Fusar-Poli et al¹ provide a scholarly and detailed overview of the state of knowledge on preventive approaches in psychiatry. Their paper should be considered an obligatory read for anyone entering or already practicing in this emerging field.

The need for preventive approaches in psychiatry is readily transparent. According to the US National Comorbidity Survey², a nationally representative population-based survey of mental disorders, one in two adults in the US suffers from the symptomatic and functional challenges of one mental disorder during his/her lifetime. Almost one in three adults will suffer from two or more mental disorders. Regrettably, like much else in psychiatry, preventive approaches are lagging behind general medicine. Fusar-Poli et al make strong arguments about several crucial challenges that critically hamper the implementation of preventive strategies. Here we elaborate on some of the key challenges mentioned in the review, and introduce a set of possible solutions to those.

The primary challenge is finding those who are at risk. Despite the longstanding history of neurobiological research, the underlying causal mechanisms of mental disorders remain mostly unknown. Symptom ratings have been widely used in psychiatry to detect individuals at risk. However, outside of specialty clinics, this strategy seems prone to failure. In a population-based study of 18 to 21-year-olds³, the presence of symptoms, while associated with subsequent hospitalization for mental disorders, had positive predictive values ranging from 0.54% to 1.99%. In other words, for every correctly identified “case”, there would be between 50 and 200 “non-cases” that would be incorrectly identified as “cases”. Such a high false-positive detection rate, often found when prodrome studies are extrapolated to the general population, questions the utility of current paradigms that aim to identify at-risk groups for large-scale preventive efforts.

Advances in genetic research have identified some syndromic cases across multiple mental disorders, yet the overwhelming majority of individuals with these disorders, and especially those with common disorders (depression, anxiety), are idiopathic, with an unknown etiology. Target­able biomarkers are unavailable to use for early detection and/or efficient early intervention. As Fusar-Poli et al¹ note, only two of 162 peripheral biomarkers were reliably associated with psychosis, bipolar disorder, or depression. Collectively, our current lack of both understanding of underlying causal mechanisms and targetable biomarkers for mental disorders that can be applied at the population level substantially limit preventive strategies.

An additional challenge is that even early intervention often comes too late. Considerable evidence from genetics, epidemiology, basic neuroscience, and neuroimaging implicates early neurodevelopment as the critical period for the risk of developing most mental disorders. Almost all mental disorders are recognizable before or during the second decade of life. Yet, atypical neurobiological development surely predates the emergence of many mental disorders. For instance, evidence suggests that the first signs of cognitive abnormalities in those who will later develop schizophrenia are detectable by the age of four – decades before the disorder is usually diagnosed⁴. Furthermore, the brain most rapidly develops in utero, and continues to do so during early childhood. Indeed, evidence in children of patients with schizophrenia implicates aberrant early, possibly prenatal, brain development⁵. Therefore, these early periods are those when preventive strategies are most likely to have an impact. Fusar-Poli et al¹ highlight this point, but it is transparent that targeting this developmental period is particularly challenging.

A final challenge underscores how we address comorbidities². Comorbidity rates are high in psychiatry and conform to a 50% rule. Approximately half of all people with one psychiatric disorder meet the criteria for a second disorder concurrently; half the people with two disorders meet the criteria for a third; and so on. Evidence based on multiple studies highlights a general underlying dimension, termed the p factor, which captures the tendency to develop psychopathology. In the Dunedin Multidisciplinary Health and Development Study, conducted in an unselected longitudinal birth cohort, higher scores on the general tendency to psychopathology were associated with compromised early-life brain function, and impairments in maturation⁶. Such findings foster the debate regarding categorical versus dimensional models that are relevant to research and in the clinic. In sum, since psychiatric disorders often co-occur, the challenge to clinicians is how to target higher-order psychopathological dimensions and the p factor without loss of specificity⁷.

