1 Introduction

When addressing disaster preparedness (DP), rural America faces unique challenges related to persistent and interrelated disparities and scarcities in health and healthcare, infrastructure, and economics. The goal of this project was to take the first steps to create a tool to help evaluate and address DP in rural areas. To achieve this goal, the researchers applied systems thinking and model-based systems engineering (MBSE) to the development of a proof-of-concept, multi-method computer simulation to assess the efficacy of disaster planning on health outcomes in rural communities, as a function of primary healthcare.

1.1 Rural Community Characteristics

Compared to their urban counterparts, rural areas are challenged in several areas, including population distribution, economic prosperity, healthcare access, and DP. Rural and urban areas generally are differentiated by population density, the concentration of human-created structures, and proximity to urban areas (Ratcliffe et al. 2016; USCB n.d.; USDA 2019). First and foremost, the population distribution in rural areas affects resource availability. While 97% of the land area in the United States is rural, only 20.0% of the population (approximately 66 million Americans) reside in a rural area (USCB 2016; USDA 2019; Human Resources & Services Administration 2022; USCB 2022). Additionally, population growth in rural counties lags behind urban and suburban counties, with half of rural counties having fewer residents than in 2000 (Parker et al. 2018).

Rural areas are behind urban areas on nearly every level of prosperity, from poverty rates to labor force participation. Almost twice as many rural counties are economically distressed than are prosperous (Economic Innovation Group 2021), and nearly 75% of employment growth is concentrated in major metropolitan cities (Hendrickson et al. 2018). Furthermore, rural residents tend to be older, sicker, and poorer than their urban counterparts.

Compared to urban areas, rural areas are also characterized by inadequate access to primary healthcare, hospitals, emergency services, and public health services. While close to 20% of Americans live in rural areas, only 10% of physicians practice in these areas (Johnt et al. 2016). Rural America accounts for two-thirds of the primary care health professional shortage areas (USDA 2019). Furthermore, these areas face a persistent shortage of clinical specialists (Johnson 2022) and qualified applicants for non-clinical positions (Medical Group Management Association 2018).

Another issue that affects rural areas is rural hospital closures, which is an acute and worsening problem. For example, since 2013, a third of hospital closures has occurred 20 or more miles from the next closest hospital (Medicare Payment Advisory Council 2018). Moreover, the U.S. Government Accountability Office (2020) found the median distance to common inpatient care for affected communities increased by 20.5 miles from 2012 to 2018. This distance increased even more for specialist services. This decline in rural hospital services represents a problematic situation, particularly in the event of a natural hazard in an affected community.

Disaster preparedness in rural communities is hampered by healthcare providers and facilities having insufficient capacity, resources, funding, equipment, and personnel. Due to the scarcity of these resources, rural areas are already challenged to meet even the daily needs of residents. This situation raises concerns about the limited ability to respond to natural hazards in rural areas (Hoard et al. 2005). Furthermore, many rural hospitals have reduced resources (for example, emergency and inpatient capacity and staff), most notably in critical access hospitals. In rural areas without a hospital, community health centers, rural health centers, and other unaffiliated providers may not participate in DP planning and activities (USA Center for Rural Public Health Preparedness 2007).

1.2 Systems Approaches and Computer Simulation

The use of systems theory and computer simulation to describe systems behavior has proven useful for public health and disaster research and planning. Computational simulation is an ideal solution for advanced disaster research because it considers hazard impacts from a systems perspective; addresses uncertainties in the inputs by examining how the system behaves under different scenarios; and provides decision-making support. Pairing a systems approach with computer simulation integrates qualitative and quantitative data and assimilates empirical data with existing interdisciplinary theories. The systems approach first can explain how interactions among agents or elements at multiple scales contribute to the system’s emergent behaviors across different levels (individual and community) or dimensions (preparedness and recovery) (Mostafavi and Ganapati 2021). The computer simulation then can assess the DP activities’ effect on post-disaster rural health outcomes. Finally, policies, interventions, and mitigation activities can be evaluated to determine which system changes could significantly affect rural health outcomes during natural hazards or other emergencies.

Model-based systems engineering is a methodology used in engineering and other fields to understand or support the development, design, analysis, verification, and validation of complex systems models (Ramos et al. 2011). Model-based systems engineering approaches incorporate different modeling techniques (for example, system dynamics, agent-based, and discrete events) and systems thinking into digital modeling environments or computer simulations. The modeling techniques differ depending on how the system or components are considered. For example, system dynamics modeling uses a reductionist philosophy, focusing on the system aggregate. In contrast, discrete event modeling uses an analytic approach, focusing on individual events and decisions (Morgan et al. 2017).

