Introducing Causal Inference Using Bayesian Networks and do-Calculus

The Target Trial: A Powerful Device Beyond Well-defined Interventions.

This study investigates a discrete causal method for nominal data (DCMND) which is one of the important issues of causal inference. It is utilized to learn the causal Bayesian network to reflect the interconnections between variables in our paper. This article also proposes a Bayesian network construction algorithm based on discrete causal inference (BDCI) and an extended BDCI Bayesian network construction algorithm based on DCMND. Furthermore, the paper studies the alarm data of mobile communication system in practice. The results suggest that decision criterion based our method is effective in causal inference and the Bayesian network constructed by our method has better classification accuracy compared to other methods.

Despite being a milestone in the history of statistical causal inference, Sewall Wright’s 1918 invention of path analysis did not receive much immediate attention from the statistical and scientific community. Through a careful historical analysis, this paper reveals some previously overlooked philosophical issues concerning the history of causal inference. Placing the invention of path analysis in a broader historical and intellectual context, I portray the scientific community’s initial lack of interest in the method as a natural consequence of relevant scientific and philosophical conditions. In addition to Karl Pearson’s positivist refutation of causation, I contend that the acceptance of path analysis faced several other challenges, including the introduction of a new formalism, conceptual barriers to causal inference, and the lack of model-based statistical thinking. The presence of these challenges shows that the delayed progress in causal inference in the early twentieth century was inevitable.

In 2002, the Agency for Healthcare Research and Quality (AHRQ) defined health services research as ''research [that] examines how people get access to health care, how much care costs, and what happens to patients as a result of this care. The main goals of health services research are to identify the most effective ways to organize, manage, finance, and deliver high quality care; reduce medical errors; and improve patient safety'' (http://www.academy health.org). AcademyHealth's (AH) definition is parallel but focuses more explicitly on the factors studied. Thus AH defines health services research as ''the multidisciplinary field of scientific investigation that studies how social factors, financing systems, organizational structures and processes, health technologies, and personal behaviors affect access to health care, the quality and cost of health care, and ultimately our health and well-being. Its research domains are individuals, families, organizations, institutions, communities, and populations'' (http://www.academyhealth.org). Explicit in both definitions is the notion that health services researchers should strive to identify and estimate the causal effects on outcomes of interest of alternative organizational structures, management approaches, financing systems, provider practices, and personal choices regarding lifestyle and behavior. Without a focus on causal effects, it would be impossible to identify the most effective ways to achieve the outcomes we seek through clinical, management, or policy interventions.

Causal inference comprises the understanding of how a certain condition would change under a specific modification of the steady state of the world. In epidemiology, causal inference attempts to understand the cause of a certain disease at the population level. Statistical inference relates to the distribution of a disease in a given sample and how closely this distribution approximates the population-level distribution. Causal inference requires us to estimate the distribution of a given disease in the population under an intervention: how changing risk factor “X” would change disease distribution “Y.” Throughout history, several theories have been formulated to define and explain causation. Although a consensus on the definition of causation has not been reached, the field has developed tools for epidemiologists in the quest to infer causality, such as the use of directed acyclic graphs and novel analytical approaches. Accordingly, this chapter aims to (1) present an overview of historical theories of causation; (2) discuss the concepts of statistical and causal inference; and (3) present new approaches to infer causality that can be used in oral health epidemiology.

To identify and evaluate the failure issues in the livestock manure-to-biomethane supply chain, this study employs a Bayesian network approach with three inference analysis methods: diagnostic analysis, sensitivity analysis, and maximum causal chain inference. First, the main hazard categories affecting the failure of the supply chain are identified, establishing risk indicators for feedstock collection, pretreatment, anaerobic digestion, purification and upgrading, transportation, and biomethane end-use. Then, the half-interval method and possibility superiority comparison are used to calculate and rank the severity of related accidents, obtaining the severity ranking of secondary indicators as well as the severity ranking of work items and risk items. Finally, Bayesian forward inference is applied to investigate the failure probability of the supply chain, combined with backward inference to identify the risk factors most likely to cause supply chain failures and trace the formation of failure hazards. The Bayesian sensitivity analysis method is ultimately applied to determine the key hazards affecting supply chain failures and the correlations between accident hazards, followed by validation. The results show that the failure probability of the supply chain through causal inference is approximately 54.76%, indicating relatively high failure risk. The three factors with the highest posterior probabilities are mechanical stirring failure C3 (88.11%), corrosion-induced ammonia leakage poisoning D6, and equipment explosion caused by excessive pressure due to overheating during dehumidification heating D9, which are the hazards most likely to cause failures in the supply chain. Improper operations and the toxicity of related chemicals are key hazards leading to supply chain failures, with the correlation between accident hazards presented as a hazard chain by integrating severity and accident probability, and the key risk points in the supply chain are identified.

