. Our hypotheses are as follows: H1 - Primary (Comparative efficacy): P-MCT will yield comparable or superior improvements than classical MCT in outcome variables measuring treatment efficacy (positive symptoms, insight, self-esteem, and treatment adherence). H2 - Exploratory (Platform validation): The ML-based predictive model will demonstrate significant predictive capacity to anticipate post-treatment response to MCT on positive symptoms (especially delusions), insight, self-esteem, and treatment adherence, as assessed by the AUC (Area Under the Curve). H3 - Exploratory (Patient satisfaction): Patient satisfaction with P-MCT will be comparable or greater than with classical MCT, as assessed through both quantitative measures and qualitative insights from focus groups. H4 - Exploratory (Platform improvement): The integration of multimodal comprehensive pre-post treatment data from the classical MCT group will refine the machine learning algorithm, resulting in improved predictive accuracy for future patient cohorts. 2. Methods and analysis 2.1. Study design This is a randomized, double-blind, parallel-group, exploratory proof-of-concept trial designed to assess the superiority of P-MCT over classical MCT. Participants are recruited from five countries (Poland, Germany, France, Chile, and Spain) and are randomly assigned to one of the two treatment conditions. Participants will include in-patient and out-patient volunteers based on center availability. Following baseline assessment, participants are randomized to either the classical Metacognitive Training (MCT) or the Personalized MCT (P-MCT) condition using a computer-generated four-block randomization procedure generated by the technological team (UPC), ensuring allocation concealment. The experimental intervention will be P-MCT. Participants receiving classical MCT will serve as an active control group following the current program. The study will be double-blind, as patients and evaluators will not be aware of the assigned treatment condition. The trial has been registered in clinicaltrials.gov (ID: NCT06603922 ; https://www.clinicaltrials.gov/study/NCT06603922 ; 19/09/24), where any deviations from this protocol will also be recorded. Both interventions will consist of 10 one-hour sessions, delivered in group format once or twice a week, depending on site feasibility. Some groups will be delivered MCT or P-MCT in online format to allow representative recruitment. Patient data will be fed into an ML platform prior to the intervention. The platform generates a personalized treatment response prediction to allow tailoring of the intervention by selecting individualized homework tasks targeting specific cognitive deficits. Allocation concealment during treatment will be ensured via sealed opaque envelopes prepared by an independent researcher not involved in recruitment or assessment. The platform does not store data, and thus predictions will be manually recorded for both treatment conditions. Only the P-MCT condition will receive personalization following the platform’s predictions, while the classical MCT predictions will be stored to analyze the platform’s accuracy at predicting the results of classical MCT. Post-treatment assessments will be conducted after the intervention and subsequently at a six-month follow-up evaluation. Full details of the study design are outlined in Supplementary Figure S2 , and platform input and output information, and an example of the homework can be found in Supplementary Figure S3 . 2.2. Participants The sample will consist of patients with a diagnosis of non-affective psychosis, according to DSM-5 criteria ( 40 ). Participants will be recruited from multiple clinical centers (associated to the following clinical partners: Fundació Sant Joan de Déu [FSJD], Universitätsklinikum Hamburg-Eppendorf [UKE], Hôpitaux Universitaires de Strasbourg [UHS], Centre Hospitalier Universitaire de Nice (CHU de Nice), Institute of Psychology, Polish Academy of Science [PAoS] and Universidad del Valparaíso [UV]), including both inpatient and outpatient services, with a minimum of 51 patients per center. Recruitment will occur in psychiatric hospitals and community mental health centers. To facilitate recruitment, clinicians at these sites will receive detailed information about the study’s inclusion and exclusion criteria through local meetings and internal communications. Additionally, announcements will be posted in the centers to raise awareness and encourage referrals during routine clinical care. Inclusion criteria are (1): A DSM-5 diagnosis of non-affective psychosis (2); presence of positive symptoms during the past year (represented by a Positive and Negative Syndrome Scale [PANSS ( 41 )] score equal or over 3 on the items measuring Delusions, Suspiciousness, or Grandiosity); (3) aged between 18 and 65 years old; (4) stable clinical condition with no expected changes in medication (as confirmed by clinical services); (5) no severe cognitive deficits (global assessment and/or information from clinical services). Exclusion criteria are: (1) previous attendance in MCT groups within the past year; (2) neurological disorders, history of head trauma, or premorbid IQ below 70 (based on medical reports and/or other sources); (3) a score equal or over 6 on the Hostility and Suspiciousness items of the PANSS Positive subscale; (4) aggressive behavior (reported by clinical services); (5) high suicide risk (verified through the DIAMOND clinical interview). Patients who do not meet the inclusion criteria will not be enrolled in the study but will be offered alternative treatment options or MCT at later stages. The inclusion and exclusion criteria were designed to balance internal validity with real-world clinical representativeness. Criteria were primarily informed by those most employed across the studies that contributed to the development of the ML predictive platform, ensuring consistency between the training dataset and prospective trial population. The requirement for positive symptoms (PANSS ≥ 3) ensures the presence of clinically relevant MCT targets, while the upper threshold for hostility/suspiciousness (PANSS ≤ 6) maintains therapeutic dynamics and avoid group issues, while maintaining group representativeness and excluding severe cases and at the same time without overly restricting sample representativeness. The stable medication criterion reduces confounding from pharmacological changes, while cognitive and safety exclusions (IQ < 70, high suicide risk, aggressive behavior) ensure participants can meaningfully engage with the intervention. This approach avoids overly stringent criteria that would limit generalizability while maintaining sufficient homogeneity to detect treatment effects ( 42 ). 