Using Multiple Sources of Evidence to Create Better Models of Complex Public Health Problems

We demonstrate a new method for creating causal loop diagrams (CLDs) that uses multiple sources of evidence to produce more comprehensive and reliable models of health problems. We illustrate this approach through a case study examining the factors influencing mood problems in healthy adults.
Using Multiple Sources of Evidence to Create Better Models of Complex Public Health Problems
Like

Share this post

Choose a social network to share with, or copy the URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

Causal loop diagrams are visual representations of complex and dynamic systems that illustrate the (bidirectional) relationships between different variables. These diagrams are especially helpful for understanding health problems, which are often complex and involve many interconnected factors. For example, a CLD could be used to map out the factors contributing to obesity, including individual behaviors like diet and exercise, as well as social and environmental factors like access to healthy food and safe places to be active. Moreover, a CLD shows causal relations between factors, making it possible to intervene more easily.

Traditional methods for constructing CLDs often rely on the knowledge of experts in a particular field. While expert input is valuable, it can be subjective and may not capture all aspects of the problem’s complexity. This paper argues that a more robust approach is needed, one that combines expert knowledge with insights from scientific literature and data analysis.

In particular, we propose to use evidence triangulation, combining three sources of evidence: Group model building (GMB), a literature review, and causal discovery. The first step (GMB) involves bringing together a group of experts to share their knowledge and build a CLD collectively based on consensus. Afterward, a review of existing research is conducted to validate or falsify established relationships between variables. Finally, statistical causal discovery analysis is used to confirm or reject potential causal links between variables using longitudinal empirical data. By triangulating these three sources, researchers can identify areas of agreement and disagreement, leading to a more comprehensive and reliable understanding of the system being studied.

Case Study: Understanding Mood Problems in Healthy Adults

To demonstrate the triangulation method, we have developed a case study on the biopsychosocial factors influencing mood problems in healthy adults. The researchers focused on a dataset created by questionnaire and biological data on 410 participants aged 30-39 from the Healthy Brain Study (Fernández, 2021) , a longitudinal study on brain function, mental health, lifestyle, and social factors.

As outlined in the triangulation design, a multidisciplinary team of experts was assembled, and the problem to be modeled was defined as the trajectory of depressive symptoms in response to stressors. The experts participated in several GMB sessions to map out the potential causal links between variables, considering factors like stress, sleep, social support, inflammation, and health behaviors. After each GMB session, the experts reviewed scientific literature to find evidence supporting or refuting the proposed links in the model. Links with no supporting evidence were either removed or flagged as "hypothetical." Finally, data from the Healthy Brain Study was analyzed using an algorithm J-PCMCI+, which can identify potential causal links between variables measured over time. Moreover, the analysis results and findings of the literature review were used by the experts to refine their initial model. This step allowed the researchers to improve the CLD, identify areas of uncertainty, and highlight potential avenues for future research.

Conclusions

The triangulation process led to a number of improvements in the CLD. By combining different sources of evidence, the researchers were able to identify links between variables that may have been missed by relying on a single method only. The triangulation process led to changes in the feedback loops identified in the model. Feedback loops are crucial for understanding how systems change over time, and identifying these loops accurately is essential for developing effective interventions. By highlighting areas of agreement and disagreement between the different sources of evidence, the triangulation process made the model more transparent regarding its uncertainties. This knowledge can help guide future research by pinpointing areas where more evidence is needed.

We therefore argue that evidence triangulation can significantly improve the quality of causal loop diagrams, leading to a more accurate and nuanced understanding of complex health problems. Ideally, these CLD’s should be open source available for continuous adaption and updates and could be regarded as a continuous iterative process.  It also opens the potential for developing more effective interventions, making it a valuable approach for researchers and policymakers.

Fernández, G. (2021). Protocol of the Healthy Brain Study: An accessible resource for understanding the human brain and how it dynamically and individually operates in its bio-social context. Plos One, 16(12), e0260952. https://doi.org/10.1371/journal.pone.0260952

 

 

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Follow the Topic

Mental Health
Humanities and Social Sciences > Behavioral Sciences and Psychology > Clinical Psychology > Mental Health
Methodology of Data Collection and Processing
Mathematics and Computing > Statistics > Methodology of Data Collection and Processing

Related Collections

With collections, you can get published faster and increase your visibility.

Information Warfare

Strategies and tools to mend our fractured information ecosystem facilitate exploring remedies for the global challenges of information warfare.

Publishing Model: Open Access

Deadline: Jun 30, 2025

Networks and Space

Whether physical or abstract, space shapes the topology of many real complex networks. This Collection explores this synergy between geometry and graph theory.

Publishing Model: Open Access

Deadline: Apr 30, 2025