O presente estudo teve por objetivo investigar a habilidade de crianças em compreender textos argumentativos a partir do estabelecimento de inferências de diferentes tipos: causais, de estado e de previsão. Crianças (8 e 9 anos) foram individualmente solicitadas a ler um texto argumentativo segmentado em cinco partes. Imediatamente, após a leitura de cada parte, eram feitas perguntas inferenciais acerca do texto e perguntas de explicação em que a criança tinha que justificar as respostas dadas. Os dados foram analisados em função do número de respostas corretas e a partir de uma análise qualitativa acerca da natureza das dificuldades apresentadas. Observou-se que as crianças de ambos os grupos etários não tinham dificuldades com as inferências causais e que as crianças de 8 anos tinham dificuldades em relação às inferências de previsão.

The predictiveness and diagnostic utility of thyroid antibodies-thyroid-stimulating hormone receptor antibody, anti-thyroglobulin antibody and anti-thyroid peroxidase antibody-were investigated in autoimmune thyroid disease. The charts of 85 patients with Hashimoto's thyroiditis or Graves' disease were retrospectively analyzed with causal inference techniques and machine learning algorithms to estimate adjusted associations and interactions among antibodies. The most robust estimated disease diagnosis association was with anti-thyroid peroxidase antibody (mean treatment effect = 0.731) which also gave the best predictive accuracy (area under the receiver operating characteristic curve = 0.875) especially in the middle-aged and older population. Predictive accuracy was enhanced by using several antibodies (area under the curve = 0.913) whereas, interaction effects were trivial. Thyroid-stimulating hormone receptor antibody positivity was highly predictive of the clinical decision to start Methimazole treatment in Graves' disease. The findings suggest that anti-thyroid peroxidase antibody preference for screening and thyroid antibody profile use in treatment planning can guide clinical decision-making, within the constraints of causal inference assumption employed.

Although causal Bayes networks are applicable to examining causal inferences about different static objects and about a changing object with different states, previous studies investigated the former, but not the latter. We propose a situation-modulated minimal change account for causal inferences. It predicts that dynamic situations are more likely to elicit minimal revisions on causal networks and adherence to the Markov assumption than static situations. Two experiments were conducted to investigate qualitative causal inferences about causal networks with binary and numerical variables, respectively. It was found that qualitative causal inferences were more likely to adhere to the Markov assumption in dynamic situations than in static situations. This finding supports the situation-modulated minimal change account rather than the other alternative accounts. We conclude that dynamic situations are more likely to elicit minimal revisions on causal networks and adherence to the Markov assumption than static situations. This conclusion is beyond the previous predominant view that causal inferences are apt to violate the Markov assumption.

Decision-analytic models (DAMs) are essentially informative yet complex tools for solving questions in medical decision making. When their complexity grows, the need for causal inference techniques becomes evident as causal relationships between variables become unclear. In this methodological commentary, we argue that graphical representations of assumptions on such relationships, directed acyclic graphs (DAGs), can enhance the transparency of decision models and aid in parameter selection and estimation through visually specifying backdoor paths (i.e., potential biases in parameter estimates) and visually clarifying structural modeling choices of frontdoor paths (i.e., the effect of the model structure on the outcome). This commentary discusses the benefit of integrating DAGs and DAMs in medical decision making and in particular health economics with 2 applications: the first examines statin use for prevention of cardiovascular disease, and the second considers mindfulness-based interventions for students' stress. Despite the potential application of DAGs in the decision science framework, challenges remain, including simplicity, defining the scope of a DAG, unmeasured confounding, noncausal aspects, and limited data availability or quality. Broader adoption of DAGs in decision science requires full-model applications and further debate.HighlightsOur commentary proposes the application of directed acyclic graphs (DAGs) in the design of decision-analytic models, offering researchers a valuable and structured tool to enhance transparency and accuracy by bridging the gap between causal inference and model design in medical decision making.The practical examples in this article showcase the transformative effect DAGs can have on model structure, parameter selection, and the resulting conclusions on effectiveness and cost-effectiveness.This methodological article invites a broader conversation on decision-modeling choices grounded in causal assumptions.