2.3. Measures Data will be collected on sociodemographic, clinical, cognitive (neurocognitive, metacognitive, social-cognitive), and psychological variables. Given the transnational nature of the study, official validated translations of all instruments were used for each participating country. When validated translations existed but item ordering differed between countries, administration was standardized according to the official English version to ensure consistency across sites. For instruments lacking official translations in some specific languages (e.g., The Internal, Personal, and Situational Attributions Questionnaire (IPSAQ), Fish Task) and for consortium-developed scales (e.g., modified Patient-Reported Impact of Symptoms in Schizophrenia (PRISS) scale to monitor symptoms during treatment, satisfaction with MCT treatment), forward-backward translation procedures were conducted by independent bilingual researchers to ensure semantic and conceptual equivalence. The instruments used and their characteristics are presented in Supplementary Table S1 . Blood samples will be drawn from participants at baseline and post-treatment. In both centers collecting blood samples (FSJD and UV), leukocyte subpopulation counts obtained from blood count analyses and C-reactive protein levels will be collected. At the FSJD site, peripheral expression levels of SP transcription factors (SP1 and SP4) will be analyzed as part of a subproject aimed at identifying biomarkers with transcriptional reprogramming capacity involved in the response to MCT. Biomarkers are included because they provide objective biological indicators of inflammatory and transcriptional processes that may influence cognitive functioning and symptom severity in psychosis. These processes are increasingly recognized as moderators of treatment response in psychological interventions. Including biomarkers enables subgroup analyses to determine whether MCT efficacy varies by biological profile. Incorporating biomarker data will refine the ML algorithm. This iterative process will determine which biomarkers consistently improve predictive accuracy, guiding future versions of the platform toward more biologically informed personalization. Following the intervention, qualitative data on patient satisfaction will be collected through focus groups at each center, exploring participants’ experiences, perceived benefits, barriers, and overall satisfaction with the intervention. This data will complement the quantitative measures collected post-treatment and at follow-up. 2.4. Data collection Pre-treatment, post-treatment, and follow-up assessments will be conducted by blinded, trained psychologists at each site. These evaluations will take place over the course of one to two interviews. Assessors will be trained in standardized administration of instruments to ensure equal collection across centers. Nursing staff will be responsible for performing cubital vein blood counts for separation of peripheral blood mononuclear cells. Blood will be stored at 4°C until processed. Additionally, the transfer of samples to the Parc Sanitari Sant Joan de Déu (PSSJD) biobank will be requested following a prior agreement for participation in the study. The analyses will be conducted at the PSSJD Molecular Psychiatry Laboratory (SP1 and SP4 analysis) and the Hospital Sant Joan de Déu Clinical Laboratory (C-reactive protein and complete blood count). Blood will be collected from participants in the sites at UV and FSJD. This study is in accordance with the World Medical Association’s Declaration of Helsinki and was approved by the Sant Joan de Déu Ethics Committee (Reference: PIC-110-24), the Institute of Psychology, Polish Academy of Science (Reference: IP.403.21.2024), Ethics Committee of Valparaíso San Antonio Health Service of Chile (Reference: N°48/2024), the local psychological ethics committee at the center for psychosocial medicine (LPEK) at the University Medical Center Hamburg-Eppendorf (Reference: LPEK-0796) and the Comité de protection des personnes Ile de France (Reference: 25.02907.000469). Since the study involves human participants, evaluations will be conducted once informed consent has been obtained. The personnel conducting the assessments, blood extractions, blood counts, and molecular analyses will be blind to the study condition assigned to each participant. All variables will be measured at baseline, post-treatment, and six months after completion, except for the DIAMOND and TEC scales, which will only be administered at baseline; the PRISS scale, which will be administered at the end of each session to monitor symptoms during treatment; patient satisfaction with treatment, which will be collected post-treatment and at follow-up; and the biological samples, which will be collected only within two weeks before the start of the intervention and within two weeks after its completion. Data will be entered into a secure, password-protected database. Coding will follow standardized formats. Range checks and validation rules will be applied. Data will be stored at each site in compliance with GDPR and local regulations. For qualitative data, each center will carry out two focus groups (one for the personalized MCT group and one for the classical MCT group), with 5–10 participants per group. Sessions will be audio-recorded, transcribed verbatim, anonymized, and translated into English. Field notes will also be kept. Monthly meetings with representatives from all centers will be held to triangulate results and ensure consistency across sites. 