A new scoring system to predict fatal accidents in General Aviation and to facilitate emergency control centre response

• Published in: Scientific reports Vol. 14; no. 1; pp. 27969 - 10
• Main Authors: Hinkelbein, Jochen, Hippler, Catherina, Liebold, Felix, Schmitz, Jan, Rothschild, Markus, Schick, Volker
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 14-11-2024; Nature Publishing Group; Nature Portfolio
• Subjects: 692/499; 692/700; Accident investigations; Accidents; Aircraft; Artificial intelligence; Emergency preparedness; Flight accidents; General Aviation; Humanities and Social Sciences; Landing behavior; multidisciplinary; Regression analysis; Science; Science (multidisciplinary); Scoring system; Survival; Flight accidents; Scoring system; General Aviation
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-024-77994-3

Numerous accidents occur with General Aviation aircraft every year. To date, pre-emptive prediction of survival or death is impossible. The current study aims to identify significant factors elementary to predict survival after General Aviation (GA) aircraft accidents. The Implementation of a scoring system, including these factors, may facilitate emergency control centre response. Data of flight accidents over a 20-year period (extracted from the German Federal Bureau of Aircraft Accident Investigation [BFU]) was analysed for fixed-wing motorized small aircrafts below 5,700 kg MTOW. Factors of interest were analysed using Chi
2
- and Mann–Whitney-U-Tests. Logistic regression was used to establish a score to calculate the probability of a fatal outcome after an aircraft accident. The BFU lists 1,595 GA aircraft accidents between 2000 and 2019. The factors “third quarter of the year” (p = 0.04), “last quarter of the year” (p = 0.002), “fire” (p < 0.0001), “distance from airport > 10 km” (p < 0.0001), “landing” (p < 0.0001) and “cruise” (p < 0.0001), significantly correlated positively or negatively with a fatal outcome. “Take-off”, “approach”, “month”, “day of the week”, “persons on board above three”, “night-time” and “icing conditions” showed no significant correlation. Using logistic regression “third quarter of the year” and “cruise” were excluded when using the B-STEP method. Including the four significant parameters, the score showed a strong effect with f
2
 = 0.709. The analysis of GA aircraft accidents in Germany enabled the identification of relevant factors and establishment of a new scoring system for survival prediction after small aircrafts accidents below 5,700 kg MTOW. The implementation of the scoring system in emergency control centres in the context of digital development and artificial intelligence can improve emergency response planning and distribution.