Abstract
Graphical models are useful tools in causal inference, and causal directed acyclic graphs (DAGs) are used extensively to determine the variables for which it is sufficient to control for confounding to estimate causal effects. We discuss the following ten pitfalls and tips that are easily overlooked when using DAGs: 1) Each node on DAGs corresponds to a random variable and not its realized values; 2) The presence or absence of arrows in DAGs corresponds to the presence or absence of individual causal effect in the population; 3) “Non-manipulable” variables and their arrows should be drawn with care; 4) It is preferable to draw DAGs for the total population, rather than for the exposed or unexposed groups; 5) DAGs are primarily useful to examine the presence of confounding in distribution in the notion of confounding in expectation; 6) Although DAGs provide qualitative differences of causal structures, they cannot describe details of how to adjust for confounding; 7) DAGs can be used to illustrate the consequences of matching and the appropriate handling of matched variables in cohort and case-control studies; 8) When explicitly accounting for temporal order in DAGs, it is necessary to use separate nodes for each timing; 9) In certain cases, DAGs with signed edges can be used in drawing conclusions about the direction of bias; and 10) DAGs can be (and should be) used to describe not only confounding bias but also other forms of bias. We also discuss recent developments of graphical models and their future directions.
Highlights
Causal diagrams have been often used among epidemiologists as a tool to describe what is already known about relevant causal structures
The use of directed acyclic graphs (DAGs) is widespread among epidemiologists; when consulting A Dictionary of Epidemiology, there was no entry for DAGs in its fourth edition that was published in 2001,9 a definition of DAGs has been included in its later editions.[10,11]
Even if one takes care to not adjust for variables affected by exposure or outcome in the traditional confounder-selection criteria, one may be led to adjust for a “collider” on the backdoor path from exposure to outcome and unnecessarily introduce bias
Summary
Causal diagrams have been often used among epidemiologists as a tool to describe what is already known about relevant causal structures. A confounder was traditionally identified based on the following three criteria[3,4,5]: a) it must be associated with the exposure; b) it must be associated with the outcome in the unexposed; and c) it must not lie on a causal pathway between exposure and outcome Because these traditional criteria sometimes fail, the graphical criteria for identifying confounders in DAGs are especially useful. This point is often explained using an example of the so-called M bias.[12] In this case, even if one takes care to not adjust for variables affected by exposure or outcome in the traditional confounder-selection criteria, one may be led to adjust for a “collider” on the backdoor path from exposure to outcome and unnecessarily introduce bias. We discuss recent developments of graphical models and their future directions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