2.5. Interventions Participants who meet inclusion criteria will be randomized into one of the two study conditions: (1) Metacognitive Training (MCT) and (2) Personalized Metacognitive Training (P-MCT). Sessions will run concurrently in all centers from October 2024 until April 2025. MCT consists of 10 one-hour group sessions (3–10 participants). It includes ten modules referring to common cognitive issues and biases in solving problems in psychosis. MCT is effective in reducing symptoms and improving cognitive biases that are involved in the onset and maintenance of psychosis ( 12 ). The topics of MCT are the following: attribution blaming and taking credit (module 1), jumping to conclusions (modules 2 & 7), changing beliefs (module 3), deficits in theory of mind and other social cognition domains (modules 4 and 6), overconfidence in errors (module 5), mood (module 8), self-esteem (module 9) and self-stigma (module 10). Each session follows a protocol defined in the manual ‘Metacognitive Training for Psychosis’ (MCT). The format (online or in-person) and frequency (once or twice a week) will vary by center depending on staff resources; this information will be considered as covariables in the analysis. Participants receive worksheets at the end of each module to continue working on the modules at home. The personalization strategy was developed through a two-day in-person consortium meeting dedicated to integrating ML predictions into treatment. Consistent with the dual objectives of personalizing treatment while maintaining accessibility, the consortium decided against full individualization of sessions, which would have reduced the number of patients able to access group-based MCT. Instead, P-MCT differs from classical MCT exclusively in homework content, while group sessions remained identical in both groups. Participants are asked to discuss their homework tasks with their MCT therapist exclusively to preserve blindness to treatment condition. After the follow-up assessment, participants are informed of which treatment group they were allocated to and the homework tasks given to the other group, as well as any extra personalized tasks not received are offered to them. A key strategic decision concerned how to personalize homework. Rather than reinforcing domains predicted to improve most, the consortium adopted a recovery-oriented approach by targeting the three outcome domains with the lowest predicted improvement out of the five domains discussed to best represent successful treatment response (positive symptoms, insight, self-esteem, and treatment adherence). personalized exercises were developed by clinicians within the consortium using (1) clinical practice guidelines and (2) predictive features identified using SHAP values from the ML model ( 39 ), ensuring alignment between clinical relevance and data-driven indicators. Sixteen exercises (four per outcome domain) were developed and standardized in length and complexity to ensure that any differences in treatment effects arise from personalization rather than workload. Homework assignment follows a fixed and reproducible rule: for each participant, the platform identifies the three outcome domains with the lowest predicted improvement. Across the ten MCT sessions, homework is assigned by cycling repeatedly through these three domains (Domain 1 → Domain 2 → Domain 3 → repeat). Because the program includes ten sessions, Domain 1 receives four exercises, while Domains 2 and 3 each receive three exercises. Within each domain, exercises are delivered sequentially (Exercise 1 to Exercise 4), ensuring a consistent rotation structure tailored to the participant’s predicted needs while maintaining standardization of the overall intervention. The established order of tasks represents the degree of relevance of the related factor detected by the ML model. For instance, in the Positive Symptom analysis, related factors associated with improvement in positive symptoms included positive symptoms themselves, negative symptoms, JTC and self-reflection, in this order. As a result, in order to target positive symptoms, the four homework tasks related to this domain include (1) psychoeducation on positive symptoms, (2) an activity planner to reduce negative symptoms, (3) a task to identify and correct JTC biases, combining self-reflection with JTC improvement, and (4) positive symptoms psychoeducation on stress management. Group sessions will be delivered by trained clinical psychologists or psychiatrists ( https://clinical-neuropsychology.de/metacognitive-training-e-training/ ) with prior experience in MCT. All facilitators will undergo standardized training in both P-MCT and classical MCT, and monthly meetings will include supervision to ensure protocol fidelity. Data from all participants’ baseline assessment will be input into the platform, and its results will be used to personalize the homework for participants in the P-MCT condition only. Results from the MCT condition will be stored and only used to confirm the platform’s prediction accuracy. In terms of adherence monitoring and ensuring protocol fidelity, participants are asked to complete a questionnaire after every session attended to record their symptomatology over the previous seven days, the completion of homework and their attendance to the session (session number and module attended). Other than the offered MCT sessions, participants will continue to receive their routine care through their usual center, including medication and any other non-MCT psychological treatment. Clinicians and MCT trainers will monitor for adverse events and discontinue the allocated intervention in case of worsening symptoms. Participants are allowed to drop out of the study at any point, and they are informed that they have the right to withdraw from their therapy group at any time when they sign the informed consent forms. Reasons for withdrawal, adherence to sessions and homework and reasons for missing data will be recorded. 2.6. Sample size calculation To achieve a two-tailed size effect (Cohen’s d ) of 0.5 with 95% power, using the total PANSS scores and balanced observations in the two groups, the estimated size is 210 in total. Considering 20% of loss to follow-up, we aim to include a total of 252 patients in the study. 2.7. Data analysis 2.7.1. Comparative efficacy (H1) To assess the interventions’ efficacy, improvements in the proximal targets of MCT, as defined in recent meta-analyses and meta-reviews ( 11 , 12 ), will be analyzed using statistical methods ( 43 ). These targets include global positive symptoms, including delusions, cognitive insight, self-esteem and treatment adherence. Regression models for repeated measures will be used for each proximal target, with the treatment group (P-MCT or classical MCT) included as a covariate for comparison. Additionally, mean differences between baseline and post-treatment scores in each group will be compared using Student’s t-tests, or Mann-Whitney tests if the assumption of normality is violated. Missing data will be handled using multiple imputation techniques. Subgroup analyses will be conducted based on sex, age, and baseline symptom severity. Sensitivity analyses will assess robustness of findings. ANCOVAs will explore whether responses differ across participant profiles ( 44 ), with treatment as the independent variable and composite participant profiles as a fixed factor. Covariates will include number of sessions attended, online vs. face-to-face delivery, and session frequency. Participant profiles will be identified using unsupervised ML techniques. 