横断面研究能否进行因果推断

基于变量调查(或测量)的共时性、统计学关联及幸存者偏倚等原因,横断面研究被认为不能进行因果推断,这是当前的"共识"。本文基于因果思维,借助因果图,首先明确定义真实截面和测量截面,并识别截面概念仅存在于理论的特性。实际横断面研究中,测量变量的共时性并不存在,而是无一例外地表现为非共时性时序,其实质上相当于测量变量间互为独立性假设,或不存在有差别错分偏倚。类似于累积病例对照研究和历史性队列研究,横断面研究均为暴露和结局已存在或发生后进行的测量,这种测量相当于对变量值的历史重建或"考古"。这类研究进行因果推断的共性前提条件之一是,测量变量与其历史变量间必须存在着因果律。测量变量均为真实变量的替代者,测量变量间的时序在因果推断上并不重要。应加强对横断面研究分析性角色的认识。     

因果推断中研究人群的形成

流行病学是对一定数量的人群进行特征描述和比较,并在此基础上进行因果推断。研究人群的形成是其第一步。本研究以观察性研究为例,首先定义个体截面和人群截面,并阐明其测量需满足的3个假设：属性真实值随时间保持不变,属性变量间互不干扰,个体间互不干扰;接着指出因果推断研究应以待定因（或暴露）的发生或状态开始的时间为标准进行统一;最后,基于人群截面的双重角色,提出人群的因果推断研究可分为2类：历史重建研究和探索未来研究,并初步梳理了研究设计框架、估计的效应及设计间的关系。从因果思维角度探讨研究人群的形成过程,可为明确因果推断研究设计类型奠定基础,选取合适的效应估计进行因果推断,值得深入研究。     

阴性对照法：原理、方法及应用

阴性对照法系基于 "阴性对照的实验检测结果必须阴性"的思想而建立,是人群研究多个比较组和关联特异度的进一步扩展。阴性对照不参与待研究的因果假设,但具有与研究假设相同的潜在偏倚结构。阴性对照法可分为阴性暴露对照法、阴性时期对照法和阴性结局对照法,主要用于探测和控制选择偏倚、混杂偏倚和测量偏倚,以补充样本量为辅。合理应用阴性对照法有助于提高试验法和观察法因果推断估计的准确性。     

界标分析在队列研究中的应用及实例分析

队列研究设计具有时序关系明确的特点，其论证因果关联的强度优于其他观察性研究，是分析性流行病学的重要研究方法之一。然而，队列研究对象的纳入过程中常采用筛检诊断或其他方式排除已出现结局事件的个体，而筛检诊断的准确性、排除的有效性会影响纳入研究个体基线状况评估的准确性，进而导致对因果效应的估计可能存在暴露-结局因果倒置。界标分析可以通过排除可能存在暴露-结局时序不明的研究对象以控制反向因果。本研究阐述界标分析的基本原理与分析步骤，并运用中国老年人健康长寿影响因素调查的数据探索体育锻炼与虚弱的关系，展示界标分析的具体应用，以期促进其在队列研究中的应用，从而更准确地推断暴露与结局的因果效应。     

孟德尔随机化研究应用于因果推断的影响因素及其结果解读面临的挑战

孟德尔随机化（Mendelian randomization,MR）研究使用遗传变异作为工具变量,推断暴露因素与结局之间的因果关系,能够有效克服混杂和反向因果问题所导致的偏倚。然而,MR研究中的工具变量须满足关联性、独立性和排他性3个核心假设。即使核心假设成立,MR研究在因果推断中的应用还受到其他局限性的影响。此外,MR研究结果的解读需要基于综合证据。本文将围绕MR研究应用于因果推断的影响因素和研究结果的解读进行综述,以期为MR研究结果应用提供指导。     

客观世界的因果关系:基于有向无环图的结构解析

客观世界因果关系的整体框架较为笼统而缺乏明晰的细节,给因果关系的研究带来困难。本文基于因果关系的时序特性结合有向无环图（DAGs）,以因和果的发生时间为界,将客观世界的时间维度划分为3个时间域和2个时间点。通过对5个时间单位上变量间存在着的完整的因果关系的病因网络DAGs进行分析发现,其病因结构由两部分叠加组成：第一部分是各个时间域间/时间点上任取一变量间的组合DAGs,为因果关系的基本结构,构成病因网络的核心,仅混杂路径影响其因果效应估计;第二部分是各个时间域内/时间点上变量间的母子或祖先-后代关系,其DAGs表现为与混杂类似的结构。本文简洁明了地构建了客观世界因果关系研究的整体框架（病因网络DAGs）,解释了控制混杂以解决因果效应估计的结构基础,为正确研究和识别因果关系奠定基础。     