A possible way to address these challenges is to identify those cases that will contribute disproportionally to morbidity and mortality. One source of intriguing evidence comes from another study of the Dunedin Multidisciplinary Health and Development sample, showing that 80% of the health burden is attributable to 20% of cases⁸. That study showed that early-life factors (familial socioeconomic characteristics, maltreatment, IQ, and self-control) clustered into 20% of the population, that accounted for disproportionately high levels of health care use (e.g., 78% of prescription fills and 57% of hospital nights). These findings imply that early life is a critical period for preventive measures for a select group in the population. However, there is potential to abuse this approach; population segments may suffer from stigma. Nevertheless, easing the effects of childhood disadvantage is a critical aim which, if attained in early life, may support families and children, as well as benefit all of society.

A second alternative is to implement universal psychiatric prevention. General medicine has advanced in this prevention (e.g., the efficacy of the COVID-19 vaccines). Evidence-based examples in psychiatry are few, but there are some, such as means restriction to prevent suicide, and physical activity to prevent incident anxiety and depression⁹. Selective universal prevention subtly differs by stratifying prevention to a large group in the population (e.g., nutrient use among pregnant women and the elderly). Better designed, easier to administer universal prevention strategies have the potential to reduce incident mental disorders. They may involve a significant financial investment, but also indirect benefits, including improvements in general health, unemployment, and even crime.

A third alternative is to target not the outcome but an effect modifier for intervention. While biomarkers for mental disorders are not yet available, it is well established that cognitive impairment accompanies, and most often predates by many years, the onset of the majority of mental disorders. There are also reliable ways to measure cognitive functioning and plausible intervention strategies. Implementing interventions to ameliorate cognitive impairments early in life may be a means for psychiatric prevention with substantial societal benefits beyond prevention of psychiatric outcomes (e.g., increasing the cognitive reserve in midlife may be a strategy to reduce dementia).

So, there are multiple challenges to implementing preventive strategies in psychiatry. There is, however, a clear need, and the time is ripe to make the leap towards primary and secondary prevention pathways in the critical period of early life and via cognition.

Fusar-Poli等人¹提供了有关精神病学预防方法知识的学术和详细概述。对于进入或已经在该新兴领域执业的任何人，应将其论文视为必读。

精神病学中对预防方法的需求是显而易见的。根据美国全国性合并症调查²，这是一项具有全国代表性的基于人群的精神障碍调查，在美国，每两个成年人中就有一个在其一生中遭受一种精神障碍的症状和功能挑战。几乎三分之一的成年人将患有两种或更多种精神疾病。令人遗憾的是，与精神病学中的其他许多方面一样，预防方法也落后于普通医学。Fusar-Poli等人就一些严重阻碍预防策略实施的关键挑战提出了有力的论据。在这里，我们详细介绍了审查中提到的一些关键挑战，并针对这些挑战介绍了一套可能的解决方案。

主要的挑战是找到处于危险之中的人。尽管神经生物学研究已有很长的历史，但精神障碍的潜在因果机制仍是未知之数。症状分级已在精神病学中广泛用于检测有风险的个体。但是，在专科诊所之外，这种策略似乎容易失败。在一项针对18至21岁³的人群的研究中，症状的出现与随后因精神障碍而住院相关，其阳性预测值为0.54％至1.99％。换句话说，对于每个正确识别的“案例”，将有50到200个“非案例”不正确被确定为“案件”。如此高的假阳性检出率（通常在将前驱研究推算到普通人群时发现），对旨在识别高风险人群进行大规模预防工作的当前范式的实用性提出了质疑。

基因研究的进展已经确定了多种精神疾病的综合症病例，但是绝大多数患有这些疾病的个体，尤其是那些患有常见疾病（抑郁，焦虑）的个体是特发性的，病因不明。无法将可靶向的生物标记物用于早期检测和/或有效的早期干预。正如Fusar-Poli等人^1所指出的那样，在162种外周生物标志物中，只有两种可靠地与精神病，躁郁症或抑郁症相关。总体而言，我们目前对潜在的病因机制和可用于人群水平的精神障碍的可靶向生物标志物既缺乏了解，也极大地限制了预防策略。

