Triangulation for causal loop diagrams: constructing biopsychosocial models using group model building, literature review, and causal discovery

Abstract

The complex nature of many health problems necessitates the use of systems thinking tools like causal loop diagrams (CLDs) to visualize the underlying causal network and facilitate computational simulations of potential interventions. However, the construction of CLDs is limited by the constraints and biases of specific sources of evidence. To address this, we propose a triangulation approach that integrates expert and theory-driven group model building, literature review, and data-driven causal discovery. We demonstrate the utility of this triangulation approach using a case example focused on the trajectory of depressive symptoms in response to a stressor in healthy adults. After triangulation with causal discovery, the CLD exhibited (1) greater comprehensiveness, encompassing multiple research fields; (2) a modified feedback structure; and (3) increased transparency regarding the uncertainty of evidence in the model structure. These findings suggest that triangulation can produce higher-quality CLDs, potentially advancing our understanding of complex diseases.