Individuals seeking to address the deficiencies in current DP and response systems—particularly in rural or non-urban settings with limited resources and capabilities—need to better understand the impacts, contributions, and interactions of the different factors that make up existing systems. Such understanding will allow for the prioritization of interventions, mitigation activities, and other investments to optimize DP and responses, with the ultimate goal of improving public health and economic outcomes. To assess the efficacy of disaster planning on the health effects in rural communities, MBSE approaches were used to develop a planning and predictive analytics tool based on a conceptual complex systems representation of the interactions of primary healthcare and DP.

1.3 Use of Systems Approaches in Disaster Preparedness

Systems theory and MBSE have been used to describe system behaviors in healthcare (Chaffee and McNeill 2007), public health (Homer and Hirsch 2006; Moore et al. 2011), and disaster management (Kapucu 2009; Coetzee et al. 2016). Computer simulation has much potential in disaster research because it can consider disasters from a systems perspective, address uncertainties by examining how the system behaves in different scenarios, and provide decision support. Simulation can integrate different types of data, assimilate empirical data with interdisciplinary theories, and explain how interactions of elements at multiple scales contribute to the emergent behaviors of the system at different levels (individual, community) or dimensions (preparedness, recovery) (Mostafavi and Ganapati 2021).

Hoard et al. (2005) called for the use of these methods in rural disaster planning over 19 years ago. However, minimal research has considered disaster preparedness in rural healthcare settings as a complex adaptive system. Only six studies were identified in a recent scoping review of MBSE applications in rural healthcare system disaster management (Berg et al. 2023). Biological disasters were the most common applications for MBSE methods, with three related to COVID-19 (Allen et al. 2020; Savitsky and Albright 2020; Kasturi et al. 2021) and one related to Dengue virus (Chovatiya et al. 2019). Bioterrorism (Patvivatsiri et al. 2007) or all-cause disaster (Toerper et al. 2018) were other MBSE applications. Studies were evenly split across disaster preparedness and response applications. However, none of the identified studies focused solely on rural areas, rather rural areas were included in the larger geographic context (for example, entire state or country) along with urban areas.

Some studies have applied such methods to create whole-hospital simulations during disaster response or recovery (Kirsch et al. 2018; TariVerdi et al. 2019; Shahverdi et al. 2022). Whole-hospital approaches often use discrete event modeling to examine the movement of patients through the hospital. For example, Kirsch et al. (2018) used a discrete event modeling approach to improve mass casualty incident response in a typical urban hospital. Other studies have used multi-method approaches to support hospital system resilience in specific types of disasters. For example, Hassan and Mahmoud (2020) used a multi-method approach to develop a hospital system framework that can be used to evaluate system-level functionality after an earthquake. Similarly, these investigators applied similar methods to support system-level performance during compound disaster events (Hassan and Mahmoud 2021). We are unaware of other investigators who have incorporated multiple response system components in their simulation approaches to support disaster management.

2 Methods

To assess whether the complex processes of rural healthcare in the overall DP context could be represented by a single computer model, a multi-disciplinary team of researchers in public health, health policy, nursing, disaster planning and preparedness, systems engineering, and computer simulation was assembled. A scoping review (Berg et al. 2023) and additional review of the literature focusing on systems thinking and systems modeling in the context of rural healthcare systems and rural DP was performed. This process provided a contextual reference to the complexities of rural healthcare and DP and helped generate a roadmap to develop the model concept.

The information obtained from the existing literature led to four conclusions. First, rural healthcare and DP can be considered dynamic complex systems consisting of a number of subsystems. Second, common variables or relationships exist between rural healthcare and DP that would allow a simulation to effectively model these systems’ interrelationships. Third, while these subsystems have some common elements, they also have structural and functional differences that would benefit from different modeling approaches. Fourth, each subsystem is complex. As a first step, a conceptual model was developed, followed by the creation of a proof-of-concept computer simulation by the systems engineering and computer simulation team members.

