Achkoski Jugoslav, Koceski S, Bogatinov D, Temelkovski B, Stevanovski G, Kocev I
Military Academy, "General Mihailo Apostolski", Skopje, Macedonia.
Faculty of Computer Sciences, University "Goce Delcev", Štip, Macedonia.
J R Army Med Corps. 2017 Jun;163(3):164-170. doi: 10.1136/jramc-2015-000616. Epub 2016 Jul 14.
This paper presents a remote triage support algorithm as a part of a complex military telemedicine system which provides continuous monitoring of soldiers' vital sign data gathered on-site using unobtrusive set of sensors.
The proposed fuzzy logic-based algorithm takes physiological data and classifies the casualties according to their health risk level, calculated following the Modified Early Warning Score (MEWS) methodology.
To verify the algorithm, eight different evaluation scenarios using random vital sign data have been created. In each scenario, the hypothetical condition of the victims was assessed in parallel both by the system as well as by 50 doctors with significant experience in the field. The results showed that there is high (0.928) average correlation of the classification results.
This suggests that the proposed algorithm can be used for automated remote triage in real life-saving situations even before the medical team arrives at the spot, and shorten the response times. Moreover, an additional study has been conducted in order to increase the computational efficiency of the algorithm, without compromising the quality of the classification results.
本文提出一种远程分诊支持算法，作为复杂军事远程医疗系统的一部分，该系统使用一套不引人注意的传感器对现场收集的士兵生命体征数据进行持续监测。
所提出的基于模糊逻辑的算法采用生理数据，并根据按照改良早期预警评分（MEWS）方法计算出的健康风险水平对伤亡人员进行分类。
为验证该算法，创建了八个使用随机生命体征数据的不同评估场景。在每个场景中，系统以及50名在该领域有丰富经验的医生同时对受害者的假设情况进行评估。结果表明分类结果的平均相关性很高（0.928）。
这表明所提出的算法甚至在医疗团队到达现场之前就可用于实际救生情况下的自动远程分诊，并缩短响应时间。此外，还进行了一项额外研究，以提高算法的计算效率，同时不影响分类结果的质量。
