Bridging the Gap Between Efficacy and Practice: A Systematic Review of Shared Decision-Making in Severe Mental Illness.

Shared decision-making (SDM) is recommended for preference-sensitive treatment decisions in severe mental illness, but its routine use in psychiatric services remains inconsistent. This systematic review synthesized evidence published between 2014 and 2024 and examined clinical, cultural, and organizational factors associated with SDM implementation. We searched PubMed, Embase, CINAHL, Cochrane Library, Web of Science, CNKI, Wanfang, and VIP for studies published from January 2014 to November 2024. Forty-two studies involving 8798 participants were included. Reporting followed PRISMA 2020 and SWiM guidance. Study quality was assessed with design-specific Joanna Briggs Institute tools, and certainty of key findings was appraised using a combined GRADE-CERQual approach. Because of heterogeneity in study design, interventions, outcomes, and settings, findings were synthesized narratively using Popay et al's framework. Patient decision aids were generally associated with lower decisional conflict, greater readiness, satisfaction, involvement, or autonomy, with little evidence of meaningful consultation-time extension. Clinician training and SDM interventions improved short-term engagement and information sharing, but evidence for sustained improvements in adherence, decisional conflict, and clinician behavior was mixed or low certainty. Family-mediated or triadic decision-making was prominent in East and Southeast Asian studies and may support adherence and participation, although it can also create tension between patient autonomy and family preferences. Across studies, clinicians and patients often emphasized different treatment priorities, particularly symptom control versus side-effect tolerability. Evidence was concentrated in Europe and East Asia, with limited data from low-resource regions. SDM in psychiatric care appears most reliable for improving immediate decision processes rather than long-term outcomes. Future implementation should test culturally adapted triadic decision aids, workflow-integrated prompts, and interprofessional decision coaching while using cautious, context-sensitive outcome evaluation.