2.1 Conceptual Model Design

Based on information obtained from our experts and the literature review process, a proof-of-concept computer model was constructed using a complex systems approach. This simulation includes several major subsystems making up or contributing to the overall rural disaster preparedness and response system. This computer simulation along with key variables and outputs represents a practical realization of a conceptual model representing disaster effects on a rural healthcare system. The researchers endeavored to include elements that would allow each subsystem to demonstrate its respective interactions. Many factors affect the subsystems and their interactions. The researchers engaged the law of parsimony and chose factors that appeared to have the most significant effect based on the literature and expert input (USA Center for Rural Public Health Preparedness 2007).

Three key decisions served to structure the model. First, the model’s operational scale was set at the rural county level to provide an accurate scale for considering the potential population affected, likely access to an acute care facility, and availability of a structured public health and emergency management function. In the United States, a “county” (or county equivalent) is a political or administrative state subdivision consisting of a geographic region with specific boundaries and some level of governmental authority (National Association of Counties n.d.; The Law Dictionary n.d.). Second, a notional generic rural county was created with the appropriate features (for example, land mass, population size, and disaster effects). Third, the model’s initial focus was the initial phase (that is, the first five days) of a natural hazard event.

With these considerations in mind, the system and major subsystem elements were defined. The subsystem elements were identified as the healthcare system, public health system, and emergency management system (EMS). A fourth subsystem identified was the emergency response (first responder) system. This model did not explicitly include this system because its key elements were subsumed into other portions of the proof-of-concept model environment.

2.2 Systems Approach and Model Definition

A system can be defined as “a collection of various structural and nonstructural elements that are connected and organized in such a way as to achieve some specific objective through the control and distribution of material resources, energy, and information” (Simonovic 2020, p. 5). Disaster preparedness and response have several independent, yet interrelated elements. A conventional approach views the systems as linear, considering only one relationship at a time. However, effective planning of operations prior to a hazard event requires evaluating each of these elements as interacting by appropriate means and at appropriate levels (Miller et al. 2006).

The envisioned model consisted of three primary subsystems that make up the typical, overall, rural disaster preparedness and response system: the rural healthcare system, the public health system, and EMS (Fig. 1). This model is a proof-of-concept computer simulation to assess the feasibility of adapting MBSE to the complex system of DP for rural healthcare systems with the overarching objective of improving the efficiency and effectiveness of DP planning for rural communities. The simulation was designed to represent the operation of each subsystem and the effect their interactions had on overall health outcomes post-disaster.

Fig. 1
figure 1

Systems structure of the rural healthcare disaster preparedness model

Full size image

In the context of the model, a healthcare system can represent any number of facilities, from a single facility to multiple facilities. A rural healthcare system is comprised of the presence or absence, of healthcare provider types (for example, rural hospitals, clinics, and primary care practices). To maintain simplicity for this initial model, a rural hospital was used to represent the functioning of a rural healthcare subsystem. The healthcare system’s performance was selected as the response variable of interest, which was defined as the availability of healthcare system resources to provide healthcare to the population affected by the natural hazard event. This variable measures the change in the healthcare system’s performance at the event occurrence, as a deviation from the system’s baseline (non-disaster) performance. Changes in this variable as the event progresses indicate how the healthcare system is responding to the event.

2.3 Model Approach Selection

The researchers explored the conceptual approaches and applied MBSE. This methodology considers each of the system’s elements and attempts to transform the system’s operations—including the appropriate subsystems, their interactions, operations, interfaces, and emergent phenomena—into a phenomenological representation using computer-based simulations to represent the system dynamically. There are a variety of MBSE computational modeling approaches, each with its features and limitations based on the application. The researchers evaluated several of these approaches to translate the conceptual model into a computational model.

2.3.1 Rationale for Hybrid Modeling

While a single modeling approach often is selected to represent a system, this approach does not need to be an either/or decision. Single modeling approaches can be combined into hybrid approaches that capitalize on the benefits of two or more techniques. We selected a multi-method or hybrid modeling approach for simulation development, integrating discrete event simulation (DES) and system dynamics (SD), to conceptualize rural disaster preparedness and response. This methodology is chosen to capture the complex interplay of healthcare delivery, emergency management, and public health infrastructure in rural settings.

Discrete event simulation is utilized to simulate specific, linear processes within the healthcare system. This includes modeling patient flow, resource allocation, and operational efficiencies in acute care settings during disasters. Discrete event simulation provides a detailed view of how healthcare facilities respond to emergencies, identifying potential bottlenecks and optimizing resource utilization (Chahal and Eldabi 2008; Brailsford et al. 2019).

