Causality, Mediation and Time: A Dynamic Viewpoint
Journal of the Royal Statistical Society Series A (Statistics in Society)Published 7 March 2012Open access
Odd O. Aalen, Kjetil Røysland, Jon Michael Gran, Bruno Ledergerber
Citations10
SJR quartileQ1
SJR score0.92
SNIP1.20
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
Time dynamics are often ignored in causal modelling and clearly, causality must operate in time and it is shown how this corresponds to a mechanistic, or system, understanding of causality.
Abstract
Time dynamics are often ignored in causal modelling. Clearly, causality must operate in time and we show how this corresponds to a mechanistic, or system, understanding of causality. The established counterfactual definitions of direct and indirect effects depend on an ability to manipulate the mediator which may not hold in practice, and we argue that a mechanistic view may be better. Graphical representations based on local independence graphs and dynamic path analysis are used to facilitate communication as well as providing an overview of the dynamic relations 'at a glance'. The relationship between causality as understood in a mechanistic and in an interventionist sense is discussed. An example using data from the Swiss HIV Cohort Study is presented.
Keywords
Mathematics
Journal of Personality and Social PsychologyThe moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations.
72,132 Citations1986Reuben M. Baron, David A. Kenny
EconometricaInvestigating Causal Relations by Econometric Models and Cross-spectral Methods
22,877 Citations1969Clive W. J. Granger
Journal of Educational PsychologyEstimating causal effects of treatments in randomized and nonrandomized studies.
9,471 Citations1974Donald B. Rubin
EpidemiologyMarginal Structural Models and Causal Inference in Epidemiology
5,548 Citations2000James M. Robins, Miguel A. Hernán +1 more
Journal of the American Statistical AssociationStatistics and Causal Inference
5,110 Citations1986Paul W. Holland
Cambridge University Press eBooksInvestigating Causal Relations by Econometric Models and Cross-Spectral Methods
4,797 Citations2010Clive W. J. Granger
Journal of the Royal Statistical Society Series B (Statistical Methodology)Local Computations with Probabilities on Graphical Structures and Their Application to Expert Systems
3,970 Citations1988Steffen L. Lauritzen, David J. Spiegelhalter
This work exploits a range of local representations for the joint probability distribution, combined with topological changes to the original network termed 'marrying' and 'filling-in', which allows efficient algorithms for transfer between representations, providing rapid absorption and propagation of evidence.
Journal of Abnormal PsychologyTesting Mediational Models With Longitudinal Data: Questions and Tips in the Use of Structural Equation Modeling.
3,453 Citations2003David A. Cole, Scott E. Maxwell
Some of the questions that arise (and misconceptions that sometimes emerge) in longitudinal tests of mediational models are described and a collection of tips for structural equation modeling (SEM) of mediations are provided.
Philosophy of ScienceThinking about Mechanisms
3,134 Citations2000Peter Machamer, Lindley Darden +1 more
Thinking in terms of mechanisms provides a new framework for addressing many traditional philosophical issues: causality, laws, explanation, reduction, and scientific change.
Nature GeneticsNetwork motifs in the transcriptional regulation network of Escherichia coli
3,058 Citations2002Shai S. Shen-Orr, Ron Milo +2 more
This work applied new algorithms for systematically detecting network motifs to one of the best-characterized regulation networks, that of direct transcriptional interactions in Escherichia coli, and finds that much of the network is composed of repeated appearances of three highly significant motifs.
Mathematical ModellingA new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect
2,498 Citations1986James M. Robins
A graphical approach to the identification and computation of causal parameters in mortality studies with sustained exposure periods is offered and an adverse effect of arsenic exposure on all-cause and lung cancer mortality which standard methods fail to detect is found.
The Annals of Mathematical StatisticsThe Method of Path Coefficients
2,297 Citations1934Sewall Wright
NatureA satellite view of aerosols in the climate system
2,292 Citations2002Yoram J. Kaufman, D. Tanré +1 more
Increases in aerosol concentration and changes in their composition, driven by industrialization and an expanding population, may adversely affect the Earth's climate and water supply.
Medical Entomology and ZoologyLatent variable models: An introduction to factor, path, and structural analysis
2,131 Citations1987C John
Princeton University Press eBooksScientific Explanation and the Causal Structure of the World
2,010 Citations2020Wesley C. Salmon
EpidemiologyIdentifiability and Exchangeability for Direct and Indirect Effects
1,818 Citations1992James M. Robins, Sander Greenland
It is shown that adjustment for the intermediate variable, which is the most common method of estimating direct effects, can be biased, and that, even in a randomized crossover trial of exposure, direct and indirect effects cannot be separated without special assumptions.
