It is easy to agree with Kirkbride et al¹ that a causal link exists between social factors and later mental health. Indeed, when the term “social factors” is defined as broadly as it is in their paper to include biological exposures due to the physical environment, we know from population genetics that social factors (i.e., the environment) are the most important causes (i.e., heritability is less than 50%) of most mental disorders². Furthermore, as genetic disorders cannot be prevented other than through lifestyle changes, it is easy to agree that broadly-defined social determinants are the most modifiable causes of mental disorders³.

Much more interesting issues are those involving complexities in the implementation of interventions. To this point, even though broadly-defined social determinants (i.e., the environment) are more modifiable than other (i.e., genetic) determinants of mental health, this broad statement provides little guidance for action. It is important to appreciate in this regard that when “social factors” are defined as broadly as they are here, any policy is an “intervention”. This means that the fields of macroeconomics, community psychology, and health care policy, as well as all policy decisions regarding such things as housing, preschool education programs, foster care and community policing, become of psychiatric interest. But these policies influence much more than mental health. And mental health is seldom a major consideration of policy makers in these areas. Even if it was, the population-level effects of these policies on mental health are largely unknown. And the complexities involved in providing even rough estimates of these effects are daunting.

Other complexities exist in designing interventions even in situations where causal effects are clear and where there are no competing interests across outcome domains. Indeed, there is often a trade-off between population optimality with respect to a point estimate and to a variance of the desired outcome. To illustrate, consider the question of where to build the next firehouse in a large metropolitan area where risk of a fire varies across neighborhoods (e.g., poor neighborhoods with older construction at higher risk), individual-level risk of death when a fire occurs also varies across neighborhoods (higher in neighborhoods with older construction), and expected number of deaths when a fire occurs varies in a different way across neighborhoods (e.g., higher expected number of deaths in high-rise buildings with many residents and exclusive egress via elevators than in smaller low-rise buildings). Given these and other inputs, operations research models can determine the optimal location for building the next firehouse to minimize overall population loss of life. However, the optimal location from that perspective might increase inequality of risk, which means that quite a different location would be selected if the goal was to equalize risk of death rather than to minimize loss of life. How do we decide which location to choose? The answer is anything but clear when competing considerations exist and resources are constrained.

Similarly difficult decisions are made every day on a smaller scale by practicing psychiatrists as they decide how to allocate their fixed clinical resources. These decisions are made in the context of higher-level decisions about allocation of health care resources (e.g., to community prevention vs. treatment). And these health care system-level decisions, in turn, are made in the context of even higher-level government decisions about the organization and financing of health care and the relative allocation of public resources across multiple sectors. Decisions at lower levels are inevitably constrained by prior decisions made at higher levels.

What are psychiatrists to do in the face of this complexity? Most psychiatrists focus on optimizing the resources available to them in their practice. Other psychiatrists consider social determinants of health in clinical decision-making⁴. And, at the extreme, some few psychiatrists change profession and become health care administrators or politicians to increase their impact on population mental health with higher-level decisions.

There is a need to focus on intervention opportunities for social determinants of mental health that are within the reach of psychiatrists in their own practices and local health care systems. For example, the Moving to Opportunity housing experiment discussed by Kirkbride et al was a massive ($100M+) macro policy intervention funded by the US Government's Department of Housing and Urban Development and carried out by an interdisciplinary team led by welfare economists, not by psychiatrists⁵. Other macro interventions shown to influence population mental health in cities and states are highlighted by the Results First Clearinghouse Network (RFCN)⁶, a network of nine clearinghouses aggregating evidence about community-level interventions of diverse sorts found to work in the US; and by the Institute of Health Equity (IHE) of University College London⁷, which has implemented and evaluated a wide range of coordinated area-level universal, selected and indicated interventions designed to advance six policy objectives⁸: give every child the best start in life; enable all children, young people and adults to maximize their capabilities and have control over their lives; create fair employment and good work for all; ensure a healthy standard of living for all; create and develop healthy and sustainable places and communities; and strengthen the role and impact of ill-health prevention.

But all the above programmes require deep buy-in by city, county and state governments, substantial coordination across sectors, and implementation of coordinated series of interventions designed to address the fact that social disadvantage is usually over-determined. Although psychiatrists could help achieve this buy-in to such interventions by participating in organized lobbying efforts, their perspective will inevitably take a back seat to those of other more powerful lobbying groups.

However, there are other interventions that individual psychiatrists could implement right now on their own by taking social factors into consideration in their practices and making use of local resources to help foster the wrap-around services that are often needed by disadvantaged patients⁴. In addition, groups of psychiatrists working in local health care systems could be instrumental in having their systems implement a mix of universal, selected and indicated interventions on social determinants of mental health that could have profound effects on population physical and mental health. A good guide in this respect is provided by the OASIS framework⁹, developed by the VA Boston Health Care System, which outlines the potential mechanisms by which health care-based social need interventions can improve health outcomes. Nonetheless, there is a need to distinguish the various levels of intervention on social determinants of mental health, and to more clearly identify and promote those that are within the reach of psychiatrists in their own practices.