System dynamics is employed to capture the broader systemic interactions and accumulative effects over time, particularly in emergency management and public health sectors. This approach aids in understanding the systemic impacts of disaster preparedness strategies and interventions. It models how policies and resource allocations evolve and influence the overall readiness and effectiveness of rural healthcare systems in disaster scenarios (Peterson and Eberlein 1994; Homer 2020).

Each modeling approach, while valuable, has inherent limitations. Discrete event simulation, although excellent for process simulation, does not capture broader system-level interactions and long-term effects. Conversely, SD provides a macro-level view but can oversimplify individual behaviors and discrete events. The hybrid approach overcomes these limitations by combining the operational detail of DES with the systemic perspective of SD, offering a more comprehensive and nuanced simulation (An 2012; Macal 2016).

The hybrid modeling approach offers flexibility and scalability, crucial in rural settings where conditions can rapidly change. This adaptability allows for adjustments and expansions in the model as new data become available or as scenarios evolve. This methodology enables the simulation of a range of scenarios, facilitating comprehensive scenario planning. The insights gained are instrumental in informed decision making and effective policy development, ensuring that strategies are not only effective but also resource-sensitive and adaptable to changing rural healthcare dynamics.

2.3.2 Healthcare System: Discrete Event Simulation (DES)

The use of DES in modeling the healthcare system is grounded in its ability to simulate sequential and time-specific processes, which are characteristic of acute care settings in healthcare systems. Discrete event simulation is particularly effective in environments where processes are linear, resource-dependent, and have definable logic, making it ideal for simulating patient flow and resource utilization in healthcare facilities (Jun et al. 1999). By tracking individual entities (for example, patients, medical supplies) through the system, DES provides a detailed view of operational efficiencies and bottlenecks, which is crucial for emergency preparedness and response planning (Pidd 2004).

2.3.3 Emergency Management: Systems Dynamics (SD)

Systems dynamics modeling is chosen for emergency management due to its strength in representing complex systems characterized by feedback loops and nonlinear interdependencies (Sterman 2000). System dynamics is adept at capturing the accumulative effects of resource changes and systemic interactions over time, essential for understanding the dynamics of emergency preparedness and response (Forrester 1961). This approach allows for the exploration of how different policies and resource allocations impact the system’s overall readiness and effectiveness in disaster scenarios (Richardson and Pugh III 1997).

2.3.4 Public Health: Systems Dynamics (SD)

In public health, SD modeling is preferred for its ability to represent complex, systemic interactions and accumulative impacts over time. Public health systems involve numerous interacting components, including healthcare policies, resource distribution, and population health trends, which are well-suited for SD’s holistic approach (Homer and Hirsch 2006). System dynamics models can simulate how interventions and resource allocations impact public health outcomes over extended periods, providing valuable insights for long-term disaster preparedness planning (Lane et al. 2000).

Each subsystem was modeled as an independent subsystem with specific links to allow interactions between them. AnyLogic® 8.7.9 simulation softwareFootnote 1 was used as the model software development platform, as it allows for the creation of integrated systems dynamics, DES, and other models.

2.4 Model Subsystems

The overall system can be divided into several subsystems. The overall system and the component subsystems along with their interrelationships are shown in Fig. 2.

Fig. 2
figure 2

Relationship diagram demonstrating subsystem relationships and interactions

Full size image

2.4.1 Public Health Subsystem Computational Model

The researchers included several main components and related subcomponents (Table 1). The main components are broken down into communications, policies and procedures, resources, and training and exercises. These components feed into improving or increasing the local public health agency’s preparedness (the main stock of the system dynamics model). The level of preparedness is then drawn down or affected by the disaster response that occurs during a natural hazard. To create a rural public health system with the flexibility to handle the varying governance types (for example, local, state, and shared/mixed), separate local and regional components were included where necessary.

Table 1 Major public health subsystem components and subcomponents
Full size table

2.4.2 Rural Healthcare Subsystem Computational Model

To model a specific natural hazard scenario, a healthcare facility was modeled to mimic the flow of people through the rural healthcare system and include the workers and facilities. The input variables include the number of facility rooms, number of healthcare providers (for example, nurses and physicians), and the baseline rate at which the affected population enters the system under normal conditions (Table 2). A natural hazard scenario can be simulated based on the duration of the event (impact time), the number of people injured, and the average injury severity (Table 3). The average severity of a person’s injury is influenced by the state of the public health and emergency management models at the time of the event, which in turn influences the outcomes of the person’s experiences and the results of their injury. Resource utilization rates help determine the overall healthcare system performance.