EpidemiologyMarginal Structural Models to Estimate the Causal Effect of Zidovudine on the Survival of HIV-Positive Men
1,727 Citations2000Miguel A. Hernán, Babette Brumback +1 more
The marginal structural Cox proportional hazards model is described and used to estimate the causal effect of zidovudine on the survival of human immunodeficiency virus-positive men participating in the Multicenter AIDS Cohort Study.
American Journal of EpidemiologyCausal Knowledge as a Prerequisite for Confounding Evaluation: An Application to Birth Defects Epidemiology
1,473 Citations2002Miguel A. Hernán
Findings are presented from the Slone Epidemiology Unit Birth Defects Study, 1992-1997, a case-control study on folic acid supplementation and risk of neural tube defects, which suggests that the crude odds ratio should be used because the adjusted odds ratio is invalid.
New England Journal of MedicineHomocysteine Lowering and Cardiovascular Events after Acute Myocardial Infarction
1,436 Citations2006Kaare Harald Bønaa, Inger Njølstad +8 more
Treatment with B vitamins did not lower the risk of recurrent cardiovascular disease after acute myocardial infarction and a harmful effect from combined B vitamin treatment was suggested.
EpidemiologyInstruments for Causal Inference
943 Citations2006Miguel A. Hernán, James M. Robins
The definition of an instrumental variable is reviewed, the conditions required to obtain consistent estimates of causal effects are described, and their implications are explored in the context of a recent application of the instrumental variables approach.
The LancetClinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study
933 Citations1999Bruno Ledergerber, Matthias Egger +10 more
The rate of virological failure of HAART was high among patients with a history of antiretroviral treatment, but the probability of clinical progression was low even in patients with viral rebound.
ACM eBooksDirect and Indirect Effects
920 Citations2022Judea Pearl
A new way of defining the effect transmitted through a restricted set of paths, without controlling variables on the remaining paths is presented, which permits the assessment of a more natural type of direct and indirect effects.
Statistics in the health sciencesSurvival and Event History Analysis
767 Citations2008Odd O. Aalen, Ørnulf Borgan +1 more
Survival and Event History Analysis: A Process Point of View
754 Citations2008Odd O. Aalen, Oernulf Borgan +1 more
Nonparametric analysis of survival and event history data and dynamic models for recurrent events and clustered survival data are presented.
The International Journal of BiostatisticsAn Introduction to Causal Inference
740 Citations2010Judea Pearl
The paper surveys the development of mathematical tools for inferring answers to three types of causal queries and defines the formal and conceptual relationships between the structural and potential-outcome frameworks and presents tools for a symbiotic analysis that uses the strong features of both.
Communication in Statistics- Theory and MethodsCorrecting for non-compliance in randomized trials using structural nested mean models
570 Citations1994James M. Robins
The LancetLong-term effectiveness of potent antiretroviral therapy in preventing AIDS and death: a prospective cohort study
558 Citations2005Jonathan A C Sterne, Miguel A. Hernán +7 more
The results, which are appropriately controlled for confounding by indication, are consistent with reported declines in rates of AIDS and death in developed countries, and provide a context in which to consider adverse effects of HAART.
International Journal of EpidemiologyFallibility in estimating direct effects
542 Citations2002Stephen R. Cole, Miguel A. Hernán
This work uses causal graphs and a partly hypothetical example from the Physicians' Health Study to explain why a common standard method for quantifying direct effects (i.e. stratifying on the intermediate variable) may be flawed.
Journal of Epidemiology & Community HealthA definition of causal effect for epidemiological research
475 Citations2004Miguel A. Hernán
A formal definition of causal effect for epidemiological studies is reviewed and it is shown why, in theory, randomisation allows the estimation of causal effects without further assumptions.
PLoS BiologyCausal Modelling and Brain Connectivity in Functional Magnetic Resonance Imaging
450 Citations2009Karl Friston
Two different state-of-the-art approaches to modeling brain connectivity using neuroimaging are examined.
EpidemiologyEstimation of Direct Causal Effects
360 Citations2006Maya L. Petersen, Sandra E. Sinisi +1 more
An estimation approach for natural direct effects that can be implemented using standard statistical software is presented, and the assumptions underlying the approach are reviewed (which are less restrictive than those proposed by previous authors).
