Direct and Indirect Effects of Sequential Treatments

By Vanessa Didelez, Philip Dawid, S. Geneletti. 2006

164 citations recorded by the source corpus.

In this paper we review the notion of direct and indirect causal effect as introduced by Pearl (2001). We show how it can be formulated without counterfactuals, using regime indicators instead. This allows to consider the natural (in)direct effect as a special case of sequential treatments discussed by Dawid & Didelez (2005) which immediately yields conditions for identifiability as well as a graphical way of checking identifiability.

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Causal inference by using invariant prediction: identification and confidence intervals (2015, methodology): “Our method is tailored for (but not restricted to) the setting where we have data from different experimental settings or regimes [Didelez et al., 2006].”
Extended Conditional Independence and Applications in Causal Inference (2015, background): “This justifies the hitherto informal or implicit application of these axioms in a number of previous works [10, 13, 19, 20, 23, 24, 25, 27, 28, 30, 36].”
Gformula: Estimating Causal Effects in the Presence of Time-Varying Confounding or Mediation using the G-Computation Formula (2011, unknown): “There has been some controversy over whether or not natural direct and indirect effects are well-defined (Robins, 2003; Didelez, 2006; Hafeman, 2009).”
Alternative Graphical Causal Models and the Identification of Direct E!ects (2010, background): “Didelez et al. (2006) (DDG) discuss an effect that they refer to as the ‘natural direct effect’ and prove it is identified under the agnostic causal model associated with the DAG in Figure 1, the difference between Eq.”
Causal mediation analysis for stochastic interventions (2019, methodology): “…of total effects of stochastic interventions was first considered by Stock (1989) and has been the subject of recent study (Robins et al., 2004; Didelez et al., 2006; Tian, 2008; Pearl, 2009; Taubman et al., 2009; Stitelman et al., 2010; Dı́az and van der Laan, 2013; Dudı́k et al., 2014;…”
Causal Bounds and Observable Constraints for Non-deterministic Models (2012, background): “Consider the controlled direct effect (CDE) (Didelez et al., 2006) and the random regime direct effect (RRDE) CDE(B) = E(C |FB = B,FA = 2)−E(C |FB = B,FA = 1) = EU{E(C |B,A = 2,U)−E(C |B,A = 1,U)} = EU(ηB2 −ηB1), RRDE(a∗) = E(C |FB = da∗ ,FA = 2)−E(C |FB = da∗ ,FA = 1) = EU [EB{E(C |B,A = 2,U)−E(C |”
Joint Causal Inference from Multiple Contexts (2016, unknown): “, 2000), “force variables” (Pearl, 1993), “decision variables” in influence diagrams (Dawid, 2002), “regime indicators” (Didelez et al., 2006), “selection variables” in selection diagrams (Bareinboim and Pearl, 2013), and “environment variable” (Peters et al.”
Identifying direct and indirect effects in a non‐counterfactual framework (2007, background): “More recently, the direct and indirect effects framework has been extended in Didelez et al. (2006).”
Variable Importance and Prediction Methods for Longitudinal Problems with Missing Variables (2015, unknown): “This type of intervention has been previously discussed in the literature (Dı́az and van der Laan, 2011a), and belongs to a wider class of interventions known as stochastic interventions (Korb et al., 2004; Didelez et al., 2006; Dawid and Didelez, 2010).”
Statistical Causality from a Decision-Theoretic Perspective (2014, background): “For further details and developments of the material in this paper, see Dawid (2000); Dawid (2002); Dawid (2003); Didelez et al. (2006); Dawid (2007a); Dawid and Didelez (2008) Dawid (2010a); Dawid (2010b); Dawid and Didelez (2010); Guo and Dawid (2010); Dawid (2011); Geneletti and Dawid (2011);…”
Controlled direct and mediated effects: definition, identification and bounds (2010, background): “A number of authors have considered different identification strategies for direct and indirect effects (Robins & Greenland, 1992; Pearl, 2001; Robins, 2003; Didelez et al., 2006; Petersen et al., 2006; Geneletti, 2007; Hafeman & VanderWeele, 2010; Imai et al., 2010).”
Causal Inference Under Multiple Versions of Treatment (2013, unknown): “The effects are thus much closer to those defined in the direct effects literature by stochastic interventions on the mediator [11, 12] and Propositions 5 and 6 would in fact follow from those results appropriately replacing the mediator by version.”
Counterfactuals, Hypotheticals and Potential Responses: A Philosophical Examination of Statistical Causality (2006, background): “• Definition of direct and indirect effects (Geneletti 2005; Didelez et al. 2006)”
Causality, mediation and time: a dynamic viewpoint (2012, background): “Didelez et al. (2006) state about the natural direct effect (here X is the treatment and Z is the mediator): ‘... we need to think of a way of intervening in X , changing it around, while Z can still be generated as if X was kept at its baseline value.”
Estimation of controlled direct effects (2008, unknown): “Standardized direct effects (Didelez et al., 2006) are obtained by fixing the mediator for each subject at a random draw KÅ from a chosen distribution function for the mediator, which does not depend on the exposure, i.”