Fusar‐Poli and eminent colleagues¹ con­clude their encyclopedic review of prevention in psychiatry by calling for governments to tackle inequalities in young ­people's men­tal health and to invest in improving its so­cial determinants: education, employment, social care, housing, criminal justice, poverty alleviation, social security/welfare benefits, community development, and immigration. We stand firmly with Fusar‐Poli et al on this position and would add social justice and public safety to the list. Academics as individuals and their institutions and professional organizations should assist governments to pursue youth mental health as a top priority.

We further commend Fusar‐Poli et al for their scholarly review of prevention concepts and in particular their noting that both the public health framework and the World Health Organization framework provide the possibility that some disorders carry risk for other disorders and that conceptual boundaries between preventive and treatment interventions can be porous. We often hear in academic discussions that an intervention must be either preventive or a treatment and that an entity must be defined and named either by risk or by severity, as in clinical high risk (CHR)² vs. attenuated psychosis, or prodromal Alzheimer's disease vs. mild cognitive impairment. Our view has long been that the same intervention can provide both treatment and prevention, and that CHR is both a disorder and an indicator of risk for future more severe disorders. In this context, the term “risk syndrome”³ may be preferable.

We may part ways, however, with Fusar‐Poli and colleagues on the relative roles of universal and indicated prevention. Notwithstanding the promise of interventions such as phosphatidylcholine and folic acid tested against surrogate biomarkers, the authors' extensive review sadly identifies few if any universal or selective interventions that meet effectiveness, cost‐effectiveness, and implementation standards for reducing the incidence of any mental disorder. The authors' contention that universal public health approaches hold the greatest potential for reducing the risk profile of the whole population does not seem predicated on empirical evidence but rather on theoretical potential.

Along those lines, we take issue with the authors' conceptual Figure 1, partly the basis for their advocacy for universal prevention. This figure shows universal prevention shifting the curve between spectrum of risk and numbers of people to the left, such that there would appear to be no people remaining in the highest risk group who would require indicated prevention. Rather than a shift of a normal curve's x‐intercepts to the left, under a universal approach we would expect to see a skewing of the curve such that the risk x‐intercepts remain fixed, the left side becomes steeper and higher, indicating a larger number of persons at lower risk, and the right side flattens, indicating a smaller but not zero number of persons at higher risk.

In our alternate conceptualization, there would be a continued need for indicated prevention even under conditions of successful universal prevention. This situation appears to be what occurred in the authors' appropriate example of reducing tobacco use in the population, where new incident cases of non‐small cell lung cancer have been reduced by anti‐tobacco measures but have not been eliminated⁴.

Fusar‐Poli et al do advocate for combining universal and indicated prevention, and we staunchly support that advocacy. The non‐small cell lung cancer example⁴, where mortality has diminished faster than incidence due to the availability of ef­fective new treatments, demonstrates the value at least of tertiary prevention and a potential role for indicated prevention even in the context of effective universal pre­vention.

With regard to the CHR syndrome as a vehicle for indicated prevention of psychosis, one of the recent criticisms of the approach, echoed by Fusar‐Poli et al, derives from the NEMESIS‐2 cohort report that antecedent mood disorders account for more of the incidence of clinical psychosis than do psychotic‐like symptoms⁵. We see three important limitations of the NEMESIS‐2 data that have received little attention. First, psychotic‐like experiences gauged through questionnaires or non‐clinical interviews in the general population are not comparable to clinician‐assessed CHR syndromes⁶. Second, the time‐points in NEMESIS‐2 were spaced three years apart. Partly‐prospective data show that the average duration of CHR symptoms is two years or less in two‐thirds of patients converting to psychosis⁷, suggesting that the development of psychotic‐like symptoms prior to psychosis may have been missed by the NEMESIS‐2 design in as many as half the cases. Third and most crucially, the average age of cohort members at the second time point was 47.7 years, far older than the 12‐to‐early 30s range where CHR has been reported to predict psychosis and where the incidence of psychosis is known to be highest⁸. As a consequence of these limitations, in our view the NEMESIS‐2 data are only partially relevant to the value of CHR as a vehicle for indicated prevention.

With regard to evidence for the success of preventive interventions for CHR, Fusar‐Poli et al rightly point out that meta‐analytic evidence so far is contradictory and that clinical trials featuring conversion to psychosis as their primary outcome require very large sample sizes. We do, however, see hope on the horizon. This past fall the US National Institute of Mental Health and the Foundation for the National Institutes of Health announced the Accelerated Medicines Partnership in Schizophrenia (AMP SCZ), a collaborative effort to advance early intervention for CHR individuals⁹. This initiative seeks to identify parameters for future clinical trials on alternate outcomes of CHR such as social functioning or attenuated positive symptoms. These alternate endpoints can also potentially serve as surrogate outcomes for reducing the incidence of psychopathology, which can then be investigated directly after entry of the new treatments into clinical practice through epidemiologic methods.

In conclusion, our view is not only that a combined universal and indicated approach is likely to be the best way to prevent psychosis in the future, but also that the CHR syndrome for psychosis continues to provide the most promising option for the indicated prevention component. We acknowledge a potential bias, working as we do in the CHR field, but we like to think we chose this field because it offers the best opportunities in psychiatry for improving public health rather than that we believe it offers the best opportunities for public health because we have chosen it.