另一个挑战是，即使早期干预也常常来不及。来自遗传学，流行病学，基础神经科学和神经影像学的大量证据表明，早期神经发育是发生大多数精神障碍风险的关键时期。在生命的第二个十年之前或之中，几乎所有的精神障碍都是可以识别的。然而，非典型的神经生物学发展肯定早于许多精神疾病的出现。例如，证据表明，在后来发展为精神分裂症的人中，认知异常的最初迹象可在通常被诊断出障碍的四岁至几十岁之前被发现⁴。此外，大脑在子宫内发育最快，并在儿童早期就一直如此。确实，精神分裂症患者儿童的证据暗示异常早期发育，可能是产前大脑发育⁵。因此，这些早期阶段是预防策略最有可能产生影响的时期。Fusar-Poli等[ ^1]强调了这一点，但是透明的是，针对这个发育时期尤其具有挑战性。

最后的挑战强调了我们如何应对合并症²。合并症的精神病学率很高，符合50％的规定。患有一种精神疾病的所有人中约有一半同时符合第二种疾病的标准；一半患有两种疾病的人符合三分之一的标准；等等。基于多项研究的证据强调了一个基本的基本维度，称为p因子，它反映了发展心理病理学的趋势。在未选择的纵向出生队列中进行的但尼丁多学科健康与发展研究中，精神病理学总体趋势得分较高与早期脑功能受损和成熟障碍有关⁶。这些发现引发了关于与研究和临床相关的分类模型与尺寸模型的争论。总而言之，由于精神疾病常常同时发生，因此临床医生面临的挑战是如何针对更高阶的精神病理学维度和p因子而又不丧失特异性⁷。

解决这些挑战的一种可能方法是确定那些会导致发病率和死亡率不成比例的案件。有趣的证据之一来自但尼丁多学科健康与发展样本的另一项研究，表明80％的健康负担可归因于20％的病例⁸。该研究表明，生命早期因素（家庭社会经济特征，虐待，智商和自我控制）聚集在20％的人口中，占医疗保健使用比例过高的比例（例如78％的处方药和57住院天数的百分比）。这些发现表明，早年是人群中某些特定人群采取预防措施的关键时期。但是，有可能滥用这种方法。人群可能遭受耻辱。然而，减轻儿童期不利的影响是一个关键目标，如果能在早年实现，就可以支持家庭和儿童，并使全社会受益。

第二种选择是实施普遍的精神病预防。普通医学在这种预防方面已经取得了进展（例如，COVID-19疫苗的功效）。精神病学中基于证据的例子很少，但也有一些例子，例如限制手段以防止自杀，进行体育活动以防止突发事件引起的焦虑和抑郁⁹。选择性全民预防的不同之处在于对人群中的一大批人群进行了分层预防（例如，孕妇和老年人之间的营养使用）。更好地设计，更易于实施的普遍预防策略有可能减少突发性精神障碍。它们可能涉及大量的财务投资，但也涉及间接利益，包括改善总体健康状况，失业甚至犯罪。

第三种选择不是针对结果，而是针对干预的效果修正。尽管尚无用于精神障碍的生物标志物，但众所周知，认知障碍伴随着大多数精神障碍的发作，而且通常早在许多年之前就已出现。还有可靠的方法来衡量认知功能和合理的干预策略。实施干预措施以改善生命早期的认知障碍，可能是精神疾病预防的一种手段，除了预防精神疾病的结果外，还具有巨大的社会效益（例如，增加中年的认知储备可能是减少痴呆的一种策略）。

因此，在精神病学中实施预防策略存在多个挑战。但是，有明确的需求，并且在早年的关键时期并通过认知朝着一级和二级预防途径迈进的时机已经成熟。