Statistical Methods in Medical ResearchMediation and moderation of treatment effects in randomised controlled trials of complex interventions
355 Citations2009Richard Emsley, Graham Dunn +1 more
London School of Economics and Political Science Research Online (London School of Economics and Political Science)Hunting Causes and Using Them: Approaches in Philosophy and Economics
345 Citations2007Nancy Cartwright
Journal of the American Statistical AssociationCausal Inference without Counterfactuals
337 Citations2000A. P. Dawid
Scandinavian Journal of StatisticsDirect and Indirect Causal Effects via Potential Outcomes*
326 Citations2004Donald B. Rubin
It is suggested that the use of principal stratification is key to understanding the meaning of direct and indirect causal effects and to clarify the situation by using potential outcomes to define causal effects.
New England Journal of MedicineHomocysteine Trials — Clear Outcomes for Complex Reasons
297 Citations2006Joseph Loscalzo
This hypothesis was later modified to include a broader population, positing that mild hyperhomocysteinemia caused by dietary deficiencies of the vitamin cofactors required for the metabolism of homocysteine is a risk factor for atherothrombosis.
EpidemiologyDirect and Indirect Effects in a Survival Context
276 Citations2011Theis Lange, Jørgen Vinsløv Hansen
A simple measure of mediation in a survival setting based on counterfactuals is proposed, and allows a causal interpretation of the mediated effect and is mathematically consistent.
Journal of the Royal Statistical Society Series B (Statistical Methodology)Chain Graph Models and their Causal Interpretations
276 Citations2002Steffen L. Lauritzen, Thomas S. Richardson
Journal of the Royal Statistical Society Series A (Statistics in Society)Causality: Some Statistical Aspects
269 Citations1992D. R. Cox
The definition is that causality is a statistical association that cannot be explained away by confounding variables and the other is based on a link with notions in the design of experiments.
Clinical TrialsThe intermediate endpoint effect in logistic and probit regression
260 Citations2007David P. MacKinnon, Chondra M. Lockwood +3 more
A limitation of a widely used approach to assessing intermediate endpoint effects is described and an alternative method, based on products of coefficients, that yields more accurate results is proposed.
Journal of EconometricsGranger causality and path diagrams for multivariate time series
232 Citations2006Michael Eichler
This paper introduces path diagrams for multivariate time series which visualize the dynamic relationships among the variables and shows that these path diagrams provide a framework for the analysis of the dependence structure of the time series.
International Journal of EpidemiologyOpening the Black Box: a motivation for the assessment of mediation
223 Citations2009Danella Hafeman, Sharon Schwartz
This work integrates literature from epidemiology and other social sciences to describe the reasons that an investigator might want to assess mediation and the connections between these reasons and specific measures of indirect and direct effects that have been previously described.
Causation and Explanation
213 Citations2014Stathis Psillos
Causation and Explanation
190 Citations2002Stathis Psillos
The Annals of StatisticsCausal Inference for Complex Longitudinal Data: The Continuous Case
177 Citations2001Richard D. Gill, James M. Robins
Robins’ theory of causal inference for complex longitudinal data is extended to the case of continuously varying as opposed to discrete covariates and treatments and it is shown that assumptions concerning counterfactual variables place no restriction on the joint distribution of the observed variables.
BiometricsCausal Mediation Analyses with Rank Preserving Models
158 Citations2007Thomas R. Ten Have, Marshall M. Joffe +4 more
Simulations show good estimation and confidence interval performance by the proposed RPM approach under unmeasured confounding relative to the standard mediation approach, but poor performance under departures from the structural interaction assumptions.
European Journal of EpidemiologyMediation and mechanism
140 Citations2009Tyler J. VanderWeele
It is shown that both concepts can be illustrated using a single causal diagram and it is also shown that mediation implies mechanism but mechanism need not imply mediation.
Osteoporosis InternationalBack pain in osteoporotic vertebral fractures
134 Citations2007Roger M. Francis, Terry Aspray +3 more
The management of pain in patients with vertebral fractures and the direct and indirect effect of osteoporosis treatments on back pain are reviewed, with emerging evidence suggesting that calcitonin, intravenous bisphosphonates and teriparatide may also have a direct effect on bone pain.
