JMIR Publications

ABSTRACT

Background:

Suicide remains a public health priority worldwide, with over 700,000 deaths annually and ranking as a leading cause of death among young adults. Traditional research methodologies have often fallen short in capturing the multifaceted nature of suicide, focusing on isolated risk factors rather than the complex interplay of individual, social, and environmental influences. Recognizing these limitations, there is a growing recognition of the value of dynamic simulation modeling to inform suicide prevention planning.

Objective:

The purpose of this systematic review is to provide a comprehensive overview of existing dynamic models of population-level suicide and suicide-related behaviors, and to summarize their methodologies, applications, and outcomes.

Methods:

Eight databases were searched, including MEDLINE, Embase, PsycINFO, Scopus, Compendex, ACM Digital Library, IEEE Xplore, and medRxiv from inception to July 2023. We developed a search strategy in consultation with a research librarian. Two reviewers independently conducted the title and abstract and full-text screenings including studies using dynamic modeling methods (e.g., System Dynamics, agent-based modeling) for suicide or suicide-related behaviors at the population level, and excluding studies on microbiology, bioinformatics, pharmacology, non-dynamic modeling methods, and non-primary modelling reports (e.g., editorials, reviews). Reviewers extracted the data using a standardized form and assessed the quality of reporting using the STRESS guidelines. A narrative synthesis was conducted for included studies. The protocol was registered through PROSPERO (CRD42022346617).

Results:

The search identified 1574 articles, with 22 studies meeting the inclusion criteria, including 15 System Dynamics (SD) models, six agent-based models (ABM), and one microsimulation model. The studies primarily targeted populations in Australia and the United States, with some focusing on hypothetical scenarios. The models addressed various interventions ranging from specific clinical and health service interventions, such as mental health service capacity increases, to broader social determinants, including employment programs and reduction in access to means. The studies demonstrated the utility of dynamic models in identifying synergistic effects of combined interventions and understanding the temporal dynamics of intervention impacts.

Conclusions:

Dynamic modeling of suicide and suicide-related behaviors, though still an emerging area, is expanding rapidly, adapting to a range of questions, settings, and contexts. While the quality of reporting was overall adequate, some lacked detailed reporting on model transparency and reproducibility.This review highlights the potential of dynamic modeling as a tool to support decision-making and to further our understanding of the complex dynamics of suicide and its related behaviors.