暴发调查设计:以急性戊型肝炎暴发为例

疾病暴发一旦确立,则意味着疾病(及其病因或疑似暴露)已经发生。暴发是疾病流行强度指标之一,虽然其本身不具备设计要素,但暴发调查可有多种设计选择。本文采用单次戊型肝炎暴发案例,调查并获得学生们对其设计类型判断的三个答案:横断面设计、病例对照设计和历史队列设计。借助因果思维和因果图,本文发现该案例并不满足或符合上述设计,即只有暴发确立型病例是已知的、可被测量两次;调查实施前暴露是未知的,暴露与结局间可混合存在着两种纵向测量时序。基于调查人群的代表性、暴发确立型病例的两次识别特征,可以将此案例视为横断面队列设计,即在假设的历史人群基础上对当前人群进行的横断面研究。对结局(和暴露)已然发生的情形,因果推断的准确性取决于其历史重建的正确性,应加强对这类基于历史重建的研究设计类型的认识。     

在因果推断中应用有向无环图识别和控制选择偏倚

在流行病学研究中，选择偏倚会导致研究样本无法代表一般人群，使研究结果偏离真实值，无法推断真实的因果关联。本文通过构建有向无环图（directed acyclic graphs，DAGs），将复杂的因果关系可视化，提供识别选择偏倚的直观方法，并通过冲撞分层偏倚的图形结构来验证不同类型的选择偏倚。在实际研究中，可能同时存在多种偏倚，对冲撞变量进行不恰当的调整会新增冲撞分层偏倚，打开后门路径，引入混杂偏倚，甚至改变原有混杂偏倚的大小与方向。为了得到暴露到结局的无偏估计，研究者可以通过构建DAGs，帮助识别冲撞变量，防止冲撞偏倚的发生。     

流行病学知识问答

Ａ　流行病学研究中的因果推论Ａ１１７　何为因果推论？其基本步骤是什么？答：因果推论是指研究者根据流行病学研究资料，对某因素与某疾病（或健康状况）之间的因果关系作出正确判断的论证过程。其基本步骤包括：１．确定两事件间是否存在统计学上的联系。两事件间若存在因果联系，则两者间必然存在统计学上的联系，即在统计学上，两者在数量关系上有显著关联。２．判断两事件间统计学联系的性质。在统计学上呈现关联的两事件间，不一定为因果关系所致。因为在统计学上的相关，除因果联系外，也可能是由于各种偏倚或机遇所致的假相关——…     

现实队列研究中暴露与结局的时序关系

暴露在发生时间上先于结局,是队列研究的优点之一,因此在因果推断上优于其他观察性设计。本文应用有向无环图（Directed Acyclic Graphs,DAGs）构建了现实队列研究中易感人群的因果结构后发现：现实的队列研究以研究人群替换易感人群进行因果效应的估计,人群的暴露与结局在时序关系上可互为先后,因果效应估计的准确性受到替换人群的易感性和基线调查时结局识别和排除有效性的影响。     

观察与实验：因果视角

观察和实验及其设计、相关的概念和内涵仍然模糊,影响着研究设计的正确认识及因果推断有效性的判断。本文借用物理学的相变概念,结合因果思维与因果图,首先建立了事物的属性及其状态、事件与现象之间的关系;接着识别了两种现象获取方式——人为观察和人为实验之间相反的因果结构,及其与暴露的干预、分配在因果推断研究中受各自的因果机制影响;最后,识别了干预系因果概念,是自然因果律基础上联系已知的现象和/或需经测量而获得的现象之间的核心纽带;基于此分析了研究设计二分类的策略,以干预法和非干预法的分类更为清晰。观察和实验提供了所有科学知识的基础,应是内涵统一的概念。基于因果律及其测量过程可能是实现研究设计精准分类的最佳选项之一,值得深入研究。     

Activity spaces in studies of the environment and physical activity: A review and synthesis of implications for causality

Activity spaces are increasingly used to understand how people interact with their environment and engage in activity but their use may raise challenges regarding causal inference. We conducted a systematic review of findings and the methodological, analytical and conceptual issues relevant to causal inference. Studies were included if they comprised a spatial summary of locations visited, assessed any part of the causal pathway between the environment, physical activity and health, and used quantitative or qualitative methods. We searched seven electronic databases in January 2018 and screened 11910 articles for eligibility. Forty-seven studies were included for review. Studies answered research questions about features of or environmental features within activity spaces using a range of spatial and temporal summary techniques. The conceptual challenge of using activity spaces to strengthen causal inference was rarely considered, although some studies discussed circularity, temporality, and plausibility. Future studies should use longitudinal and experimental designs and consider the potential and actual use of spaces for physical activity, and their relationship with total levels of activity.     