人们很容易同意 Kirkbride 等人的观点¹，即社会因素与后来的心理健康之间存在因果关系。事实上，当“社会因素”一词被广泛定义为包括物理环境引起的生物暴露时，我们从群体遗传学中得知，社会因素（即环境）是最重要的原因（即，大多数精神障碍的遗传力低于50% ²。此外，由于遗传性疾病只能通过改变生活方式来预防，因此很容易同意广义的社会决定因素是精神障碍最可改变的原因³。

更有趣的问题是那些涉及实施干预措施的复杂性的问题。就这一点而言，尽管广义的社会决定因素（即环境）比心理健康的其他决定因素（即遗传）更容易改变，但这一广义的陈述几乎没有为行动提供指导。在这方面，重要的是要认识到，当“社会因素”被定义得如此广泛时，任何政策都是“干预”。这意味着宏观经济学、社区心理学和医疗保健政策领域，以及有关住房、学前教育计划、寄养和社区警务等所有政策决策，都成为精神病学的兴趣。但这些政策的影响远不止心理健康。心理健康很少成为这些领域政策制定者的主要考虑因素。即使是这样，这些政策对心理健康在人口层面上的影响在很大程度上还是未知的。即使对这些影响进行粗略估计，其复杂性也令人望而生畏。

即使在因果效应明确且结果领域之间不存在竞争利益的情况下，设计干预措施也存在其他复杂性。事实上，关于点估计的总体最优性和期望结果的方差之间经常存在权衡。为了说明这一点，请考虑在大都市区的何处建造下一个消防站的问题，其中各个社区的火灾风险各不相同（例如，建筑陈旧的贫困社区风险较高），发生火灾时的个人死亡风险也不同各个社区之间的预期死亡人数有所不同（建筑较旧的社区中的预期死亡人数较高），并且各个社区发生火灾时的预期死亡人数也有不同的差异（例如，在有许多居民且通过电梯专用出口的高层建筑中，预期死亡人数高于在有电梯的高层建筑中）。在较小的低层建筑中）。考虑到这些和其他输入，运筹学模型可以确定建造下一个消防站的最佳位置，以最大限度地减少总体人口生命损失。然而，从这个角度来看，最佳地点可能会增加风险的不平等，这意味着如果目标是均衡死亡风险而不是最大限度地减少生命损失，就会选择完全不同的地点。我们如何决定选择哪个位置？当存在相互竞争的考虑因素且资源受到限制时，答案绝非明确。

执业精神科医生在决定如何分配固定临床资源时，每天都会做出类似的困难决定。这些决定是在有关医疗保健资源分配的更高级别决策的背景下做出的（例如，社区预防与治疗）。反过来，这些医疗保健系统层面的决策是在更高级别的政府决策的背景下做出的，这些决策涉及医疗保健的组织和融资以及跨多个部门的公共资源的相对分配。下级决策不可避免地受到上级决策的制约。

面对这种复杂性，精神科医生该怎么办？大多数精神科医生专注于优化他们在实践中可用的资源。其他精神科医生在临床决策中考虑健康的社会决定因素⁴。而且，在极端情况下，一些精神科医生会改变职业，成为医疗保健管理者或政治家，以通过更高级别的决策来增加对人口心理健康的影响。

有必要关注心理健康社会决定因素的干预机会，这些干预机会是精神科医生在自己的实践和当地医疗保健系统的能力范围内的。例如， Kirkbride 等人讨论的转向机会住房实验是一项大规模（1 亿美元以上）宏观政策干预，由美国政府住房和城市发展部资助，由福利经济学家而非精神病学家领导的跨学科团队实施⁵ . ^{结果第一信息交换所网络 (RFCN) 6}强调了影响城市和州人口心理健康的其他宏观干预措施，该网络由九个信息交换所组成，汇集了在美国有效的各种社区层面干预措施的证据；伦敦大学学院健康公平研究所 (IHE) ⁷实施并评估了一系列协调一致的地区级普遍性、选择性和指示性干预措施，旨在推进六项政策目标⁸：为每个儿童提供最好的开始生活; 使所有儿童、青少年和成人能够最大限度地发挥自己的能力并掌控自己的生活；为所有人创造公平就业和良好工作；确保所有人的健康生活水平；创建和发展健康和可持续的场所和社区；加强疾病预防的作用和影响。

但所有上述计划都需要市、县和州政府的深度支持，跨部门的实质性协调，以及实施一系列协调一致的干预措施，以解决社会劣势通常是被过度决定的事实。尽管精神病学家可以通过参与有组织的游说活动来帮助实现对此类干预措施的认可，但他们的观点将不可避免地让位于其他更强大的游说团体的观点。

然而，个别精神科医生现在可以自行实施其他干预措施，在他们的实践中考虑社会因素，并利用当地资源来帮助促进弱势患者经常需要的全方位服务⁴。此外，在当地卫生保健系统中工作的精神科医生小组可以帮助他们的系统对心理健康的社会决定因素实施普遍的、有选择的和有针对性的干预措施，这些干预措施可能对人口的身心健康产生深远的影响。^{VA 波士顿医疗保健系统开发的 OASIS 框架9}提供了这方面的良好指南，该框架概述了基于医疗保健的社会需求干预措施可以改善健康结果的潜在机制。尽管如此，有必要区分对心理健康社会决定因素的不同干预水平，并更明确地识别和促进精神科医生在自己的实践中力所能及的干预措施。