Table 2 Major rural healthcare subsystem components and subcomponents
Full size table
Table 3 Major inputs describing the natural hazard event
Full size table

2.4.3 Emergency Management Subsystem Computational Model

The subsystem architecture was modeled by explicitly including the major components of an emergency management function (Table 4). The associated system dynamics model is the aggregation of these functions, which were connected by creating feedback loops and establishing relationships controlling their interactions. These components relate these functions in system dynamics parlance, with the key stocks being resources and EMS infrastructure. These stocks are conditioned by the initial number of resources, external influences, resource utilization over time, and the EMS infrastructure’s ability to sustain functioning over time.

Table 4 Major emergency medical assistance (EMA) system functions
Full size table

2.4.4 Model Operations

The simulation model was constructed as an integrated simulation with each model element directly interfaced with among all elements. Model inputs are controlled via user inputs to demonstrate system interactions. These inputs can be changed either as file inputs or real-time during model operation. Model time was set to represent actual disaster event time. The nominal duration of the disaster event was set at four weeks representing the time from the start of the effects of the disaster event during the initial acute phase. Standard model inputs were determined by information available in the literature and expert input.

2.5 Verification and Validation

Several approaches were used to verify and validate the proof-of-concept model. Verification was accomplished through a detailed assessment of the computer code and the model’s logic flow. The software development platform furnished a graphical user interface providing a direct means of viewing the logic flow of each model element. The model’s computer code was built with internal traces for error checking.

Validation was performed by developing and running use cases with expected outcomes of a disaster scenario in a rural community. Use cases were designed based on information from representative disaster events, expert input, and government resources. These use cases were developed to test the model’s expected boundary conditions. Actual disaster events in rural communities were identified, and the available associated data were used to evaluate the model’s output against the outcome of the example disaster event.

This research project was evaluated by the University of Tennessee Institutional Review Board (IRB), which determined that it did not require IRB review since the proposed research does not involve human subjects as defined by federal regulations.

3 Results and Discussion

This project makes four important contributions to the literature on rural healthcare systems and DP and planning. First, the researchers demonstrated that a multi-method simulation model can represent a rural healthcare system’s operation, including the effects of the pre-event public health and disaster management system elements. Second, the researchers created a proof-of-concept model that represents the major subsystems and their respective interactions, which have been reduced to a computational model. Third, this model provided initial results demonstrating the effects of planning changes in the public health and emergency planning subsystems on rural healthcare system performance. Fourth, the project indicated that an MBSE approach is an effective method for combining the subsystem components of healthcare and DP.

This project’s hybrid modeling approach aligns with the public health field’s desire to increase the use of modeling and simulation to accelerate innovation (Maglio et al. 2014). This model also provides a quantitative output of system element behaviors. This was accomplished by providing an output in the form of an index for healthcare system performance that indicates the causal relationship between subsystem factor changes and healthcare system effects. In a sense, this index is a measure of “goodness” of the system’s performance. The model accurately represents the effects of communication, planning interactions, funding availability, community resilience, and initial preparedness. The index is a relative measure of system performance that is a ratio of pre-disaster performance to post-disaster performance. The Emergency Management and Public Health indices are defined similarly.

A representative notional model output derived from simulation of the validation is shown for each subsystem (Fig. 3). These outputs are driven by the inputs to the simulation, which are based on data obtained and expert insight on response to disaster events. Multiple simulations were performed to create an overall composite representation of the outputs given the predefined input values. In each instance, te is the time of the event occurrence. The healthcare subsystem’s output analysis revealed an initial decline in performance at the onset of a natural hazard event. This is particularly significant in the rural context where acute care hospitals are primary providers. The decline is believed to reflect the immediate strain on limited resources, such as healthcare providers and facility rooms—critical challenges in rural healthcare settings. As the event progressed, the sustained lower performance underscored the rural healthcare system’s difficulties in coping with sudden increases in demand due to disasters. However, the system began to recover as the impact of the event diminished, indicating not only the replenishment of resources but also the resilience inherent in the rural community and healthcare staff.

