Hofmann Andreas B, Spiller Tobias R, Rapaport Jeremy A, Homan Philipp, Seifritz Erich, Vetter Stefan, Egger Stephan T, Burrer Achim
Department of Adult Psychiatry and Psychotherapy, University Hospital of Psychiatry Zurich (PUK), Lenggstrasse 31, 8032, Zurich, Switzerland.
Psychiatric Services St. Gallen, Wil, SG, Switzerland.
BMC Psychiatry. 2025 Jul 17;25(1):711. doi: 10.1186/s12888-025-07149-9.
The length of stay (LoS) in psychiatric facilities is a critical metric for healthcare planning and resource allocation. While previous research has established that LoS distributions are typically right-skewed across medical specialties, detailed characterizations of these distributions within psychiatric settings remain limited, particularly regarding variations across diagnostic categories.
We conducted a retrospective cross-sectional analysis of 17,687 psychiatric hospitalizations at the University Hospital of Psychiatry Zurich between 2013 and 2020. Using both linear and logarithmic visualizations, we examined LoS distribution patterns across diagnostic groups based on ICD-10 classifications.
Following identified distribution patterns, patients could be categorized into short-stay (1-10 days) and long-stay (11-200 days) groups for comparative analysis. LoS distribution demonstrated a bimodal pattern when visualized on a logarithmic scale, with distinct peaks representing short-term crisis interventions and longer therapeutic hospitalizations. This bimodal distribution was particularly evident in anxiety and stress-related disorders and major depressive disorder. Diagnostic categories differed significantly in their LoS-distribution patterns, with schizophrenia spectrum disorders, organic mental disorders, and bipolar disorders more frequently requiring extended hospitalizations. Long-stay patients exhibited higher admission HoNOS scores (median 20 vs. 18) and were significantly older (median 49 vs. 39 years) than short-stay patients.
Our findings reveal that psychiatric hospitalization durations follow a bimodal rather than simply right-skewed distribution, suggesting two distinct patient populations requiring fundamentally different treatment approaches. This pattern varies systematically across diagnostic categories but transcends diagnostic boundaries, indicating that factors beyond primary diagnosis influence treatment duration. These results support the development of differentiated care structures addressing both acute crisis intervention and extended therapeutic needs within psychiatric care systems.
精神科机构的住院时长是医疗规划和资源分配的关键指标。虽然先前的研究已证实，住院时长分布在各医学专科中通常呈右偏态，但在精神科环境中对这些分布的详细特征描述仍然有限，尤其是关于不同诊断类别的差异。
我们对苏黎世大学精神病医院2013年至2020年间的17687例精神科住院病例进行了回顾性横断面分析。我们使用线性和对数可视化方法，根据国际疾病分类第十版（ICD - 10）分类检查了不同诊断组的住院时长分布模式。
根据确定的分布模式，患者可分为短期住院（1 - 10天）和长期住院（11 - 200天）两组进行比较分析。当以对数尺度可视化时，住院时长分布呈现双峰模式，不同的峰值分别代表短期危机干预和较长时间的治疗性住院。这种双峰分布在焦虑和应激相关障碍以及重度抑郁症中尤为明显。不同诊断类别在住院时长分布模式上存在显著差异，精神分裂症谱系障碍、器质性精神障碍和双相情感障碍更常需要延长住院时间。长期住院患者的入院健康需求和问题量表（HoNOS）得分更高（中位数为20分对18分），且明显比短期住院患者年龄更大（中位数为49岁对39岁）。
我们的研究结果表明，精神科住院时长遵循双峰分布而非简单的右偏态分布，这表明存在两类截然不同的患者群体，需要根本不同的治疗方法。这种模式在不同诊断类别中系统变化，但跨越了诊断界限，表明除主要诊断外的因素会影响治疗时长。这些结果支持在精神科护理系统中发展差异化护理结构，以满足急性危机干预和长期治疗需求。
