Network Analysis as a Tool to Understand Comorbidity Over Time
A. C. Zegarra, D. E. Prieto-Molinari, B. Canessa-Lohmann, A. L. Delgado-TenorioIntroduction
Network analysis provides a methodological alternative to traditional latent variable approaches by conceptualizing psychiatric disorders as systems of interacting symptoms. While cross-sectional network studies map comorbidity at a single point in time, repeated measures are needed to understand how these structures evolve. Longitudinal and repeated cross-sectional designs offer distinct strategies: the former tracks intra-individual symptom dynamics, while the latter captures population-level changes across independent samples. Methodological guidance on how to apply network analysis to comorbidity over time remains scarce, despite increasing demand to study the impact of contextual factors such as pandemics or socioeconomic stressors.
Objectives
Demonstrate methodological strategies for applying network analysis to the study of comorbidities across time, and to illustrate their application with a repeated cross-sectional analysis of depressive symptoms and alcohol use disorder.
Methods
We outline a stepwise framework for temporal network studies of comorbidity. Key stages include: (1) design selection (cross-sectional, repeated cross-sectional, longitudinal), (2) consistent measurement of symptom sets across time, (3) network estimation using regularized models, (4) evaluation of central and bridge symptoms with indices tailored to cross-disorder connectivity, (5) assessment of accuracy and stability via bootstrapping, and (6) network comparison across time points with formal tests of structural invariance. As an applied example, we employed a repeated cross-sectional study using Peruvian survey data on depressive symptoms and alcohol use disorder between 2018 and 2023, with more than 30,000 cases in each year.
Results
he framework enabled detection of stable within-disorder symptom clusters and identification of bridge nodes linking depression and alcohol use. In Peru, worthlessness and guilt consistently emerged as central depressive connectors, while withdrawal acted as a key alcohol-related bridge. Cross-disorder associations intensified in 2020 during the COVID-19 pandemic, illustrating how external stressors can reshape comorbidity structures. Formal comparison confirmed overall invariance across years, highlighting both the sensitivity of the approach and the boundaries of repeated cross-sectional designs.
Image 1: Long description.