Development of a drone integrated mass casualty triage algorithm

This study, which aims to develop a triage algorithm to determine the urgency level of casualties through the acquisition and processing of casualty images by drones, was carried out within the scope of the project numbered 222S343, supported by the TÜBİTAK 1001 - Scientific and Technological Research Projects Support Program. The study was completed in two phases between March and December 2024: (1) the development of the draft DIMaCTA Algorithm and (2) testing the validity and reliability of the algorithm. The development process included three steps: (1) a literature review based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol, (2) the creation of a draft algorithm by a working group led by the thesis author, combining findings from the literature and field experience, and (3) expert evaluation. In the first stage of evaluation, a Delphi process was conducted with 15 emergency medicine specialists (93.3% male; mean age 46.8 ± 4.7 years). In the second stage, 100 case scenarios were designed and assigned reference triage categories. These scenarios were then evaluated by four emergency medicine specialists using the DIMaCTA, START, and drone-based triage algorithm ARTS. The triage categories assigned to the cases by the experts were compared with the reference categories in terms of accuracy, overtriage and undertriage rates, sensitivity, and specificity. In addition, the reliability of the DIMaCTA algorithm was examined using interrater reliability and the Intraclass Correlation Coefficient (ICC). The undertriage rate for the three compared algorithms was found to be 5.3%. The accuracy rate of the DIMaCTA algorithm was determined as 86.8%, and the overtriage rate as 7.9%. These rates indicate that DIMaCTA performed better than the START (82.9%, 11.8%) and ARTS (78.9%, 15.8%) algorithms; and the Intraclass Correlation Coefficient (0.95) showed a high level of agreement among the experts.