2.7.2. Platform validation (H2) In the classical MCT group, the predictive accuracy of the platform will be evaluated by comparing actual results with predictions in the five key variables (positive symptoms, delusions, insight, self-esteem, and treatment completion), with predictive performance assessed using the AUC for classification models ( 45 ) and RMSE and the R 2 determination coefficient for regression models ( 46 ). In the P-MCT group, the analysis will focus on whether this approach led to improved outcomes compared to baseline predictions. Effect sizes will be calculated using Cohen’s d. 2.7.3. Patient satisfaction (H3) Patient satisfaction will be assessed using a mixed-methods approach, combining quantitative measures collected post-treatment and at follow-up with qualitative data from focus groups conducted at each center. Focus groups will be transcribed verbatim, coded and analyzed using content analysis using ATLAS.ti, and grouped into categories and subcategories reflecting recurrent themes (satisfaction, perceived benefits, and barriers). Methodological rigor will follow Guba and Lincoln’s criteria of dependability, credibility, transferability, and confirmability ( 47 ). Triangulation across researchers and sites will be performed in monthly consortium meetings to enhance reliability and credibility. Representative quotes will be reported while maintaining anonymity. 2.7.4. Platform improvement (H4) Exploratory analyses using unsupervised ML techniques will examine whether integration of comprehensive pre-post distal treatment data, including quantitative clinical, cognitive, psychological measures and blood biomarker data (e.g., depression, trauma, cognitive bias, SP1/SP4 expression), can refine the ML platform and improve predictive accuracy for future patient cohorts. Qualitative data will not be used for platform improvement. This combined approach allows for integration of quantitative outcomes with qualitative insights, providing a comprehensive assessment of intervention efficacy and patient experiences across the five participating countries. 3. Discussion The present study constitutes a pioneering effort in the development of a platform aimed at optimizing psychosis treatment strategies by integrating a personalized approach centered on the individual characteristics of each patient. While existing literature supports the need to tailor interventions to the specific attributes of each individual, current clinical practice seldom implements this approach due, among other factors, to economic constraints and the limited scientific knowledge regarding the criteria that should guide treatment personalization. MCT emerges as an ideal candidate for evaluating the feasibility of a personalized psychological treatment approach in psychosis for several reasons. Firstly, its efficacy has been extensively demonstrated both in the short and long term ( 12 ). Secondly, it is recommended in therapeutic guidelines as a first-line intervention for individuals with psychosis, addressing the psychological underpinnings of the disorder. Furthermore, MCT has shown positive outcomes across diverse psychosis subpopulations, suggesting its applicability to various diagnoses and disease stages. Moreover, its manualized format ensures interinstitutional comparability and facilitates implementation by mental health professionals with appropriate training. Additionally, its group-based application enhances cost-effectiveness compared to individual therapies. The PERMEPSY project demonstrates a high degree of feasibility, supported by the consortium’s expertise in researching psychological interventions for psychosis, particularly MCT, as well as in the development of digital platforms and the application of machine learning. Prior to designing the randomized trial and the development of the predictive platform, the consortium conducted comprehensive systematic reviews of the literature to identify the most relevant variables associated with treatment response ( 11 , 15 , 19 , 37 , 38 ). These reviews, carried out by all partners according to their specific areas of expertise, ensured that the variables selected for inclusion in the retrospective harmonized database and the trial were both evidence-based and clinically meaningful. Variables identified by previous meta-analyses from researchers outside the consortium were also considered ( 12 , 13 , 48 ). This preliminary work strengthens the scientific foundation of the project and guarantees that the data collected are relevant for both the machine learning analysis and clinical decision-making. Moreover, the consortium has already collected and harmonized the necessary data for ML-based analyses and the creation of the platform prototype ( 39 , 49 , 50 ). The clinical team’s capacity to recruit patients in both past and ongoing studies further strengthens the feasibility of the trial. Despite robust evidence confirming the efficacy of psychological interventions in symptom reduction and relapse prevention ( 51 ), only 10% to 25% of individuals with psychosis have access to such treatments ( 52 ). This is particularly concerning given that 50% to 75% of patients discontinue antipsychotic medication ( 5 ). The PERMEPSY initiative seeks to bridge this gap by providing a personalized psychological intervention tailored to patients’ specific needs. It is estimated that profiling patients and delivering individualized treatment could reduce relapse rates by over 70%, thereby helping patients regain a sense of purpose, enhancing their functional capacities, and increasing their overall recovery prospects. This approach is also expected to yield a significant reduction in both direct and indirect costs associated with psychosis. Machine learning based technological solutions have the potential to enhance the identification of predictive variables for treatment response, thereby facilitating clinical decision-making. In the PERMEPSY study, a harmonized retrospective database was used to train a machine learning algorithm capable of predicting changes in key clinical and functional outcomes, such as positive symptoms, delusions, insight, self-esteem, between others. These predictions are generated based on composite patient profiles derived from pre-treatment data ( 39 , 50 ). The platform thus supports clinicians to recognize objective problem areas, such as cognitive biases to address in the therapy. By combining the recognition of object problem areas from the ML algorithm with information about subjective concerns reported by patients, the MCT can be tailored to the individual needs. Beyond these clinical and cognitive predictors, this study also exploratorily