The use of directed acyclic graphs (DAGs) in physical activity and nutrition research
A scoping review of the literature
DOI:
https://doi.org/10.51224/SRXIV.556Keywords:
causal inference, confounding, collider bias, epidemiology, kinesiologyAbstract
Introduction: Directed acyclic graphs (DAGs) illustrate causal structures, but their application in physical activity and nutrition research remains unclear. We aimed to characterise DAG use in this literature, highlighting best practices and areas for improvement.
Methods: We conducted a scoping review of DAG use in physical activity and nutrition-related articles published between 1999 and 2024, extracting data on study topic, design, DAG justification and construction, number of arcs, nodes, exposures, outcomes, confounders, mediators, mediator-outcome confounders, competing exposures, and instrumental variables.
Results: Of 115 included studies, 110 contained extractable DAG data. Five could not be extracted due to DAG size or unfixed nodes. Among the 110 studies, 86 (78%) made their DAG available. Most (61, 55%) did not specify methods for identifying variables or causal arcs. When specified, the most common approach was literature review (32, 29%). DAGitty software was used in 68 studies (62%). A total of 96 DAGs were identified, with the majority addressing nutritional exposures (75, 68%). DAGs had a median number of 13 nodes; 2 causal paths; 6 confounders; 1 mediator; and 0 mediator-outcome confounders, instrumental variables and competing exposures.
Conclusion: DAGs support causal inference but their value depends on accurately representing the true causal structure. Many studies lacked a systematic approach for DAG construction and omitted potentially informative nodes such as mediators and mediator-outcome confounders. We provide recommendations to improve the use and transparency of DAGs in physical activity and nutrition research.
Metrics
References
Pearl J. Causal Inference in the Health Sciences: A Conceptual Introduction. Health Serv Outcomes Res Methodol. 2001;2(3/4):189–220.
Hernán MA, Hsu J, Healy B. A Second Chance to Get Causal Inference Right: A Classification of Data Science Tasks. CHANCE. 2019 Jan 2;32(1):42–9.
Grimes DA, Schulz KF. Bias and causal associations in observational research. The Lancet. 2002 Jan;359(9302):248–52.
Hernán MA, Hernández-Díaz S, Robins JM. A Structural Approach to Selection Bias. Epidemiology. 2004 Sep;15(5):615–25.
Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999 Jan;10(1):37–48.
Tennant PWG, Murray EJ, Arnold KF, Berrie L, Fox MP, Gadd SC, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int J Epidemiol. 2021 May 17;50(2):620–32.
Knüppel S, Stang A. DAG program: identifying minimal sufficient adjustment sets. Epidemiology. 2010 Jan;21(1):159.
Hernan MA, Cole SR. Invited Commentary: Causal Diagrams and Measurement Bias. Am J Epidemiol. 2009 Oct 15;170(8):959–62.
Peters MM, Godfrey C, McInerney P, Munn Z, Tricco AC, Khalil H. Chapter 11: Scoping reviews. In: JBI Manual for Evidence Synthesis. JBI; 2020.
Clark JM, Sanders S, Carter M, Honeyman D, Cleo G, Auld Y, et al. Improving the translation of search strategies using the Polyglot Search Translator: a randomized controlled trial. Journal of the Medical Library Association. 2020 Apr 1;108(2).
Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med. 2018 Oct 2;169(7):467–73.
Textor J, Hardt J, Knüppel S. DAGitty. Epidemiology. 2011 Sep;22(5):745.
Ferguson KD, McCann M, Katikireddi SV, Thomson H, Green MJ, Smith DJ, et al. Evidence synthesis for constructing directed acyclic graphs (ESC-DAGs): a novel and systematic method for building directed acyclic graphs. Int J Epidemiol. 2020 Feb 1;49(1):322–9.
Tennant PW, Arnold K, Berrie L, Ellison GT, Gilthorpe MS. ADVANCED MODELLING STRATEGIES Challenges and pitfalls in robust causal inference with observational data [Internet]. 2017. Available from: http://lida.leeds.ac.uk/
Choi KW, Chen CY, Stein MB, Klimentidis YC, Wang MJ, Koenen KC, et al. Assessment of Bidirectional Relationships Between Physical Activity and Depression Among Adults. JAMA Psychiatry. 2019 Apr 1;76(4):399.
Pearl J. Causal Diagrams for Empirical Research. Biometrika. 1995 Dec;82(4):669.
Campbell T, Cullen B. Estimating the effect of physical activity on cognitive function within the UK Biobank cohort. Int J Epidemiol. 2023 Oct 5;52(5):1592–611.
Foraita R, Witte J, Börnhorst C, Gwozdz W, Pala V, Lissner L, et al. A longitudinal causal graph analysis investigating modifiable risk factors and obesity in a European cohort of children and adolescents. Sci Rep. 2024 Mar 21;14(1):6822.
Peng Z, Apfelbacher C, Brandstetter S, Eils R, Kabesch M, Lehmann I, et al. Directed acyclic graph for epidemiological studies in childhood food allergy: Construction, user’s guide, and application. Allergy. 2024 Aug 17;79(8):2051–64.
Schwartz MH, Steele KM, Ries AJ, Georgiadis AG, MacWilliams BA. A model for understanding the causes and consequences of walking impairments. PLoS One. 2022 Dec 28;17(12):e0270731.