BiometrikaEstimation of treatment effects in randomised trials with non-compliance and a dichotomous outcome using structural mean models
115 Citations2004James M. Robins
Journal of the Royal Statistical Society Series B (Statistical Methodology)Estimation of Controlled Direct Effects
106 Citations2008Sylvie Goetgeluk, Stijn Vansteelandt +1 more
Lifetime Data AnalysisOn Granger causality and the effect of interventions in time series
98 Citations2009Michael Eichler, Vanessa Didelez
This paper considers the effect of external, possibly multiple and sequential, interventions in a system of multivariate time series, the Granger causal structure of which is taken to be known.
Genome ResearchEstimating Coarse Gene Network Structure from Large-Scale Gene Perturbation Data
98 Citations2002Andreas Wagner
Bristol Research (University of Bristol)Direct and Indirect Effects of Sequential Treatments
96 Citations2012Vanessa Didelez, A. P. Dawid +1 more
International Journal of CancerEducation and risk of breast cancer in the Norwegian‐Swedish women's lifestyle and health cohort study
89 Citations2004Tonje Braaten, Elisabete Weiderpass +3 more
The results of this study suggest a clear positive gradient in risk for breast cancer by level of education, which can be fully explained by established breast cancer risk factors.
FEBS LettersThe dynamic systems approach to control and regulation of intracellular networks
89 Citations2005Olaf Wolkenhauer, Mukhtar Ullah +2 more
The case for differential equation models as a ‘natural’ representation of causal entailment in pathways is reviewed and block‐diagrams, commonly used in the engineering sciences, are introduced and compared to pathway maps.
Statistics in MedicineA sequential Cox approach for estimating the causal effect of treatment in the presence of time‐dependent confounding applied to data from the Swiss HIV Cohort Study
88 Citations2010Jon Michael Gran, Kjetil Røysland +7 more
A simple and intuitive approach to estimate similar treatment effects, using observational data to mimic several randomized controlled trials, and offers an alternative to the use of inverse probability of treatment weights, which is unstable in certain situations.
Evaluation ReviewGraphical Models for Causation, and the Identification Problem
83 Citations2004David A. Freedman
Journal of the Royal Statistical Society Series B (Statistical Methodology)Graphical Models for Marked Point Processes Based on Local Independence
81 Citations2008Vanessa Didelez
A new notion of graph separation, which is called δ‐separation, is introduced and implications for the underlying model as well as for likelihood inference are explored.
Journal of the Royal Statistical Society Series C (Applied Statistics)Analysis of Longitudinal Data with Drop-Out: Objectives, Assumptions and a Proposal
81 Citations2007Peter J. Diggle, Daniel Farewell +1 more
It is argued that more attention could be given to study objectives and to the relevant targets for inference in the analysis of data from a trial into the treatment of schizophrenia, in which approximately 50% of recruited subjects dropped out before the final scheduled measurement time.
Current Opinion in HIV and AIDSRepositioning HIV protease inhibitors as cancer therapeutics
74 Citations2008Wendy B. Bernstein, Phillip A. Dennis
Although mechanisms of action for the antitumor effects of HIV protease inhibitor are complex, their broad spectrum of activity, minimal toxicity, and wide availability make protease inhibitors ideal candidates for repositioning as cancer therapeutics.
ScholarlyCommons (University of Pennsylvania)2007): “Defining and estimating intervention effects for groups that will develop an auxiliary outcome
72 Citations2007Marshall M. Joffe, Dylan Small +1 more
BloodHIV protease inhibitors restore impaired T-cell proliferative response in vivo and in vitro: a viral-suppression–independent mechanism
69 Citations2000Wei Lu, Jean‐Marie Andrieu
Findings explain the discrepancies between virologic and immunologic responses that are increasingly reported in patients receiving HAART, and may provide insights into the pathogenesis of HIV infection.
Statistical ScienceDefining and Estimating Intervention Effects for Groups that will Develop an Auxiliary Outcome
66 Citations2007Marshall M. Joffe, Dylan Small +1 more
Journal of Theoretical BiologyAn abstract cell model that describes the self-organization of cell function in living systems
64 Citations2007Olaf Wolkenhauer, Jan‐Hendrik S. Hofmeyr
The proof of the theorem supports Rosen's argument that living cells have non-simulable properties, as it shows that for a mathematical model of a self-organizing cell the associated category must be cartesian closed.
Lifetime Data AnalysisDynamic path analysis—a new approach to analyzing time-dependent covariates
60 Citations2006Johan Fosen, Egil Ferkingstad +2 more
This article introduces a general approach to dynamic path analysis to the situation where variables may be time-dependent and where the outcome of main interest is a stochastic process, and focuses on the survival and event history analysis setting where the main outcome is a counting process.
