4  Motivation

4.1 Mediation analysis in Epidemiology

Mediation analysis is:

The quantitative study of pathways and mechanisms through which an exposure or intervention impacts an outcome

In clinical and epidemiological research, the primary focus is often on determining whether a specific exposure or intervention has an effect on a disease or health outcome (aetiological epidemiology). Once this effect is established, the next natural question is to explore the “black box”—the underlying mechanisms that explain how the exposure (or intervention) leads to the observed outcome. As we are not only interested in whether an exposure/intervention has an effect, but in how this effect works (1).

Mediation analysis helps us to define an answer to this HOW question, by quantitatively exploring the pathways and mechanisms that explain the causal relationship between the exposure (or intervention or treatment) and the outcome. By doing so, mediation analysis helps to open the black box.

4.2 Motivation for mediation analysis

  1. Explanation and Understanding:

    Unpacking how and why certain factors (e.g., physical activity) impact outcomes (e.g., diabetes risk) through intermediate variables (e.g., weight loss).

    Example: Does body mass index mediate the relationship between diet and type 2 diabetes in mid-life?

  2. Confirmation and Refutation of Theory:

    Testing theories about the mechanisms driving health outcomes and determining whether certain factors mediate the relationships between exposures and outcomes.

    Example: Is the relationship between lower socioeconomic status and cardiovascular disease explained (mediated) through high lipid levels?

  3. Refining Interventions:

    Mediation analysis may help us to improve the effectiveness of interventions by allowing us to target key mediators.

    Example: Let’s consider the evidence from the landmark diabetes prevention trials: The Diabetes Prevention Programme (2), The Diabetes Prevention Study (3), and The Da Qing study (4). These studies randomized high diabetes risk individuals to either a lifestyle intervention (consisting of diet modfication in combination with physical activity) or routine health advice and care. The main finding was a ~58% lower risk of incidence type 2 diabetes in the lifestyle intervention group compared to the routine care group.

We might be interested in further refining the intervention so as to increase the magnitude of the effect. It would be interesting to know if the lifestyle changes reduce the risk of diabetes only because they lower BMI (effect mediated through BMI) or if lifestyle has a direct impact on diabetes risk that does not require any change in BMI. This may help us understand how important BMI targets are in future lifestyle modification programs.

Can you find a relevant mediation question in your own research?

Spend 3 minutes talking to your neighbour and discuss an mediation question that may be relevant in your own research.

4.3 Motivations for running this workshop

The methodological framework for conducting mediation analysis has evolved rapidly in recent years, making it a crucial tool for understanding complex relationships in research. Initially rooted in basic regression techniques, mediation analysis has advanced into causal mediation analysis, enabling researchers to unravel the intricate pathways between exposures, mediators, and outcomes.

This short workshop aims to introduce the basic statistical framework, key assumptions, and practical applications of both traditional and modern mediation analysis techniques. Through real-world data examples, we’ll explore how mediation analysis can enhance your research and understanding of causal relationships.

Despite its growing importance, mediation analysis is often used incorrectly. A recent systematic review on the reporting quality of studies applying mediation analysis revealed that many studies provide incomplete reporting, which hinders both reproducibility and the translation of findings into practice (5) .

Our goal is for you to learn how to apply mediation analysis correctly and leverage it as a powerful tool to advance your research. By improving the quality and rigor of mediation analysis in your work, you can ensure more reliable findings and meaningful insights.

4.4 References

1.
VanderWeele TJ. Mediation Analysis: A Practitioner’s Guide. Annual Review of Public Health [Internet] 2016 [cited 2022 Aug 17];37:17–32. Available from: https://doi.org/10.1146/annurev-publhealth-032315-021402
2.
Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM, Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. The New England Journal of Medicine 2002;346:393–403.
3.
Tuomilehto J, Lindström J, Eriksson JG, Valle TT, Hämäläinen H, Ilanne-Parikka P, Keinänen-Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. The New England Journal of Medicine 2001;344:1343–50.
4.
Pan XR, Li GW, Hu YH, Wang JX, Yang WY, An ZX, Hu ZX, Lin J, Xiao JZ, Cao HB, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 1997;20:537–44.
5.
Rrn R, Ag C, Mk B, Sm G, H L, Jh M. A Systematic Review of the Reporting Quality of Observational Studies That Use Mediation Analyses. Prevention science : the official journal of the Society for Prevention Research [Internet] 2022 [cited 2023 Sep 27];23. Available from: https://pubmed.ncbi.nlm.nih.gov/35167030/