Horgan BG, Drew MK, Halson SL, Piromalli LE, Drinkwater EJ, Chapman DW, et al. Impaired recovery is associated with increased injury and illness: A retrospective study of 536 female netball athletes. Scand J Med Sci Sports. 2021 Mar 18;31(3):691–701.
Amaral YN di V do, Marano D, Filha MMT, Moreira MEL. Pre-gestational overweight and polyunsaturated fatty acids in human milk: theoretical causality model. Cien Saude Colet. 2024;29(2).
Tang D, Xiao X, Chen L, Kangzhu Y, Deng W, Basang, et al. Association of dietary patterns with obesity and metabolically healthy obesity phenotype in Chinese population: a cross-sectional analysis of China Multi-Ethnic Cohort Study. British Journal of Nutrition. 2022 Dec 14;128(11):2230–40.
Xiao X, Qin Z, Lv X, Dai Y, Ciren Z, Yangla Y, et al. Dietary patterns and cardiometabolic risks in diverse less-developed ethnic minority regions: results from the China Multi-Ethnic Cohort (CMEC) Study. Lancet Reg Health West Pac. 2021 Oct;15:100252.
Preußel K, Milde-Busch A, Schmich P, Wetzstein M, Stark K, Werber D. Risk Factors for Sporadic Non-Pregnancy Associated Listeriosis in Germany—Immunocompromised Patients and Frequently Consumed Ready-To-Eat Products. PLoS One. 2015 Nov 23;10(11):e0142986.
Kwok WS, Khalatbari-Soltani S, Dolja-Gore X, Byles J, Oliveira JS, Pinheiro MB, et al. Differences in Falls and Physical Activity in Older Women From Two Generations. J Gerontol A Biol Sci Med Sci. 2024 Apr 1;79(4).
Campbell T, Ferguson K, Whyte J, Cullen B. Constructing a graphical model of the relationship between physical activity and cognitive function based on a systematic review of prospective evidence. 2021.
Foraita R, Witte J, Börnhorst C, Gwozdz W, Pala V, Lissner L, et al. A longitudinal causal graph analysis investigating modifiable risk factors and obesity in a European cohort of children and adolescents. Sci Rep. 2024 Mar 21;14(1):6822.
Kalisch M, Ch BME. Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm Peter B ¨ uhlmann. Vol. 8, Journal of Machine Learning Research. 2007.
Gururaghavendran R, Murray EJ. Can algorithms Replace Expert Knowledge for Causal Inference? A Case Study on Novice Use of Causal Discovery. Am J Epidemiol. 2024 Aug 31;
Ferguson KD, McCann M, Katikireddi SV, Thomson H, Green MJ, Smith DJ, et al. Evidence synthesis for constructing directed acyclic graphs (ESC-DAGs): a novel and systematic method for building directed acyclic graphs. Int J Epidemiol. 2020 Feb 1;49(1):322–9.
Poppe L, Steen J, Loh WW, Crombez G, De Block F, Jacobs N, et al. How to develop causal directed acyclic graphs for observational health research: a scoping review. Health Psychol Rev. 2024 Sep 27;1–21.
Tomova GD, Arnold KF, Gilthorpe MS, Tennant PWG. Adjustment for energy intake in nutritional research: a causal inference perspective. Am J Clin Nutr [Internet]. 2022 Jan 11;115(1):189–98. Available from: https://academic.oup.com/ajcn/article/115/1/189/6329041
Tomova GD, Gilthorpe MS, Tennant PW. Theory and performance of substitution models for estimating relative causal effects in nutritional epidemiology. Am J Clin Nutr. 2022 Nov;116(5):1379–88.
Ibsen DB, Dahm CC. Food substitutions revisited. Am J Clin Nutr. 2022 Nov;116(5):1195–8.
PEARL J. Causal diagrams for empirical research. Biometrika. 1995;82(4):669–88.
Lee KJ, Carlin JB, Simpson JA, Moreno-Betancur M. Assumptions and analysis planning in studies with missing data in multiple variables: moving beyond the MCAR/MAR/MNAR classification. Int J Epidemiol. 2023 Aug 2;52(4):1268–75.
Bulbulia JA. Methods in causal inference. Part 4: confounding in experiments. Evol Hum Sci. 2024 Sep 27;6:e43.
Evanchuk JL, Kozyrskyj A, Hanas N, Goruk S, Vaghef-Mehrabani E, Archundia-Herrera CM, et al. Maternal Iron Status Is Dynamic Throughout Pregnancy and Might Predict Birth Outcomes in a Sex Dependent Manner: Results from the Alberta Pregnancy Outcomes and Nutrition (APrON) Cohort Study. J Nutr. 2023 Sep;153(9):2585–97.
Nilsson A, Bonander C, Strömberg U, Björk J. A directed acyclic graph for interactions. Int J Epidemiol. 2021 May 17;50(2):613–9.
Downloads
Additional Files
Posted
License
Copyright (c) 2025 Gabriel Perri Esteves, Lucas Porto Santos, Joanna Shim, Jake Slaton, Eimear Dolan, Paul Swinton (Author)
This work is licensed under a Creative Commons Attribution 4.0 International License.