examines whether blood biomarkers (inflammatory markers and SP transcription factors at selected sites) can contribute to refining predictive models of MCT response, offering preliminary insights into potential biological correlates of therapeutic change. Furthermore, incorporating a gender-sensitive approach in the early stages of treatment could further optimize outcomes. The development of the prototype platform and its implementation in clinical practice therefore represents a significant methodological advancement, enabling patient stratification based on data-driven insights and technology-assisted personalization ( 39 ). While PERMEPSY constitutes an innovative initiative in applying personalized medicine principles to psychological interventions in psychosis, its implementation presents certain methodological challenges. The multicentric nature of the trial may introduce variability in MCT delivery, a factor that will be mitigated through a standardized training program, monthly supervision sessions, and staff mobility for trainers. Additionally, although the total sample size ( n = 252) is sufficient to detect overall group differences, statistical power may be inadequate for conducting detailed subgroup analyses (e.g., by gender or comorbidities). To address this limitation, ML models prioritizing the most predictive variables will be employed, thereby reducing the risk of overfitting. Finally, this study will also contribute to the development of knowledge regarding the ethical and legal aspects of patient stratification, as well as the regulatory frameworks for ML-based clinical decision support systems across various European Union countries and beyond. In fact, the consortium is already working on a regulatory roadmap in order to meet requirements for medical device certification, allowing the platform to be used in clinical settings. The findings may redefine mental health care standards by promoting more precise, cost-effective, and patient-centered interventions, in contrast to conventional therapeutic models. In line with this, the project provides an opportunity to adhere to current regulatory frameworks with the aim of facilitating the future translation of the developed platform into a certified medical device. PERMEPSY represents a groundbreaking initiative in translating personalized medicine principles into the domain of psychological interventions for psychosis. By integrating advanced ML techniques within a transnational clinical trial, the project seeks not only to validate a predictive model but also to establish methodological foundations for future research in this emerging field. The potential benefits, ranging from resource optimization to reducing chronicity, justify the logistical and technical challenges inherent to its execution. The results of this trial could redefine mental health care standards, prioritizing precise, patient-centered interventions over outdated therapeutic paradigms. 4. Ethics and dissemination Patient data included in the PERMEPSY database will be collected and stored in compliance with the Ethical standards established by the World Medical Association in the Declaration of Helsinki at its current version (75 th ) ( 53 , 54 ). The study was approved by the Sant Joan de Déu Ethics Committee (Reference: PIC-110-24), the Institute of Psychology, Polish Academy of Science (Reference: IP.403.21.2024), Ethics Committee of Valparaíso San Antonio Health Service of Chile (Reference: 48/2024), the local psychological ethics committee at the center for psychosocial medicine (LPEK) at the University Medical Center Hamburg-Eppendorf (Reference: LPEK-0796) and the Comité de protection des personnes Ile de France (Reference: 25.02907.000469). Patient safety will be monitored throughout the study. Case report forms (CRFs) at each site have been designed to systematically record adverse events voluntarily reported by participants at each assessment point. Additionally, a modified version of the Patient-Reported Impact of Symptoms in Schizophrenia (PRISS) scale will be administered after each MCT session to monitor symptom changes during treatment and identify potential adverse effects. Any serious adverse events will be immediately reported to the designated responsible physician at each center, who will implement appropriate medical measures according to local clinical protocols and regulatory requirements. Serious adverse events will also be reported to local ethics committees in accordance with institutional and national regulations. Personal data will be codified and stored securely and confidentially, not leaving each center’s archives. Upon completion of data collection from all centers, patient identification of patients recruited outside one’s own center will be unavailable through pseudonymization. The final prospective database will be anonymized and will not contain any personal data from participants. The anonymized prospective database will be openly available 10 years after the end of the project from the Cora repository ( https://cora.csuc.cat/en/rdr-research-data-repository/ ). Access to the anonymized final dataset will be granted to consortium partners. The dataset will be made publicly available 10 years after project completion. The length of stored data will be 10 years after the end of the project, and destruction of stored data in open access is 10 more years after the end of the embargo period (20 years after the end of the project). All participants included in the study will be asked to give informed consent for their data to be used in future related projects, for their data to be anonymized and shared in open access after 10 years, and for their pseudonymized data to be shared with consortium partners during the project. An additional informed consent form will be collected to gather blood samples and store them in the PSSJD biobank. Patients will also be informed that their data will be destroyed at the 20-year mark. The treatment, communication and transfer of data of all participants will comply with Regulation EU 2016/679 of the European Parliament and of the Council of April 27, 2016, on the protection of natural persons with regard to the processing of personal data and the free movement of such data and with the Organic Law 3/2018, of December 5th, 2018, on the Protection of Personal Data and guarantee of digital rights. The justifying legal basis of data processing is the patient’s signed consent, in accordance with the provisions of article 9 of EU Regulation 2016/679. The project’s outcomes will be communicated in expert scientific conferences, peer-reviewed publications in scientific journals (in open access where applicable) and in community events or as plain-language summaries to reach participants and other lived experienced individuals, healthcare professionals, the public, and other relevant stakeholders. Trial outcomes will be registered and updated in public trial registries. This study protocol and all future publications will adhere to the Consolidated Standards of Reporting Trials (CONSORT) guidelines for transparent reporting of randomized controlled trials. The Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist ( 55 ) can be found in the Supplementary Table S4 . The project’s past, present and future dissemination actions can be viewed from the project’s website: permepsy.org . Funding Statement The author(s) declared that financial support was received for this work and/or its publication. This work is part of the European ERAPERMED 2022-292 call for international collaboration projects of the 2021–2023 Strategic Action in Health funded by European Union - Next Generation EU resources under the Recovery, Transformation and Resilience Plan, supported as grant AC22/0010 by the Instituto de Salud Carlos III (ISCIII) in Spain. The PERMEPSY project was supported under the frame of ERA PerMed by: Instituto de Salud Carlos III (ISCIII), Spain, Grant numbers AC22/00010 & AC22/0053; German Federal Ministry of Education and Research (BMBF), Germany, Grant number 01KU2306; Agence Nationale de la Recherche (ANR), France, Grant number ANR-22-PERM-0009-05.; National Centre for Research and Development (NCBR) Grant number: PerMed/V/82/PERMEPSY/2023, Poland; Agencia Nacional de Investigación y Desarrollo (ANID), Chile, Grant number REC+ERAPERMED2022-292.Moreover, CR is the beneficiary of of the FI-STEP predoctoral programme scholarship of the Secretariat for Universities and Research of the Department of Research and Universities of the Government of Catalonia, under grant reference 2025 STEP 00458, co-financed by the European Union through the European Social Fund Plus (ESF+). Footnotes Edited by: Esben Strodl , Queensland University of Technology, Australia Reviewed by: Michele Poletti , IRCCS Local Health Authority of Reggio Emilia, Italy Hamish McLeod , University of Glasgow, United Kingdom Ethics statement This study is in accordance with the World Medical Association’s Declaration of Helsinki and was approved by the Sant Joan de Déu Ethics Committee (Reference: PIC-110-24), the Institute of Psychology, Polish Academy of Science (Reference: IP.403.21.2024), Ethics Committee of Valparaíso San Antonio Health Service of Chile (Reference: N°48/2024), the local psychological ethics committee at the centre for psychosocial medicine (LPEK) at the University Medical Center Hamburg-Eppendorf (Reference: LPEK-0796) and the Comité de protection des personnes Ile de France (Reference: 25.02907.000469). The participants provided their written informed consent to participate in this study. Author contributions ML: Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing. CR: Investigation, Writing – original draft, Writing – review & editing. AA: Investigation, Methodology, Writing – review & editing. AG: Investigation, Writing – review & editing. MK: Data curation, Writing – review & editing. HG: Investigation, Methodology, Writing – review & editing. JP: Investigation, Methodology, Writing – review & editing. RF: Methodology, Writing – review & editing. MS: Data curation, Investigation, Writing – review & editing. AC: Writing – review & editing. VA: Methodology, Supervision, Writing – review & editing. FB: Methodology, Supervision, Writing – review & editing. SM: Methodology, Supervision, Writing – review & editing. CK: Formal Analysis, Methodology, Resources, Software, Supervision, Validation, Writing – review & editing. ŁG: Methodology, Supervision, Writing – review & editing. SO: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing. Conflict of interest The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision. Generative AI statement The author(s) declared that generative AI was not used in the creation of this manuscript. Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyt.2026.1711659/full#supplementary-material DataSheet1.pdf (602.2KB, pdf) References 1. Ochoa S, Haro J, Torres J, Pintomeza A, Palacin C, Bernal M, et al. What is the relative importance of self reported psychotic symptoms in epidemiological studies? Results from the ESEMeD—Catalonia Study. Schizophr Res. (2008) 102:261–9. doi: 10.1016/j.schres.2008.04.010, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 2. Moreno-Küstner B, Martín C, Pastor L, McKenna PJ. Prevalence of psychotic disorders and its association with methodological issues. A systematic review and meta-analyses. PloS One. (2018) 13:e0195687. doi: 10.1371/journal.pone.0195687, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Jongsma HE, Turner C, Kirkbride JB, Jones PB. International incidence of psychotic disorders, 2002–17: a systematic review and meta-analysis. Lancet Public Health. (2019) 4:e229–44. doi: 10.1016/S2468-2667(19)30056-8, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Kovács G, Almási T, Millier A, Toumi M, Horváth M, Kóczián K, et al. Direct healthcare cost of schizophrenia – European overview. Eur Psychiatr. (2018) 48:79–92. doi: 10.1016/j.eurpsy.2017.10.008, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 5. Moritz S, Veckenstedt R, Randjbar S, Vitzthum F, Woodward TS. Antipsychotic treatment beyond antipsychotics: metacognitive intervention for schizophrenia patients improves delusional symptoms. Psychol Med. (2011) 41:1823–32. doi: 10.1017/S0033291710002618, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 6. Morrison AP, Turkington D, Pyle M, Spencer H, Brabban A, Dunn G, et al. Cognitive therapy for people with schizophrenia spectrum disorders not taking antipsychotic drugs: a single-blind randomized controlled trial. Lancet. (2014) 383:1395–403. doi: 10.1016/S0140-6736(13)62246-1, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 7. World Health Organization . Package of interventions for rehabilitation: module 8: mental health: web annex: literature reviews and evidence tables. In: Package of interventions for rehabilitation: module 8: mental health: web annex: literature reviews and evidence tables. Geneva, Switzerland: (2023). [ Google Scholar ] 8. Gaebel W, Hasan A, Falkai P. S3-leitlinie schizophrenie. Heidelberg Germany: Springer-Verlag; (2019). [ Google Scholar ] 9. Lincoln T, Pedersen A, Hahlweg K, Wiedl KH, Frantz I. Evidenzbasierte Leitlinie zur Psychotherapie von Schizophrenie und anderen psychotischen Störungen Vol. 5. Göttingen, Germany: Hogrefe Verlag GmbH & Company KG; (2019). [ Google Scholar ] 10. Galletly C, Castle D, Dark F, Humberstone V, Jablensky A, Killackey E, et al. Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for the management of schizophrenia and related disorders. Aust New Z J Psychiatry. (2016) 50:410–72. doi: 10.1177/0004867416641195, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 11. Meinhart A, Sauvé G, Schmueser A, Penney D, Berna F, Gawęda Ł, et al. Metacognitive training for psychosis (MCT): a systematic meta-review of its effectiveness. Transl Psychiatry. (2025) 15:156. doi: 10.1038/s41398-025-03344-0, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Penney D, Sauvé G, Mendelson D, Thibaudeau >É, Moritz S, Lepage M. Immediate and sustained outcomes and moderators associated with metacognitive training for psychosis: A systematic review and meta-analysis. JAMA Psychiatry. (2022) 79:417. doi: 10.1001/jamapsychiatry.2022.0277, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Hotte-Meunier A, Penney D, Mendelson D, Thibaudeau É, Moritz S, Lepage M, et al. Effects of metacognitive training (MCT) on social cognition for schizophrenia spectrum and related psychotic disorders: a systematic review and meta-analysis. Psychol Med. (2024) 54:914–20. doi: 10.1017/S0033291723002611, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 14. Jeffrey C, Penney D, Sauvé G, Mendelson D, Thibaudeau É, Moritz S, et al. Does metacognitive training for psychosis (MCT) improve neurocognitive performance? A systematic review and meta-analysis. Schizophr Res. (2025) 275:79–86. doi: 10.1016/j.schres.2024.12.004, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 15. Goncalves A, Ochoa S, Moritz S, Gawęda Ł, Cavieres A, König C, et al. Metacognitive training (MCT) for psychosis: a systematic review and grade recommendations. Eur Psychiatr. (2025) 68:e80. doi: 10.1192/j.eurpsy.2025.10027, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. González-Blanch C, Birulés I, Pousa E, Barrigon ML, López-Carrilero R, Lorente-Rovira E, et al. Moderators of cognitive insight outcome in metacognitive training for first-episode psychosis. J Psychiatr Res. (2021) 141:104–10. doi: 10.1016/j.jpsychires.2021.06.040, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 17. Leanza L, Studerus E, Bozikas VP, Moritz S, Andreou C. Moderators of treatment efficacy in individualized metacognitive training for psychosis (MCT+). J Behav Ther Exp Psychiatry. (2020) 68:101547. doi: 10.1016/j.jbtep.2020.101547, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 18. Moritz S, Menon M, Andersen D, Woodward TS, Gallinat J. Moderators of symptomatic outcome in metacognitive training for psychosis (MCT). Who Benefits Who Does Not? Cognit Ther Res. (2018) 42:80–91. doi: 10.1007/s10608-017-9868-3, PMID: 41784122 [ DOI ] [ Google Scholar ] 19. Gelner H, Krężołek M, Aleksandrowicz A, Piwińska J, Lamarca M, Berna F, et al. The role of demographic variables in the effectiveness of metacognitive training for psychosis: A meta-analysis. Schizophr Res. (2025) :283:85–97. doi: 10.1016/j.schres.2025.06.017, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 20. Salas-Sender M, López-Carrilero R, Barajas A, Lorente-Rovira E, Pousa E, Barrigón ML, et al. Gender differences in response to metacognitive training in people with first-episode psychosis. J Consulting Clin Psychol. (2020) 88:516–25. doi: 10.1037/ccp0000468, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 21. Insel TR. The NIMH research domain criteria (RDoC) project: precision medicine for psychiatry. AJP. (2014) 171:395–7. doi: 10.1176/appi.ajp.2014.14020138, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 22. Niculescu AB, Le-Niculescu H. Precision medicine in psychiatry: biomarkers to the forefront. Neuropsychopharmacol. (2022) 47:422–3. doi: 10.1038/s41386-021-01183-3, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Fernandes BS, Steiner J, Bernstein HG, Dodd S, Pasco JA, Dean OM, et al. C-reactive protein is increased in schizophrenia but is not altered by antipsychotics: meta-analysis and implications. Mol Psychiatry. (2016) 21:554–64. doi: 10.1038/mp.2015.87, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 24. Núñez C, Stephan-Otto C, Usall J, Bioque M, Lobo A, González-Pinto A, et al. Neutrophil count is associated with reduced gray matter and enlarged ventricles in first-episode psychosis. Schizophr Bulletin. (2019) 45:846–58. doi: 10.1093/schbul/sby113, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Fusté M, Meléndez-Pérez I, Villalta-Gil V, Pinacho R, Villalmanzo N, Cardoner N, et al. Specificity proteins 1 and 4, hippocampal volume and first-episode psychosis. Br J Psychiatry. (2016) 208:591–2. doi: 10.1192/bjp.bp.114.152140, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 26. Vila È, Huerta-Ramos E, Núñez C, Usall J, Ramos B. Specificity proteins 1 and 4 in peripheral blood mononuclear cells in postmenopausal women with schizophrenia: a 24-week double-blind, randomized, parallel, placebo-controlled trial. Eur Arch Psychiatry Clin Neurosci. (2019) 269:941–8. doi: 10.1007/s00406-018-0938-7, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 27. Zhou X. Over-representation of potential SP4 target genes within schizophrenia-risk genes. Mol Psychiatry. (2022) 27:849–54. doi: 10.1038/s41380-021-01376-8, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Kaushik P, Bansal K, Kumar Y, Changela A. Mental disorders prognosis and predictions using artificial intelligence techniques: a comprehensive study. SN Comput Sci. (2024) 5:1048. doi: 10.1007/s42979-024-03416-w, PMID: 41784122 [ DOI ] [ Google Scholar ] 29. Meehan AJ, Lewis SJ, Fazel S, Fusar-Poli P, Steyerberg EW, Stahl D, et al. Clinical prediction models in psychiatry: a systematic review of two decades of progress and challenges. Mol Psychiatry. (2022) 27:2700–8. doi: 10.1038/s41380-022-01528-4, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Koutsouleris N, Kahn RS, Chekroud AM, Leucht S, Falkai P, Wobrock T, et al. Multisite prediction of 4-week and 52-week treatment outcomes in patients with first-episode psychosis: a machine learning approach. Lancet Psychiatry. (2016) 3:935–46. doi: 10.1016/S2215-0366(16)30171-7, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 31. Nebot À, Domènech S, Albino-Pires N, Mugica F, Benali A, Porta X, et al. LONG-REMI: an AI-based technological application to promote healthy mental longevity grounded in reminiscence therapy. IJERPH. (2022) 19:5997. doi: 10.3390/ijerph19105997, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Vellido A, Biganzoli E, Lisboa P. (2008). Machine learning in cancer research: implications for personalised medicine, in: 16th European Symposium on Artificial Neural Networks, Bruges (Belgium): ESANN. 