Scholarlink

<p>This paper critically examines how Human–AI collaboration and metacognitive occupational therapy (MCT-OT) contribute to schizophrenia rehabilitation. This study integrates a quasi-experimental clinical trial with a qualitative systems evaluation of AI-enabled tools, including an EMR-integrated clinical decision support system (CDSS) for risk flagging, an NLP-based progress-monitoring module, and a clinician dashboard. A mixed-methods design was used. Quantitative phase: <i>n</i> = 80 male participants (experimental <i>n</i> = 40, control <i>n</i> = 40) received a 12-session MCT-OT versus standard care; primary outcomes were MCQ-30 and RDAS, analyzed using paired/independent t-tests and ANCOVA with effect sizes (Cohen’s d) reported. Qualitative phase: thematic analysis of interviews and workflow observations in three rehabilitation centers. The experimental group showed statistically significant improvements on lower MCQ-30 scores indicating reduced dysfunctional metacognitive beliefs; mean change = 8.64 (SD = 4.23); Cohen’s d = 1.06; <i>p</i> &lt; .001 and on RDAS (mean change = 8.22; <i>p</i> &lt; .001). Effective schizophrenia rehabilitation requires a hybrid Human–AI model that preserves clinician oversight and therapeutic flexibility while leveraging AI for monitoring and decision support. Limitations include male-only sampling and single-region recruitment; further studies should test longitudinal and more diverse samples.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Human–AI Collaboration and Metacognitive Occupational Therapy in Schizophrenia Rehabilitation

  • 期刊论文

  • 发表日期：

    2025年12月17日

查看全文
  • Narjes Houshangi

摘要

This paper critically examines how Human–AI collaboration and metacognitive occupational therapy (MCT-OT) contribute to schizophrenia rehabilitation. This study integrates a quasi-experimental clinical trial with a qualitative systems evaluation of AI-enabled tools, including an EMR-integrated clinical decision support system (CDSS) for risk flagging, an NLP-based progress-monitoring module, and a clinician dashboard. A mixed-methods design was used. Quantitative phase: n = 80 male participants (experimental n = 40, control n = 40) received a 12-session MCT-OT versus standard care; primary outcomes were MCQ-30 and RDAS, analyzed using paired/independent t-tests and ANCOVA with effect sizes (Cohen’s d) reported. Qualitative phase: thematic analysis of interviews and workflow observations in three rehabilitation centers. The experimental group showed statistically significant improvements on lower MCQ-30 scores indicating reduced dysfunctional metacognitive beliefs; mean change = 8.64 (SD = 4.23); Cohen’s d = 1.06; p < .001 and on RDAS (mean change = 8.22; p < .001). Effective schizophrenia rehabilitation requires a hybrid Human–AI model that preserves clinician oversight and therapeutic flexibility while leveraging AI for monitoring and decision support. Limitations include male-only sampling and single-region recruitment; further studies should test longitudinal and more diverse samples.

Graphical Abstract