Abstract
BackgroundDirected acyclic graphs (DAGs) are an increasingly popular approach for identifying confounding variables that require conditioning when estimating causal effects. This review examined the use of DAGs in applied health research to inform recommendations for improving their transparency and utility in future research.MethodsOriginal health research articles published during 1999–2017 mentioning ‘directed acyclic graphs’ (or similar) or citing DAGitty were identified from Scopus, Web of Science, Medline and Embase. Data were extracted on the reporting of: estimands, DAGs and adjustment sets, alongside the characteristics of each article’s largest DAG.ResultsA total of 234 articles were identified that reported using DAGs. A fifth (n = 48, 21%) reported their target estimand(s) and half (n = 115, 48%) reported the adjustment set(s) implied by their DAG(s). Two-thirds of the articles (n = 144, 62%) made at least one DAG available. DAGs varied in size but averaged 12 nodes [interquartile range (IQR): 9–16, range: 3–28] and 29 arcs (IQR: 19–42, range: 3–99). The median saturation (i.e. percentage of total possible arcs) was 46% (IQR: 31–67, range: 12–100). 37% (n = 53) of the DAGs included unobserved variables, 17% (n = 25) included ‘super-nodes’ (i.e. nodes containing more than one variable) and 34% (n = 49) were visually arranged so that the constituent arcs flowed in the same direction (e.g. top-to-bottom).ConclusionThere is substantial variation in the use and reporting of DAGs in applied health research. Although this partly reflects their flexibility, it also highlights some potential areas for improvement. This review hence offers several recommendations to improve the reporting and use of DAGs in future research.
Highlights
Estimating causal effects is a key aim of applied health research.[1]
This review explicitly considered the use of Directed acyclic graphs (DAGs) in applied health research, where they were used not to discuss causal inference theory but to ‘identify variables necessary for adjustment’
We offer several recommendations for improving the reporting, specification and application of DAGs in applied health research where causal effect estimates are sought from observational data
Summary
Estimating causal effects is a key aim of applied health research.[1]. One approach is to conduct a randomized controlled experiment, but practical and ethical constraints mean this is only possible for a limited range of exposures.[2]. Many approaches are available to assist with deciding which variables to adjust for from a list of potential candidates, including various theory-free statistical criteria and algorithms.[5] few of these conventional approaches explicitly consider the role of each variable in relation to the exposure and outcome, and it is often unclear why some variables were chosen for consideration and others not Without this information, many of the reported associations are uninterpretable, since estimating a specific causal effect requires conditioning on a specific set of variables that are determined by strong theoretical principles.[6,7] This is exemplified by the ‘Table 2 fallacy’, which occurs when the coefficients for two or more ‘risk factors’ in a multivariable regression model are (mistakenly) interpreted as estimates for meaningful causal effects.[8].
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