This Chapter discusses the issue of irregular, covariate-dependent measurement times when drawing causal inferences with non-randomized data. We call measurement times the times when a variable of interest is observed. The focus is on a longitudinal outcome, e.g., a clinical measure of interest, that is observed only sporadically. The reasons why irregular, covariate-dependent measurements can bias causal inference of the conditional or the marginal treatment effects on the longitudinal outcome are discussed. Two types of methods to account for irregular measurement times are distinguished, those using inverse weights to create a pseudo-population in which covariates are balanced across measured outcomes and those which are not measured, and methods based on the imputation of the longitudinal outcome process at a set of prespecified time points. Modeling of the measurement times and the causal assumptions that are required when using inverse weights are briefly discussed. We consider the advantages and pitfalls of imputation methods over weighting methods. A toy case study is presented, and R code is provided on the companion website to implement the two main methods discussed in this Chapter to deal with irregular measurement times.

Understanding causal mechanisms among variables is critical to efficient management of complex biological systems such as animal agriculture production. The increasing availability of data from commercial livestock operations offers unique opportunities for attaining causal insight, despite the inherently observational nature of these data. Causal claims based on observational data are substantiated by recent theoretical and methodological developments in the rapidly evolving field of causal inference. Thus, the objectives of this review are as follows: 1) to introduce a unifying conceptual framework for investigating causal effects from observational data in livestock, 2) to illustrate its implementation in the context of the animal sciences, and 3) to discuss opportunities and challenges associated with this framework. Foundational to the proposed conceptual framework are graphical objects known as directed acyclic graphs (DAGs). As mathematical constructs and practical tools, DAGs encode putative structural mechanisms underlying causal models together with their probabilistic implications. The process of DAG elicitation and causal identification is central to any causal claims based on observational data. We further discuss necessary causal assumptions and associated limitations to causal inference. Last, we provide practical recommendations to facilitate implementation of causal inference from observational data in the context of the animal sciences.

Causal discovery and effect estimation for time series provide scientists with a way to extract causal information from observational studies when possible. But the high dimensionality of raw climate data causes computational problems for most analysis methods, and causal inference is no exception. To address this problem, climate scientists usually pre-process climate data using dimension reduction techniques (including seasonal and regional averaging and principle component analysis) that may result in the loss of valuable information before the true analysis even begins. For example, climate scientists often represent El Niño Southern Oscillation variability (ENSO) using the uni-variate Nino3.4 index, which cannot distinguish between central Pacific and eastern Pacific El Niño events, which are believed to impact global climate varaibility in different ways. This study introduces a method for avoiding premature data dimension reduction in causal effect estimation, implemented in tigramite. The method allows the researcher to define multi-variate climate indices, reducing the dimensionality of the causal inference problem via the causal assumptions instead of losing information from the data itself. To investigate the performance of this approach on climate data, we examine the effect of ENSO on the North Atlantic Oscillation (NAO) in simulated data from the Coupled Model Intercomparison Project, phase 6. We choose this as our case study because different types of El Nino are believed to have very different effects on NAO, to the extent that the impact may be completely undetectable in observations when no distinction between the types of ENSO is made. By comparing estimated effects using uni- and multi-variate climate indices, we demonstrate that this method retains valuable information that would be lost in uni-variate analysis, and make recommendations for best practices when using multi-variate climate indices in causal effect estimation.

ObjectivesIt remains unclear whether non-alcoholic fatty liver disease (NAFLD) is a cause or a consequence of metabolic syndrome (MetS). We proposed a simplified Bayesian network (BN) and attempted to confirm...

False Data Injection Attack (FDIA) has become a growing concern in modern cyber-physical power systems. Many learning-based approaches have utilized the statistical correlation patterns between measurements to facilitate the detection and localization of FDIA. However, these correlation patterns are susceptible to the distribution drift of measurement data, which can be induced by changes in system operating points or variations in attack strength, leading to degraded model performance. Causal inference serves as a promising solution to this problem, as it can embed the physical relationship between measurements as causal patterns that are robust against data distribution drifts. However, causal inference is also computationally demanding. To leverage its advantages and address the computational cost issue, this paper proposes a lightweight framework based on causal inference and Graph Attention Networks (GATs) to extract causal patterns between measurements and locate FDIAs. The proposed framework consists of two levels. The lower level uses an X-learner algorithm to estimate the causality strength between measurements and generate Measurement Causality Graphs (MCGs). The upper level then applies a GAT to identify the anomaly patterns in the MCGs. Since the extracted causality patterns are intrinsically related to the measurements, it is easier for the upper level model to identify the attacked nodes than the existing FDIA localization approaches. A physical neighbor masking strategy is implemented to cut down the computational cost of both levels. The performance of the proposed framework is evaluated on the IEEE 39-bus and 118-bus systems. Experimental results show that the causality-based FDIA localization mechanism provides a lightweight solution to interpretable measurement causality extraction and robust FDIA localization.