Many articles in psychiatric journals start their introduction by focusing on the gravity of mental disorders. Thus, for example, one article may start like ‘Depression is the leading cause of disability around the world. The number of people living with depression increased by around 18% between 2005 and 2015, and now affects 322 million people or about 4% of the world’s entire population’, just to quote one recent case of mine (1). Similar tonality can be easily found in other academic articles and popular books, touting such sensational words as ‘epidemic’ (2, 3), ‘plague’ (4), or ‘pandemic’ (5). What good evidence is there for such claims, however verisimilar and alarming they may sound? It has been repeatedly pointed out that prescriptions of antidepressants have doubled in the highincome countries between 2000 and 2015 (6) and currently more than 10% of adult populations are prescribed antidepressants in such countries as USA or Australia. The proportion of people with diagnosable mental disorders actually receiving treatment has also increased from approximately 20–33% in the 1990s in USA (7) or from 37% to 46% in the 2000s in Australia (8). These facts would certainly be contributing to the impression of an epidemic or a plague. However, we must remember that, on the one hand, antihypertensive and antidiabetic drug consumption has similarly doubled and that cholesterol-lowering drug consumption has quadrupled in the same 15 years (6) and that, on the other hand, there appears to be serous overdiagnosis and overprescribing in the case of depression (9). Have the prevalences of depression and anxiety increased? – this is the question, and the question is not easy to answer because, logically speaking, it requires large-scale surveys, repeated over the years, with similar enough methodologies. It is then no wonder that the literature has reported mixed and conflicting findings. We must remember that, in some cases, the degree of variability of estimated prevalences across countries is beyond reasonable explanation even when the studies had apparently used similar methodologies: For example, a report from the World Mental Health Survey Initiative summarized that the annual prevalence of depression is only 3.0% in Germany or Italy but it reaches 5.9% in France and 8.3% in USA (10). Should the figure in Germany rise to 5.9% in the next survey, could we argue that the prevalence has doubled in the meantime? Richter and his colleagues in this issue (11) set out to answer this challenging yet extremely important question by systematically reviewing epidemiological surveys that used the same methodology in the same geographical region repeatedly. They found 42 such studies, which compared prevalences of various conditions including depression, distress, alcohol dependence, and general mental illness, between two time points of, on average, 9.9 years apart. The pooled OR was 1.18 (95% confidence interval: 1.07–1.31). The authors concluded that the prevalence of adult mental illness has only modestly increased. The heterogeneity among the studies was extremely high, with I-squared of 96%. The subgroup analyses indicated strong effect modification (relative OR close to 0.5 or 2.0) because of the survey methodology (clinical diagnosis vs self-report symptoms), target mental condition (general mental illness vs substance dependence), decades (1980s vs. 2000s or later), or study quality (low vs high). In view of the extreme heterogeneity and the existence of apparent effect moderators, probably the overall OR of 1.18 is not very informative. I wonder what the estimates may have been if limited to clinical diagnosis of depression in Western countries after 2000s based on moderate to high quality scores. In the discussion, the authors point out the similarity of their findings with those from the latest Global Burden of Disease study. The latter reported 13.5% increase in the ten years between 2007 and 2017 in the total years lived with disability (YLD) because of mental disorders (excluding substance use disorders): When standardized for age, however, the increase disappeared and the change was 1.1% (12). By way of reference, it is good to remember that the all-cause YLD increased by 17.0% in raw counts and by 0.9% when age-adjusted during the same 10 years.

中文翻译：

---

一场流行病还是一场常见精神障碍的瘟疫？

精神病学期刊上的许多文章都是通过关注精神障碍的严重性来开始介绍的。因此，例如，一篇文章的开头可能是“抑郁症是世界范围内导致残疾的主要原因”。2005 年至 2015 年间，患有抑郁症的人数增加了约 18%，现在影响了 3.22 亿人，约占世界总人口的 4%”，仅引用我最近的一个案例 (1)。在其他学术文章和流行书籍中可以很容易地找到类似的调性，吹捧诸如“流行病”（2、3）、“瘟疫”（4）或“大流行”（5）等耸人听闻的词。这种说法有什么好的证据，无论它们听起来多么逼真和令人震惊？一再指出，2000 年至 2015 年期间，高收入国家抗抑郁药的处方量翻了一番（6），目前美国或澳大利亚等国家有超过 10% 的成年人口服用抗抑郁药。实际接受治疗的可诊断精神障碍患者的比例也从 1990 年代美国的大约 20-33% (7) 或澳大利亚 2000 年代的 37% 增加到 46% (8)。这些事实肯定会造成流行病或瘟疫的印象。然而，我们必须记住，一方面，抗高血压和抗糖尿病药物的消费量同样翻了一番，而降胆固醇药物的消费量在同样的 15 年中翻了两番 (6)，另一方面，在抑郁症的情况下，似乎存在严重的过度诊断和过度处方 (9)。抑郁症和焦虑症的患病率是否增加了？——这就是问题，这个问题并不容易回答，因为从逻辑上讲，它需要大规模的调查，多年来重复，使用足够相似的方法。难怪文献报道了混合和相互矛盾的发现。我们必须记住，在某些情况下，即使这些研究显然使用了类似的方法，各国估计患病率的差异程度也无法合理解释：例如，世界心理健康调查倡议的一份报告总结说，每年抑郁症在德国或意大利仅为 3.0%，但在法国达到 5.9%，在美国达到 8.3% (10)。如果在下一次调查中德国的这一数字上升到 5.9%，我们是否可以认为在此期间患病率翻了一番？Richter 和他的同事在本期 (11) 开始通过系统地回顾在同一地理区域重复使用相同方法的流行病学调查来回答这个具有挑战性但极其重要的问题。他们发现了 42 项此类研究，比较了平均相隔 9.9 年的两个时间点之间各种疾病的患病率，包括抑郁症、痛苦、酒精依赖和一般精神疾病。合并 OR 为 1.18（95% 置信区间：1.07–1.31）。作者得出结论，成人精神疾病的患病率仅略有增加。研究之间的异质性非常高，I 平方为 96%。亚组分析表明，由于调查方法（临床诊断与自我报告症状）、目标精神状况（一般精神疾病与物质依赖）、几十年（1980 年代与 2000 年代或稍后），或学习质量（低与高）。鉴于极端的异质性和明显效应调节剂的存在，1.18 的总体 OR 可能不是非常有用。我想知道如果仅限于 2000 年代之后西方国家基于中等至高质量评分的抑郁症临床诊断，估计会是什么。在讨论中，作者指出他们的发现与最新的全球疾病负担研究的发现相似。后者报告了13。在 2007 年至 2017 年间的十年间，因精神障碍（不包括物质使用障碍）而导致残疾 (YLD) 的总年数增加了 5%：然而，当对年龄进行标准化时，这种增加消失了，变化为 1.1%（12 ）。作为参考，最好记住，在同一 10 年中，原始计数的全因 YLD 增加了 17.0%，年龄调整后增加了 0.9%。