55–64. [ Google Scholar ] 33. Vellido A. The importance of interpretability and visualization in machine learning for applications in medicine and health care. Neural Comput Applic. (2020) 32:18069–83. doi: 10.1007/s00521-019-04051-w, PMID: 41784122 [ DOI ] [ Google Scholar ] 34. The right treatment for each patient: unlocking the potential of personalized psychiatry. Nat Ment Health. (2023) 1:607–8. doi: 10.1038/s44220-023-00131-y, PMID: 41775740 [ DOI ] [ Google Scholar ] 35. International Consortium for Personalised Medicine (ICPerMed) . Value and evaluation of personalised medicine (2025). Available online at: https://www.icpermed.eu (Accessed December 24, 2025). 36. Comai S, Manchia M, Bosia M, Miola A, Poletti S, Benedetti F, et al. Moving toward precision and personalized treatment strategies in psychiatry. Int J Neuropsychopharmacol. (2025) 28:pyaf025. doi: 10.1093/ijnp/pyaf025, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Acuña V, Cavieres Á, Arancibia M, Escobar C, Moritz S, Gaweda L, et al. Assessing patient satisfaction with metacognitive training (MCT) for psychosis: A systematic review of randomized clinical trials. Clin Psychol Psychoth. (2024) 31:e3065. doi: 10.1002/cpp.3065, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 38. Lamarca M, Espinosa V, Acuña V, Vila-Badia R, Balsells-Mejia S, Moritz S, et al. Reducing self-stigma in psychosis: A systematic review and meta-analysis of psychological interventions. Psychiatry Res. (2024) 342:116262. doi: 10.1016/j.psychres.2024.116262, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 39. König C, Copado P, Vellido A, Nebot À, Angulo C, Lamarca M, et al. An artificial intelligence-based platform for personalized predictions of Metacognitive Training effectiveness. Comput Struct Biotechnol J. (2025) 28:281–93. doi: 10.1016/j.csbj.2025.07.051, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. American psychiatric association . Diagnostic and statistical manual of mental disorders. Fifth Edition. Washington DC USA: American Psychiatric Association; (2013). doi: 10.1176/appi.books.9780890425596, PMID: [ DOI ] [ Google Scholar ] 41. Kay SR, Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull. (1987) 13:261–76. doi: 10.1093/schbul/13.2.261, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 42. Tan YY, Papez V, Chang WH, Mueller SH, Denaxas S, Lai AG. Comparing clinical trial population representativeness to real-world populations: an external validity analysis encompassing 43–895 trials and 5 685–738 individuals across 989 unique drugs and 286 conditions in England. Lancet Healthy Longevity. (2022) 3:e674–89. doi: 10.1016/S2666-7568(22)00186-6, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 43. Mansmann U, Boulesteix AL. Modelling individual response to treatment and its uncertainty:A review of statistical methods and challenges for future research. In: LaCaze A, Osimani B, editors. Uncertainty in pharmacology. Springer International Publishing, Cham: (2020). p. 319–44. (Boston Studies in the Philosophy and History of Science; vol. 338). doi: 10.1007/978-3-030-29179-2_14, PMID: [ DOI ] [ Google Scholar ] 44. Vickers AJ, Altman DG. Analysing controlled trials with baseline and follow up measurements. BMJ. (2001) 323:1123–4. doi: 10.1136/bmj.323.7321.1123, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 45. White N, Parsons R, Collins G, Barnett A. Evidence of questionable research practices in clinical prediction models. BMC Med. (2023) 21:339. doi: 10.1186/s12916-023-03048-6, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 46. Cohen J, Cohen P, West SG, Aiken LS. Applied multiple regression/correlation analysis for the behavioral sciences. 3rd edn. Routledge: (2013). doi: 10.4324/9780203774441, PMID: [ DOI ] [ Google Scholar ] 47. Lincoln YS, Guba EG. Naturalistic inquiry. Dublin, Ireland: sage; (1985). [ Google Scholar ] 48. Lopez-Morinigo JD, Ajnakina O, Martínez ASE, Escobedo-Aedo PJ, Ruiz-Ruano VG, Sánchez-Alonso S, et al. Can metacognitive interventions improve insight in schizophrenia spectrum disorders? A systematic review and meta-analysis. Psychol Med. (2020) 50:2289–301. doi: 10.1017/S0033291720003384, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. König C, Guendouz W, Copado P, Angulo C, Nebot À, Ochoa S, et al. Group discovery in a clinical database of patients with psychosis who have undergone metacognitive training. In: Cerulo L, Napolitano F, Bardozzo F, Cheng L, Occhipinti A, Pagnotta SM, editors. Computational intelligence methods for bioinformatics and biostatistics. Springer Nature Switzerland, Cham: (2025). p. 269–80. (Lecture Notes in Computer Science; vol. 15276). doi: 10.1007/978-3-031-89704-7_21, PMID: [ DOI ] [ Google Scholar ] 50. König C, Copado P, Lamarca M, Guendouz W, Fischer R, Schlechte M, et al. Data harmonization for the analysis of personalized treatment of psychosis with metacognitive training. Sci Rep. (2025) 15:10159. doi: 10.1038/s41598-025-94815-3, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Bighelli I, Rodolico A, García-Mieres H, Pitschel-Walz G, Hansen WP, Schneider-Thoma J, et al. Psychosocial and psychological interventions for relapse prevention in schizophrenia: a systematic review and network meta-analysis. Lancet Psychiatry. (2021) 8:969–80. doi: 10.1016/S2215-0366(21)00243-1, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 52. Haddock G, Eisner E, Boone C, Davies G, Coogan C, Barrowclough C. An investigation of the implementation of NICE-recommended CBT interventions for people with schizophrenia. J Ment Health. (2014) 23:162–5. doi: 10.3109/09638237.2013.869571, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 53. World Medical Association . World medical association declaration of helsinki: ethical principles for medical research involving human participants. JAMA. (2025) 333:71. doi: 10.1001/jama.2024.21972, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 54. World Medical Association Inc . World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. (2013) 310:2191–4. doi: 10.1001/jama.2013.281053, PMID: [ DOI ] [ PubMed ] [ Google Scholar ] 55. Chan AW, Boutron I, Hopewell S, Moher D, Schulz KF, Collins GS, et al. SPIRIT 2025 statement: updated guideline for protocols of randomized trials. BMJ. (2025) 389:e081477. doi: 10.1001/jama.2025.4486, PMID: [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials DataSheet1.pdf (602.2KB, pdf) Articles from Frontiers in Psychiatry are provided here courtesy of Frontiers Media SA ACTIONS View on publisher site PDF (569.8 KB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top