Research on the value screening of combat rescue information based on the Markov decision tree model

In future digital battlefields, the systematic collection of combat casualty care information is essential for improving trauma outcomes and guiding medical resource allocation. However, a significant methodological gap exists in determining which specific data points are most valuable to record when prospective wartime data are unavailable. Rather than constructing a physical information collection system, this study aims to develop a simulation-based screening framework that can identify and prioritize high-value examination and treatment information across the continuum of tactical combat casualty care.

A Markov decision tree model was used to simulate the effects of different treatment techniques on injury-state transitions during casualty evacuation, spanning line-of-fire care, tactical medical care, and damage-control treatment at rescue institutions. Inspection and treatment techniques recorded in battlefield medical documents were screened and ranked based on expert-informed evaluation.

Simulation outputs indicate that the on-site treatment phase exerts the most substantial impact on injury state evolution, with comparatively smaller changes observed during ground and air evacuation. The model differentiated the information value of various diagnostic and therapeutic techniques, yielding a prioritized list of essential data elements specific to each care echelon. These include immediate life-saving interventions (e.g., tourniquet application time, needle decompression status) and continuous monitoring parameters (e.g., pulse oximetry trends, serial Glasgow Coma Scale assessments) that should be captured in field medical documentation.