Tellus A Dynamic Meteorology and OceanographyGranger causality and Atlantic hurricanes
57 Citations2007James B. Elsner
Journal of the Royal Statistical Society Series B (Statistical Methodology)A General Dynamical Statistical model with Causal Interpretation
53 Citations2009Daniel Commenges, Anne Gégout‐Petit
A general dynamical model as a framework for causal interpretation is developed and a definition of causal influence using the concepts of a ‘physical system’ is proposed, which makes it possible to link descriptive and explicative statistical models, and encompasses quantitative processes and events.
Journal of the Royal Statistical Society Series B (Statistical Methodology)Estimation of Direct Effects for Survival Data by using the Aalen Additive Hazards Model
52 Citations2011Torben Martinussen, Stijn Vansteelandt +2 more
Health Services and Outcomes Research MethodologyCausal mediation analyses for randomized trials
51 Citations2008Kevin G. Lynch, Mark Cary +2 more
Econometric TheoryNoncausality in Continuous Time Models
51 Citations1996Fabienne Comte, Éric Renault
Scandinavian Journal of StatisticsGraphical Models for Composable Finite Markov Processes
51 Citations2006Vanessa Didelez
It is shown that the graph allows to read off further independencies characterizing the underlying Markov process, and some standard methods for inference are adapted to exploit the graphical representation, e.g. for testing local independence.
Cognitive ProcessingA new method for detecting causality in fMRI data of cognitive processing
49 Citations2005Alessandro Londei, Alessandro D’Ausilio +2 more
This paper presents a completely data-driven analysis method that applies both independent components analysis and the Granger causality test, performed in two separate steps, to extract both spatial and temporal activities from fMRI recordings.
Scandinavian Journal of StatisticsWhat can Statistics Contribute to a Causal Understanding?
49 Citations2007Odd O. Aalen, Arnoldo Frigessi
The LancetTime and chance: the stochastic nature of disease causation
46 Citations2005DIW Coggon, CN Martyn
The nature of disease causation is reviewed, and it is argued that the role of stochastic processes is underappreciated.
Biometrical JournalDynamic Analysis of Recurrent Event Data Using the Additive Hazard Model
39 Citations2006Johan Fosen, Ørnulf Borgan +2 more
A method for analysis of recurrent event data using information on previous occurrences of the event as a time-dependent covariate is proposed, and concepts like direct, indirect and total effects of the fixed covariates may be defined in an analogous way as for traditional path analysis.
Statistics in MedicineEvent history analysis and inference from observational epidemiology
34 Citations1999Niels Keiding
Two sets of tools are concerned: event history analysis and structural nested failure time models, both applied to a particularly intricate problem in observational epidemiology, of empirically assessing the graft-versus-leukaemia effect after bone marrow transplantation.
BernoulliA martingale approach to continuous-time marginal structural models
34 Citations2011Kjetil Røysland
Journal of the Franklin InstituteRevealing the dynamic causal interdependence between neural and muscular signals in Parkinsonian tremor
33 Citations2006Shuncai Wang, Yanqing Chen +5 more
Lifetime Data AnalysisDynamic path analysis for event time data: large sample properties and inference
25 Citations2009Torben Martinussen
Large sample properties of the estimator of the cumulative indirect effect that may be used to draw inferences of the total effect into a sum of direct and indirect effects holds under certain assumptions are given.
BiostatisticsA dynamic approach for reconstructing missing longitudinal data using the linear increments model
19 Citations2010Odd O. Aalen, Nina Gunnes
It is proposed to reconstruct responses for missing data and discuss the required assumptions needed for both monotone and nonmonotone missingness, which can be used to estimate causal effects in the presence of time-dependent confounding.
Wiley series in probability and statisticsCausal Inference from Observational Data: A Bayesian Predictive Approach
19 Citations2012Elja Arjas
Bristol Research (University of Bristol)Asymmetric separation for local independence graphs
8 Citations2012Vanessa Didelez
The properties of this graph separation as well as local independence are investigated in detail within a framework of asymmetric (semi)graphoids allowing insight into what information can be read off these graphs.
Lancaster EPrints (Lancaster University)Analysis of longitudinal data with drop-out: objectives, assumptions and a proposal (with discussion).
5 Citations2007Diggle, Peter J., Farewell, Daniel +1 more
PhilSci-Archive (University of Pittsburgh)Varieties of Modularity for Causal and Constitutive Explanations
2 Citations2008Jaakko Kuorikoski
…