Building causal models in pain research: the case of executive functioning and transitions in pain states

Abstract

Pain states fluctuate over time and across situations. Similarly, there is variation in risk and protective factors and how they impact on these pain-related transitions. We are interested in whether such variations are more than random, and whether they can be accounted for by observed variables. The availability of large longitudinal datasets, such as UK Biobank ( https://www.ukbiobank.ac.uk/ ), offers a unique opportunity to study these variations at scale. However, such datasets bring a high risk of bias (eg, confounding) and danger of over-interpretation. It is therefore important to be transparent about our causal thinking. Directed acyclic graphs (DAGs) are graphical representations of the hypothesized causal relationships between variables. They are used to identify the smallest set of variables that need to be adjusted to remove confounding bias in estimating the causal effect of an exposure on an outcome. However, use of DAGs in pain research is not common, despite their potential to guide study design and data analysis. In this article, we present a workflow for building a DAG using domain knowledge from 3 different sources: researchers (theory-based), people with lived experience (person-based), and the literature (evidence-based). We created a DAG for the putative effect of executive function on the maintenance of chronic high-impact pain. The resulting DAG provides a valuable framework for guiding future research on the role of executive functioning in pain, and it underscores the broader potential of using DAGs to improve causal inference in pain research.

More information

Identifiers

ORCID for John McBeth: orcid.org/0000-0001-7047-2183

Catalogue record

Export record

Altmetrics