Fig. 3
figure 3

Model subsystem outputs representing subsystem response from event onset to end of simulation

Full size image

In the public health subsystem, the performance index showed a decrease followed by oscillations after the event’s onset. These oscillations reflect the challenges faced by rural public health systems, including inadequate access to services and persistent shortages of clinical specialists. The hysteresis observed in the recovery process underscores the time lag in implementing public health interventions and adapting to changing scenarios in rural settings. This may be due to limited resources, infrastructure, and varied governance types (local, state, shared/mixed).

The emergency management subsystem exhibited more moderate diversions compared to public health, suggesting a more resilient structure. This could be attributed to structured public health and emergency management functions at the county level, reflecting the aggregation of resources and infrastructure in emergency management. These interactions between the emergency management subsystem and the healthcare and public health subsystems are crucial, especially in rural settings where distances to facilities and availability of resources significantly impact disaster response effectiveness.

The model’s adaptability to simulate various disaster scenarios through adjustments in parameters like disaster severity and duration is instrumental in understanding and planning for diverse disaster situations in rural contexts. This flexibility is a crucial tool for policymakers and practitioners, guiding the development of versatile and effective DP strategies that are sensitive to the unique challenges of rural healthcare.

Our findings contribute to the broader discussion on challenges faced by rural healthcare systems, such as professional shortages and limited service access. The model provides a valuable tool for addressing these challenges, offering data-driven insights to inform policy and practice improvements in rural disaster preparedness.

In our subsystem analysis, we observed a decline in healthcare system performance during the onset of disasters, highlighting the vulnerability of rural healthcare infrastructures. This underscores the need for robust preparedness strategies, focusing on enhancing resilience and adaptive capacity in the face of disasters. The fluctuating public health performance post-disaster onset and the relative stability in emergency management operations revealed through our model provide insights into potential areas for strengthening disaster response mechanisms in rural settings.

These findings also underscore interest in addressing social and infrastructural aspects of disaster risks. The inclusion of graphical representations of model outputs enhances the comprehensibility of complex subsystem interactions.

4 Assumptions, Strengths, and Limitations

Several key assumptions were made for this model’s design. As noted, the hospital system was represented as an acute care hospital. This approach limits the considerations of other elements that may exist in rural healthcare systems, such as primary care and urgent care facilities. However, the model is flexible enough to allow for minor adjustments to the healthcare inputs (that is, number of doctors, nurses, and facility rooms), which would enable the simulation to model other types of healthcare facilities.

More work is needed to understand each subsystem’s actual operation and map these operations into the model. In particular, the emergency response subsystem needs to be added to more accurately simulate an actual natural hazard event. Including additional information about the broader healthcare system will enhance the model’s ability to adapt to different circumstances. Unfortunately, data related to the subsystems and their operations are either difficult to obtain or nonexistent. Expert opinion or input from our research team—such as that used in developing the conceptual model—may have to be the default in such situations.

While our model offers contributions to disaster risk science, particularly in rural healthcare, it also presents limitations in terms of broader applicability and accuracy. Future research should incorporate additional and expanded data sources from actual preparedness activities or hazard events, which is often inaccessible due to national security concerns, to refine the model’s accuracy and extend its scope to encompass a wider range of healthcare facilities, deepening our understanding of rural healthcare dynamics in disaster preparedness and response.

Finally, this model was developed as a proof-of-concept design to assess the feasibility of the methodology and design approach. As such, the end-user community was not extensively engaged in all aspects of its design. For future model iterations, the developers will solicit the direct involvement of the end-user community to improve and refine the simulation for broader applicability.

5 Conclusion

This computational model serves as a prototypical representation of the operation of rural healthcare systems in the context of disaster preparedness and planning. By using the hybrid modeling approach, the researchers were able to capture the key operating elements of rural healthcare, public health, and emergency management functions. This model serves as a first step to a more comprehensive approach to supporting rural communities in more efficient and effective healthcare planning for responding to natural hazard events.

The researchers envision that this model will evolve to represent the system with higher fidelity. Logical next steps in advancing the model include expanding the data used in designing and validating the model, representing the rural healthcare system beyond the operations of a single hospital/healthcare facility (including offering more detail on population impacts), and enhancing the subsystems to provide better measures.

More immediate efforts will be to seek data sources to validate or eliminate the input variables’ influence on the overall healthcare system performance during normal operation and during a disaster. Such efforts will promote the creation of a more parsimonious model with a simple user interface that provides actionable results. Machine learning and artificial intelligence will be considered to support the model in adapting to changes in resources or disaster conditions.