Public Health Computer Simulation Tool to Support Disaster Preparedness for Flooding in Rural Communities

Background: Natural hazards have complex effects on human systems. When these types of events occur in rural communities, the resources of local public health, healthcare, and emergency response organizations can be quickly overwhelmed. Effective planning that considers the unique characteristics and needs of such under-resourced rural areas can help to mitigate their impact. Model-based systems engineering (MBSE) methods can assist decision makers by considering disasters from a systems perspective and examining how the system behaves under different scenarios. MSBE is particularly well-suited for exploring how different approaches to preparedness and levels of resources (e.g., stockpiles, emergency responders, and community volunteers) affect outcomes. The purpose of our research was to develop a computer simulation tool that provides rural communities with an in-depth understanding of how rural disaster preparedness systems interact and improves response capabilities of their healthcare systems to natural hazard events.

Methods: We defined an effective and efficient public health emergency response as one which is well-coordinated, timely, and minimizes losses and then explored the independent, system-dependent, and critical factors that help produce successful responses in rural communities. The definition of healthcare system was holistic and inclusive of all organizations, institutions, resources, and people whose main goal is to improve health and well-being of individuals. We narrowed the broad set of natural hazards on a subset that had acute onset, short duration of impact, and that occurred frequently in the United States, with flooding being an ideal starting point. We conducted stakeholder engagement workshops in two case study sites in Nebraska and Tennessee where actual flood events had recently impacted rural communities. We incorporated the outputs of these workshops with MBSE computer simulation methods, including system dynamics and agent-based modeling, to design a computer simulation model that accurately characterizes the key systems involved in a disaster response. The simulation model was also developed to track key public health and healthcare outputs that occur during and after such an event (e.g., number of individuals in need of rescue or medical care, patient wait times, surge, resource management).

Results: The first case study was a flood event that occurred in Waverly, Tennessee in August 2021; the Waverly flood was an acute event, geographically isolated, occurred with little warning, impacted a rural community, and had a high fatality rate. The second case study was a flood event that occurred in Nebraska in March 2019; the Nebraska flood was an acute event, geographically widespread, occurred with more warning, impacted many rural communities, and resulted in extensive damage but a low fatality rate.

Based on stakeholder feedback, we built a simulation that included the following functional elements, key factors, and variables. Functional elements include several preparedness and response subsystems (i.e., public health, healthcare, emergency management, and emergency response) that were integrated into a “system of systems”. Key factors included interagency trust of the healthcare and emergency resource organizations involved, the level of interagency preparation for the event, the strength of the relationships between these organizations, the experience of the organizations with the event, the communication levels between subsystems and with potential outside organizations that can provide support, and the resilience of the area. The variables included a variety of user-modifiable components across the subsystems, such as healthcare (e.g., amount of healthcare resources, location of healthcare centers), hazard-related (e.g., areas affected, level of inundation), population-related factors (e.g., age, race, gender, and other social determinants of health of those affected), among others.

Conclusion: Our simulation was able to accurately represent the interactions of these system elements using a multi-method systems modeling approach showing the respective interactions and relationships, as well as predict the actual outcomes experienced by these communities as the case study events unfolded. We were able to use the simulation to depict best- and worst-case flood scenarios for these communities that yielded plausible outputs. This public health disaster preparedness tool provides an in-depth understanding and simulation of how rural preparedness systems interact to prepare for and respond to a disaster. It will be useful to test existing rural preparedness policies and procedures, determine how to use existing resources most effectively, and identify where additional resources are needed and will have the greatest impact.