Capturing Prevention Science Complexities with Systems Science: Modeling and Simulating Program Implementation and Sustainment with System Dynamics

Few would argue that mental health prevention and program implementation/sustainment are complex and challenging. This stems from a multifaceted system of diverse parts (elements, agents, actors); multilevel, embedded ecological factors; and an array of poorly understood or unacknowledged system behaviors (interactions, processes, feedback loops) – all interacting with one another over time. Yet, existing models and methods often fail to capture/address the dynamic complexity of program implementation – controlling or simplifying key variables and processes – inevitably diminishing model scope and utility and further limiting understanding of counterintuitive and paradoxical behaviors of complex systems. As promoted by NIH-OBSSR, engineering designs/techniques may be better suited to address prevention science complexities; systems science may offer remarkable potential. System Dynamics (SD), in particular, offers strategies/tools to capture the systemic nature of program implementation/sustainment and examine “big pictures” with interconnected parts and evolving patterns.

Striking a balance between a conceptual presentation and a mathematical application, we illustrate the utility of SD by modeling and simulating hypothetical data (grounded in major empirical findings) as applied to an existing model of program implementation and sustainment delivered in schools by teachers (i.e., Han & Weiss [HW], 2005). Based on the HW qualitative model (conceptual feature of SD), we outline the SD modeling process by demonstrating/illustrating:

CAUSAL MAPPING & DIAGRAMMING – Map built using 4 stocks & flows (teacher skill, attribution & motivation-each contributing to implementation) by developing component structures to represent system dynamics of implementation process; Used to develop quantitative SD models with numerical processes underlying diagram components by specifying parameters & relationships

STOCK-FLOW CHANGES, LOOPS & BoT – Process of modeling behavior over time/BoT; Feedback loops among system elements intensify & shape even small changes arising from spontaneously emerging dynamics, subsequently generating new patterns over time (often unexpected/unintended/counterintuitive; Examples at 2 phases presented w/SD model)

EXTENSIONS – Additional “what if” scenarios, altering intervention characteristics (type, timing, frequency) & exploring dynamic influences of model determinants/outcomes (child behavior)

Public health services and systems constantly evolve as new challenges present, demanding innovative solutions in dynamically complex settings. While not perfect (nor appropriate for all issues), SD can enable prevention stakeholders across disciplines to make sense of real-world issues in real-time, to explore insights and simulate experiments about complex system phenomena otherwise unknown/unpredictable over time. While our model pertains to school mental health, various settings/circumstances are applicable. As such, we discuss how other frameworks may be integrated to provide a more comprehensive model. We conclude by discussing implications for mental health and prevention, SD strengths/limitations and future directions for research, policy and practice.