Two-Tiered Ambulance Dispatch and Redeployment considering Patient Severity Classification Errors.

Area of Science:

• Operations Research
• Emergency Medical Services Management
• Healthcare Systems Engineering

Background:

• Two-tiered ambulance systems (advanced and basic life support) offer cost-efficiency but require accurate patient severity classification.
• Existing ambulance dispatch models often overlook patient classification errors and diverse patient/ambulance types.

Purpose of the Study:

• To address the two-tiered ambulance dispatch and redeployment problem considering patient classification errors.
• To develop an efficient algorithm for managing ambulance resources in dynamic, uncertain environments.

Main Methods:

• Formulation of the problem as a semi-Markov decision process.
• Development and application of a mini-batch monotone-approximate dynamic programming (ADP) algorithm.
• Computational experiments using historical data from Seoul.

Main Results:

• The proposed ADP algorithm reduced the risk level index (RLI) by 11.2% compared to a greedy policy.
• Analysis identified the significant impact of undertriage rates on patient RLI, exceeding that of overtriage.
• System parameters like the proportion of advanced life support units were evaluated.

Conclusions:

• The developed algorithm provides an efficient strategy for two-tiered ambulance management.
• Findings offer practical guidelines for minimizing undertriage rates in patient severity classification.
• Accurate patient classification is crucial for optimizing emergency medical service performance.