Nurse-Led Applications of Artificial Intelligence in Schizophrenia Care: A Scoping Review

1. Research Background Schizophrenia (SCZ) is a severe chronic mental disorder characterized by cognitive, thinking, emotional, and behavioral abnormalities, affecting approximately 24 million people worldwide according to the World Health Organization (WHO). Patients with schizophrenia require long-term, personalized nursing interventions due to the risk of dangerous behaviors such as self-injury, suicide, and violence, as well as complex physical and psychological disorders. However, the global distribution of mental health resources is highly uneven, and the shortage of psychiatric nurses is a prominent issue—these factors collectively exacerbate the challenges of schizophrenia nursing. Traditional nursing models are plagued by repetitive labor, low efficiency, and poor flexibility, making it difficult to meet the growing and diverse nursing needs of patients. In recent years, the rapid advancement of artificial intelligence (AI) technology has brought transformative opportunities to mental health care. AI technologies, including machine learning, deep learning, natural language processing, computer vision, and intelligent robots, have demonstrated significant potential in disease data processing, risk prediction, and auxiliary intervention. They can break through the limitations of traditional nursing, such as heavy reliance on human resources and constraints of time and space. However, existing research on AI in schizophrenia care primarily focuses on the technology itself (e.g., diagnostic algorithms or predictive models) and rarely emphasizes the "nurse-led" core nursing context. Nurses, as frontline providers of schizophrenia care, play an irreplaceable role in bridging technology and patient-centered care, yet their specific roles and coordination mechanisms in AI-enabled care remain understudied and poorly defined. This research gap highlights the need for a systematic synthesis of existing evidence to clarify the panorama of nurse-led AI applications in schizophrenia care. 2. Research Purpose This study aims to conduct a comprehensive scoping review to systematically analyze the application status, main forms, and potential challenges of AI technology in schizophrenia care under the nurse-led model. The core research objectives are threefold: 1. To map the key application areas of nurse-led AI technology in schizophrenia care (e.g., assessment, diagnosis, prediction, intervention, monitoring); 2. To identify the main types of AI technologies involved, their specific applications, and reported effectiveness; 3. To clarify the roles of nurses in AI-enabled care and summarize the current limitations and future development directions of this field. By addressing these objectives, the study seeks to fill the perspective gap in the interdisciplinary field of AI technology and schizophrenia nursing, provide evidence-based references for clinical nursing practice, and lay a foundation for the optimization of intelligent nursing models. 3. Research Design and Methodology 3.1 Study Design This study adopts a scoping review methodology, guided by the Joanna Briggs Institute (JBI) Scoping Review Framework and reported in strict accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) checklist. Scoping reviews are selected due to the nascent and heterogeneous nature of research on nurse-led AI in schizophrenia care—this approach allows for comprehensive coverage of existing knowledge, clarification of conceptual definitions, and identification of research gaps, rather than focusing on in-depth analysis of a limited number of studies. 3.2 Information Sources and Search Strategy • Databases: A systematic search was conducted across 7 core academic databases, including EMBASE, PubMed, CINAHL, Cochrane Library, PsycINFO, Web of Science, and Cochran. • Search Period: Literature published from the establishment of each database to November 30, 2025, was included to ensure the timeliness and comprehensiveness of evidence. • Search Terms: Key terms included "schizophrenia," "artificial intelligence," "AI," "machine learning," "deep learning," "nurse-led," "psychiatric nursing," "mental health care," etc. The search strategy was initially developed and validated in EMBASE, then adjusted for compatibility with other databases. • Supplementary Search: Reference lists of included studies and relevant systematic reviews were manually screened to identify additional eligible literature. 3.3 Eligibility Criteria • Inclusion Criteria: (1) Studies focusing on patients with schizophrenia as the main population; (2) Studies applying AI-related technologies in nursing-related areas (e.g., care assessment, intervention implementation, risk monitoring); (3) Studies explicitly mentioning or implying a "nurse-led" or nurse-involved model; (4) Original studies (randomized controlled trials, observational studies, qualitative studies) and systematic/scoping reviews; (5) Studies published in English. • Exclusion Criteria: (1) Literature with unavailable full text or incomplete data; (2) Studies not related to nursing practice; (3) Technical studies focusing solely on AI algorithm development without clinical nursing application; (4) Duplicate publications. 