针对华裔成人常见精神障碍的心理干预随机对照试验：试验质量与可信度的系统评价

Randomised Controlled Trials of Psychological Interventions Treating Common Mental Disorders in Adults of Chinese Descent: A Systematic Review of Trial Quality and Credibility

——《柳叶刀-精神病学》第13卷第7期，2026年7月——

< The Lancet / Psychiatry>, Volume 13, Issue 7, July 2026

【摘要】背景：自2000年以来，针对华人常见精神障碍的心理干预随机对照试验数量有所增加，这些研究可为心理干预的跨文化有效性提供重要证据。既往针对心理干预随机对照试验的系统评价（主要基于北美和欧洲的研究）一致指出其存在方法学缺陷；然而，针对华人人群的证据质量尚未得到全面评估。本系统评价旨在描述针对华人常见精神障碍心理干预随机对照试验的方法学特征及可信度。方法：本系统评价检索了MEDLINE、Embase、PsycINFO、CNKI和万方（WANFANG）数据库，纳入了自数据库建库起至2025年9月1日期间发表的符合条件的英文或中文随机对照试验。纳入标准包括：针对年龄在18岁及以上、被诊断患有常见精神障碍的华人成年人（汉族血统占比至少80%）进行的心理干预试验。12名评审员独立且双人进行筛选（依据纳入标准）及数据提取（涵盖研究特征、汇总结果数据及期刊特征）。利用结构化质量框架评估纳入研究的试验设计、实施、报告及结果。我们计算了干预后的组间标准化均数差，以描述所报告效应的分布情况。我们在研究与撰写过程中纳入了具有相关生活经验的人士。结果：通过数据库检索共识别出37,787项研究，最终纳入857项随机对照试验，涉及75,143名参与者。结果显示：34项（4%）研究报告了已注册的试验方案；398项（46%）采用了充分的随机化方法；21项（2%）采用了有效的分配隐藏；78项（9%）对结果评估者实施了盲法；32项（4%）进行了意向性分析（ITT分析）；78项（9%）在实施干预时参考了治疗手册。 417项（49%）研究在数据真实性方面存在问题，其效应量中位数（g=2.06）是440项（51%）未发现此类问题的研究（g=0.96）的两倍以上。关于充分随机化方法的报告情况有所改善：在2006年之前发表的18项研究中，无一报告了此类方法；而在检索时段最后5年发表的257项研究中，有179项（70%）报告了此类方法。纳入的研究分布在278种期刊上，其中188种（68%）期刊未提供关于试验报告的公开指导意见，仅41种（15%）引用了既定的报告指南。解读：大量随机对照试验在试验报告、方法学严谨性和数据真实性方面存在局限性，引发了人们对证据质量的担忧。需要采取系统性措施——包括方法学培训、监管基础设施建设以及期刊监管——以提升这些日益增多的证据的可信度及其贡献价值。

[Summary] Background: Since 2000, there has been an increase in the number of randomised controlled trials of psychological interventions for common mental disorders in people of Chinese descent, which could provide important evidence regarding the cross-cultural effectiveness of psychological interventions. Systematic reviews of randomised controlled trials of psychological interventions, predominantly carried out in North America and Europe, have consistently identified methodological shortcomings; the quality of the evidence for people of Chinese descent has not been comprehensively evaluated. This systematic review aims to characterise the methodology and credibility of randomised controlled trials of psychological interventions for common mental disorders in people of Chinese descent (ie, ethnic Chinese populations). Methods: In this systematic review, we searched MEDLINE, Embase, PsycINFO, CNKI, and WANFANG to identify eligible randomised controlled trials published in English or Chinese from database inception to Sept 1, 2025. We included trials of psychological interventions in adults of Chinese descent (with at least 80% of Han Chinese descent) aged 18 years and older with a diagnosis of a common mental disorder. 12 reviewers independently and in duplicate judged selection against inclusion criteria and extracted data on study characteristics, aggregated outcome data, and journal characteristics. Trial design, conduct, reporting, and outcomes of included studies were evaluated using a structured quality framework. We calculated post-intervention standardised mean differences between groups to characterise the distribution of reported effects. We involved people with related lived experience in the research and writing process. Findings: We identified 37 787 studies through database search, of which 857 randomised controlled trials, encompassing 75 143 participants, were included in the review. We found that 34 (4%) studies reported a registered trial protocol, 398 (46%) had adequate methods of randomisation, 21 (2%) used valid allocation concealment, 78 (9%) used outcome assessor blinding, 32 (4%) had an intention-to-treat analysis, and 78 (9%) referenced a treatment manual for the interventions delivered. 417 (49%) studies raised concerns over data authenticity, with a median effect size (g=2·06) more than double that of the 440 (51%) studies with no identified concerns (g=0·96). The reporting of adequate randomisation methods improved from none of 18 studies published before 2006 to 179 (70%) of 257 studies published in the last 5 years of the search range. Included studies were published across 278 journals, of which 188 (68%) journals provided no accessible guidance on trial reporting and only 41 (15%) referenced established reporting guidelines. Interpretation: The large volume of randomised controlled trials with limitations in trial reporting, methodological rigour, and data authenticity raised concerns over the quality of the evidence. Systemic efforts including methodological training, regulatory infrastructure, and journal oversight are needed to improve the credibility and contribution of this growing body of evidence.

论文原文：Siying Li, Zhenyang Xi, Yiran Wei, er al. (2026). Randomised controlled trials of psychological interventions treating common mental disorders in adults of Chinese descent: a systematic review of trial quality and credibility. The Lancet / Psychiatry, Volume 13, Issue 7, Pages 581-589. July 2026.

https://doi.org/10.1016/S2215-0366(26)00133-1

（翻译兼责任编辑：MARY）

（需要英文原文的朋友，请联系微信：millerdeng95或iacmsp）