Fusar-Poli 和杰出同事¹结束了对精神病学预防的百科全书式审查，呼吁各国政府解决年轻人心理健康方面的不平等问题，并投资于改善其社会决定因素：教育、就业、社会关怀、住房、刑事司法、扶贫、社会保障/福利、社区发展和移民。我们坚定地支持 Fusar‐Poli 等人的这一立场，并将把社会正义和公共安全纳入其中。学术界作为个人及其机构和专业组织应协助政府将青少年心理健康作为首要任务。

我们进一步赞扬 Fusar-Poli 等人对预防概念的学术审查，特别是他们指出，公共卫生框架和世界卫生组织框架都提供了一种可能性，即某些疾病可能会带来其他疾病的风险，并且预防性和预防性之间的概念界限也很重要。治疗干预措施可能是多孔的。我们经常在学术讨论中听到，干预措施必须是预防性的或治疗性的，并且必须根据风险或严重性来定义和命名实体，例如临床高风险 (CHR) ²与减轻精神病或前驱阿尔茨海默病与轻度认知障碍。长期以来，我们的观点是，同样的干预措施可以提供治疗和预防，CHR 既是一种疾病，也是未来更严重疾病风险的指标。在这种情况下，术语“风险综合症” ³可能更合适。

然而，我们可能与 Fusar-Poli 及其同事在普遍预防和针对性预防的相对作用上存在分歧。尽管磷脂酰胆碱和叶酸等干预措施有望针对替代生物标志物进行测试，但令人遗憾的是，作者的广泛审查发现，很少有普遍或选择性的干预措施能够满足降低任何精神障碍发病率的有效性、成本效益和实施标准。作者认为全民公共卫生方法最有可能降低整个人群的风险状况，这一观点似乎不是基于经验证据，而是基于理论潜力。

沿着这些思路，我们对作者的概念性图 1 提出质疑，该图在一定程度上是他们倡导普遍预防的基础。该图显示，普遍预防将风险谱与人数之间的曲线向左移动，这样一来，最高风险群体中似乎就没有需要指定预防的人了。在通用方法下，我们期望看到曲线的倾斜，使得风险 x 截距保持固定，左侧变得更陡和更高，而不是正常曲线的 x 截距向左移动，这表明风险更大风险较低的人数，右侧变平，表明风险较高的人数较少但不为零。

在我们的替代概念中，即使在成功进行普遍预防的情况下，也仍然需要有针对性的预防。这种情况似乎是作者减少人群烟草使用的适当示例中所发生的情况，其中非小细胞肺癌的新发病例已通过反烟草措施减少，但尚未消除⁴ 。

Fusar-Poli 等人确实主张将普遍预防和针对性预防相结合，我们坚决支持这一主张。非小细胞肺癌示例⁴中，由于有效新疗法的出现，死亡率下降速度快于发病率下降速度，这至少证明了三级预防的价值，以及即使在有效普遍预防的背景下，指示性预防的潜在作用。

关于 CHR 综合征作为预防精神病的一种手段，最近对该方法的批评之一（Fusar-Poli 等人对此表示赞同）源自 NEMESIS-2 队列报告，即先前的情绪障碍占发病率的大部分临床精神病的发生率高于精神病样症状⁵ 。我们发现 NEMESIS-2 数据的三个重要局限性很少受到关注。首先，通过问卷调查或非临床访谈在一般人群中评估的精神病样经历与临床医生评估的 CHR 综合征不可比较⁶ 。其次，NEMESIS-2 中的时间点间隔三年。部分前瞻性数据显示，三分之二转变为精神病的患者中，CHR 症状的平均持续时间为两年或更短⁷ ，这表明 NEMESIS-2 设计可能忽略了精神病之前精神病样症状的发展在多达一半的情况下。第三，也是最关键的是，第二个时间点队列成员的平均年龄为 47.7 岁，远高于据报道 CHR 可以预测精神病且精神病发病率最高的 12 岁至 30 岁出头的范围。 ⁸ .由于这些限制，我们认为 NEMESIS-2 数据仅与 CHR 作为指示性预防工具的价值部分相关。

关于 CHR 预防性干预措施成功的证据，Fusar-Poli 等人正确地指出，迄今为止的荟萃分析证据是相互矛盾的，并且以转变为精神病为主要结果的临床试验需要非常大的样本量。然而，我们确实看到了希望。去年秋天，美国国家心理健康研究所和美国国立卫生研究院基金会宣布了精神分裂症加速药物合作伙伴关系 (AMP SCZ)，这是一项旨在推进 CHR 个体早期干预的合作项目⁹ 。该计划旨在确定未来关于 CHR 替代结果（例如社会功能或减弱的阳性症状）的临床试验的参数。这些替代终点也可以作为减少精神病理学发生率的替代结果，然后可以在新疗法进入临床实践后通过流行病学方法直接进行调查。

总之，我们的观点是，不仅普遍性和指示性方法相结合可能是未来预防精神病的最佳方法，而且治疗精神病的 CHR 综合征继续为指示性预防成分提供最有希望的选择。我们承认，正如我们在 CHR 领域所做的那样，存在潜在的偏见，但我们喜欢认为我们选择这个领域是因为它为改善公共卫生提供了精神病学的最佳机会，而不是因为我们认为它为公共卫生提供了最佳机会，因为我们选择了它。