因果图模型及其在营养流行病学研究中的应用

不良饮食是慢性非传染性疾病最重要的可控危险因素之一,但通过随机对照试验定量阐明具体饮食因素与健康结局的因果关联面临很多困难。近年来,因果推断的迅速发展为充分利用和发掘观察性研究数据,产生高质量的营养流行病学研究证据提供了有力的理论和方法工具。其中,因果图模型通过整合大量先验知识将复杂的因果关系系统可视化,提供了识别混杂和确定因果效应估计策略的基础框架,基于不同的因果图,可选择调整混杂、工具变量或中介分析等不同的分析策略。本文对因果图模型的思想和各种分析策略的特点及其在营养流行病学研究中的应用进行介绍,旨在促进因果图模型在营养领域的应用,为后续研究提供参考和建议。     

A Bayesian mixture of semiparametric mixed‐effects joint models for skewed‐longitudinal and time‐to‐event data

In longitudinal studies, it is of interest to investigate how repeatedly measured markers in time are associated with a time to an event of interest, and in the mean time, the repeated measurements are often observed with the features of a heterogeneous population, non‐normality, and covariate measured with error because of longitudinal nature. Statistical analysis may complicate dramatically when one analyzes longitudinal–survival data with these features together. Recently, a mixture of skewed distributions has received increasing attention in the treatment of heterogeneous data involving asymmetric behaviors across subclasses, but there are relatively few studies accommodating heterogeneity, non‐normality, and measurement error in covariate simultaneously arose in longitudinal–survival data setting. Under the umbrella of Bayesian inference, this article explores a finite mixture of semiparametric mixed‐effects joint models with skewed distributions for longitudinal measures with an attempt to mediate homogeneous characteristics, adjust departures from normality, and tailor accuracy from measurement error in covariate as well as overcome shortages of confidence in specifying a time‐to‐event model. The Bayesian mixture of joint modeling offers an appropriate avenue to estimate not only all parameters of mixture joint models but also probabilities of class membership. Simulation studies are conducted to assess the performance of the proposed method, and a real example is analyzed to demonstrate the methodology. The results are reported by comparing potential models with various scenarios. Copyright © 2015 John Wiley & Sons, Ltd.     

Causal inference from randomized trials in social epidemiology

Social epidemiology is the study of relations between social factors and health status in populations. Although recent decades have witnessed a rapid development of this research program in scope and sophistication, causal inference has proven to be a persistent dilemma due to the natural assignment of exposure level based on unmeasured attributes of individuals, which may lead to substantial confounding. Some optimism has been expressed about randomized social interventions as a solution to this long-standing inferential problem. We review the causal inference problem in social epidemiology, and the potential for causal inference in randomized social interventions. Using the example of a currently on-going intervention that randomly assigns families to non-poverty housing, we review the limitations to causal inference even under experimental conditions and explain which causal effects become identifiable. We note the benefit of using the randomized trial as a conceptual model, even for design and interpretation of observational studies in social epidemiology.     

General linear contrasts on latent variable means: structural equation hypothesis tests for multivariate clinical trials

Structural equation models articulate the assumed measurement and causal relations among variables, imposing discipline on otherwise unstructured and redundant associations that arise in correlational studies. Biomedical research has eschewed such methods, relying on the generally superior causal inference afforded by randomized controlled trials. Increasingly, however, clinical trials incorporate numerous covariates that are measured but unmanipulated. Most clinical trials now also include multiple correlated endpoints, which can generate ambiguous outcome patterns refractory to simple statistical analysis and interpretation. Modern clinical trials are really multivariate longitudinal studies with at best a component of randomized control; as such, structural equation approaches can add rigour and clarity. The analysis of latent variance (LANOVA) conception combines structural equation and experimental analysis of variance legacies. It allows, for any design that can be decomposed into between-group and within-person models, tests on latent means (that is, the means of the unobserved factors) that are directly analogous to their analysis of variance counterparts. LANOVA variables are either outcomes (which may have a time structure), varying covariates (which may have a time structure), or background covariates (which are static). Allowable causal relations are set-recursive: background covariates can affect all other variables; varying covariates can affect current outcomes and later outcomes and covariates; outcomes can affect only later outcomes. All standard ANOVA and ANCOVA hypotheses can then be tested by proper restrictions among the means and intercepts of these latent variables. Structural equation modelling programs can be used to estimate the models and test the hypotheses. I demonstrate the approach by expressing and testing hypotheses appropriate for two clinical studies that evaluate patient-reported outcomes in populations with solid and haematological malignancies.     

Causation and causal inference in epidemiology

Concepts of cause and causal inference are largely self-taught from early learning experiences. A model of causation that describes causes in terms of sufficient causes and their component causes illuminates important principles such as multi-causality, the dependence of the strength of component causes on the prevalence of complementary component causes, and interaction between component causes. Philosophers agree that causal propositions cannot be proved, and find flaws or practical limitations in all philosophies of causal inference. Hence, the role of logic, belief, and observation in evaluating causal propositions is not settled. Causal inference in epidemiology is better viewed as an exercise in measurement of an effect rather than as a criterion-guided process for deciding whether an effect is present or not.     

Activity spaces in place and health research: Novel exposure measures,data collection tools,and designs

Activity space research provides a framework to consider mobility while linking environments to behaviors in the study of neighborhood effects on health. Increased use of wearable location sensors provides new opportunities to observe and analyze fine-grained spatial and temporal information on individuals' mobility patterns, environmental exposures and behaviors; however, these analysis does not easily translate into causal inference. Additional dimensions underlying behavioral decision-making likely influence or even modify environmental effects on behaviors. This commentary discusses how further progresses in exposure measurement, integration of data collection tools, and development of study designs could support future interventions to optimize how environments shape health profiles and inequities.