Modelling Hospital Strategies in City-Scale Ambulance Dispatching

Xinyu Fu, Valeria V. Krzhizhanovskaya, A. A. Yakovlev, Sergey V. KovalchukarXiv 2022 · DOI · arXiv

Abstract

The optimisation in the ambulance dispatching process is significant for patients who need early treatments. However, the problem of dynamic ambulance redeployment for destination hospital selection has rarely been investigated. The paper proposes an approach to model and simulate the ambulance dispatching process in multi-agents healthcare environments of large cities. The proposed approach is based on using the coupled game-theoretic (GT) approach to identify hospital strategies (considering hospitals as players within a non-cooperative game) and performing discrete-event simulation (DES) of patient delivery and provision of healthcare services to evaluate ambulance dispatching (selection of target hospital). Assuming the collective nature of decisions on patient delivery, the approach assesses the influence of the diverse behaviours of hospitals on system performance with possible further optimisation of this performance. The approach is studied through a series of cases starting with a simplified 1D model and proceeding with a coupled 2D model and real-world application. The study considers the problem of dispatching ambulances to patients with the ACS directed to the PCI in the target hospital. A real-world case study of data from Saint Petersburg (Russia) is analysed showing the better conformity of the global characteristics (mortality rate) of the healthcare system with the proposed approach being applied to discovering the agents' diverse behaviour.

Introduction

In major metropolitan areas, hospital overcrowding is a critical factor influencing patient outcomes, particularly for time-sensitive conditions like Acute Coronary Syndrome (ACS). For these patients, every minute of delay in receiving treatment, such as stenting, directly increases mortality risk. Traditional ambulance dispatching models often rely on the "principle of proximity," assuming patients are always sent to the nearest facility. However, empirical data suggests a more complex reality: over 40% of patients are not treated at the nearest hospital. This discrepancy arises because hospitals are active agents that may redirect patients based on internal constraints like workload, specialist availability, and equipment status. Most existing simulations fail to account for this strategic behavior, assuming static, globally optimal rules that do not reflect the decentralized decision-making inherent in urban healthcare systems.

Methodology: The GT-DES Framework

To bridge this gap, we propose a hybrid GT-DES model that couples Discrete-Event Simulation (DES) with Game Theory (GT). This approach allows for the modeling of both the physical logistics and the strategic decision-making of the healthcare system.

  1. Discrete-Event Simulation (DES): This component models the system's operational aspects, including patient arrivals (modeled as a Poisson process), ambulance travel times through dynamic city traffic, and the internal queuing and treatment processes within hospitals.
  2. Game Theory (GT): This component models hospitals as independent players in a non-cooperative game. Each hospital chooses between two primary strategies: "Accepting" (A) or "Redirecting" (R). The model seeks to identify the Nash Equilibrium, a state where no hospital can improve its own performance metric (a score balancing patients served against service time) by unilaterally changing its strategy.
Integrated flowchart of the patient, EMS, and hospital perspectives within the GT-DES model.

The integration of these frameworks allows the simulation to account for the "psychology" of the system, where the Emergency Medical Service (EMS) acts as a messenger, but the final destination is determined by the strategic responses of the receiving hospitals.

Case Study: Saint Petersburg

The model was validated using a comprehensive dataset from Saint Petersburg, Russia, encompassing:

  • Patient Records: 5,124 ACS cases from 2015.
  • Hospital Network: 13 hospitals with Percutaneous Coronary Intervention (PCI) facilities.
  • Service Times: 1,866 "stent time" records to establish realistic treatment duration distributions.
  • Traffic Data: A network of 143 departure nodes to estimate ambulance travel times, accounting for urban factors like rush hours and bridge openings.

Results and Discussion

The GT-DES model demonstrated significant improvements in predictive accuracy over standard models. A key finding is that "Accepting" is not always the globally or locally optimal strategy. As patient flow increases, hospitals may rationally choose to redirect patients to prevent system-wide overcrowding and "explosions" in wait times.

Global system performance (time) vs. patient flow for various hospital strategy combinations (A: Accept, R: Redirect).

Key insights from the analysis include:

  • Mortality Correlation: The model achieved a Pearson correlation of 0.80 (p=0.0064) between simulated and observed mortality rates.
  • Accuracy Boost: Accounting for diverse hospital behaviors improved mortality prediction precision by 12–14%.
  • Sensitivity: Sobol sensitivity analysis revealed that arrival and service rates are more critical to system performance than ambulance speed.
  • Non-Cooperative Reality: The global optimum (best for the city) often differs from the Nash Equilibrium (what hospitals actually do). Modeling this "selfish" behavior is essential for accurate system representation.

Conclusion

The GT-DES framework provides a more realistic lens for understanding city-scale ambulance dispatching by treating hospitals as strategic agents. While currently limited by binary strategies and a focus on emergency patients, future work will incorporate more granular data, such as physician shift changes and non-emergency patient loads. This model serves as a foundation for developing practical decision-support tools for urban healthcare policy and management.

Summary updated

Abstract

The optimisation in the ambulance dispatching process is significant for patients who need early treatments. However, the problem of dynamic ambulance redeployment for destination hospital selection has rarely been investigated. The paper proposes an approach to model and simulate the ambulance dispatching process in multi-agents healthcare environments of large cities. The proposed approach is based on using the coupled game-theoretic (GT) approach to identify hospital strategies (considering hospitals as players within a non-cooperative game) and performing discrete-event simulation (DES) of patient delivery and provision of healthcare services to evaluate ambulance dispatching (selection of target hospital). Assuming the collective nature of decisions on patient delivery, the approach assesses the influence of the diverse behaviours of hospitals on system performance with possible further optimisation of this performance. The approach is studied through a series of cases starting with a simplified 1D model and proceeding with a coupled 2D model and real-world application. The study considers the problem of dispatching ambulances to patients with the ACS directed to the PCI in the target hospital. A real-world case study of data from Saint Petersburg (Russia) is analysed showing the better conformity of the global characteristics (mortality rate) of the healthcare system with the proposed approach being applied to discovering the agents' diverse behaviour.
BibTeX
@misc{fu2022modelling,
  author = {Xinyu Fu and Valeria V. Krzhizhanovskaya and A. A. Yakovlev and Sergey V. Kovalchuk},
  title = {{Modelling Hospital Strategies in City-Scale Ambulance Dispatching}},
  year = {2022},
  eprint = {2201.01846},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2201.01846},
}
Figure