3.4 Literature Screening and Data Extraction • Screening Process: All retrieved records were imported into EndNote 21 for deduplication. Two independent reviewers conducted two-stage screening: first based on titles and abstracts, then full-text screening. Disagreements were resolved through discussion or arbitration by a third reviewer. • Data Extraction: A standardized extraction table (see Appendix 1) was used to collect key information from included studies, including title, author, publication year, study design, sample size, AI application area, AI technology type, nurse roles, research results, and conclusions. The extraction table was pre-tested to ensure consistency, and data extraction was performed independently by two reviewers with cross-validation. 3.5 Data Synthesis Given the descriptive nature of scoping reviews, quantitative meta-analysis was not conducted. Instead, a thematic classification and narrative synthesis approach was adopted. Extracted data were organized into three core themes: (1) Application fields of AI technology; (2) Types and effectiveness of AI technologies; (3) Roles of nurses and existing challenges. Visualizations (e.g., flow diagrams, distribution charts) were used to present the synthesis results for clarity. 4. Expected Outcomes This study is expected to generate the following key outcomes: 1. Comprehensive mapping of application areas: Identify and categorize the main application scenarios of nurse-led AI in schizophrenia care (e.g., disease prediction, cognitive intervention, symptom assessment, medication adherence monitoring), and quantify the distribution of literature across these areas (preliminary results indicate prediction and intervention are the most studied fields). 2. Classification of AI technology types: Summarize the main AI technologies involved (e.g., machine learning, deep learning, robotics) and their specific applications (e.g., risk prediction models, remote cognitive training tools, intelligent monitoring platforms), along with their reported effectiveness (e.g., improvement in assessment accuracy, reduction in nursing workload). 3. Clarification of nurse roles: Systematically define the roles of nurses in AI-enabled care (e.g., information collectors, intervention coordinators, clinical evaluators) and quantify the prevalence of these roles in existing literature (preliminary findings show only 12 out of 42 included studies explicitly mention nurse roles). 4. Identification of challenges and limitations: Synthesize the main challenges in the field, including technical bottlenecks (e.g., data standardization issues, "black box" algorithms), ethical and legal concerns (e.g., patient privacy protection, algorithmic bias), and practical barriers (e.g., nurse training gaps, cost-effectiveness issues). 5. Evidence-based recommendations: Provide targeted suggestions for clinical practice (e.g., prioritizing AI tools for high-risk prediction and remote intervention), nursing education (e.g., integrating AI literacy training into psychiatric nursing curricula), and future research (e.g., strengthening large-sample, multi-center studies on nurse-AI collaboration). 5. Significance of the Research 5.1 Academic Significance • Originality: This study is the first to adopt a "nurse-led" framework to systematically synthesize AI applications in schizophrenia care, filling the gap in existing reviews that focus solely on technology rather than nursing practice. • Methodological rigor: By adhering to JBI guidelines and PRISMA-ScR standards, the study ensures transparency and reproducibility, providing a methodological reference for future scoping reviews in intelligent nursing. • Knowledge integration: The thematic synthesis of heterogeneous evidence clarifies the conceptual landscape of nurse-led AI care, laying the foundation for interdisciplinary collaboration between nursing and AI technology. 5.2 Practical Significance • Clinical practice: The results will help clinical nurses select appropriate AI tools, optimize care processes, and alleviate the burden of staffing shortages. For example, identifying mature machine learning-based risk prediction models can enable nurses to proactively intervene in suicidal or violent behaviors. • Policy and education: The findings will inform the development of training programs for psychiatric nurses (e.g., enhancing data literacy and AI technology familiarity) and provide evidence for health policy-makers to promote the standardized application of AI in mental health care. • Patient outcomes: By optimizing the nurse-led AI model, the study aims to improve the accessibility, personalization, and effectiveness of schizophreni