MATHEMATICAL MODEL OF FORECASTING FOR OUTCOMES IN VICTIMS OF METHANE-COAL MIXTURE EXPLOSION

BACKGROUND. The severity of the victims’ state  in the early period after the combined  trauma (with the prevalence of a thermal  injury) is associated with the development of numerous  changes  in all organs and systems  which make proper  diagnosis  of complications and estimation of lethal  outco...

Full description

Bibliographic Details
Main Authors: E. Y. Fistal, V. G. Guryanov, V. V. Soloshenko
Format: Article
Language:Russian
Published: Sklifosovsky Research Institute for Emergency Medicine, Public Healthcare Institution of Moscow Healthcare Department 2016-09-01
Series:Неотложная медицинская помощь
Subjects:
Online Access:https://www.jnmp.ru/jour/article/view/294
Description
Summary:BACKGROUND. The severity of the victims’ state  in the early period after the combined  trauma (with the prevalence of a thermal  injury) is associated with the development of numerous  changes  in all organs and systems  which make proper  diagnosis  of complications and estimation of lethal  outcome  probability extremely  difficult to be performed.MATERIAL AND METHODS. The article  presents a mathematical model  for predicting  lethal  outcomes  in victims of methanecoal mixture explosion, based on case histories of 220 miners who were treated at the Donetsk Burn Center in 1994–2012.RESULTS. It was revealed  that  the  probability  of lethal  outcomes  in victims of methane-coal mixture  explosion was statistically significantly affected  with the  area  of deep  burns  (p<0.001), and  the  severe traumatic brain injury (p<0.001). In the probability of lethal  outcomes,  tactics  of surgical treatment for burn wounds in the early hours after the injury was statistically significant (p=0.003). It involves the primary debridement of burn wounds in the period of burn shock with the simultaneous closure of affected  surfaces with temporary biological covering.CONCLUSION. These neural network models are easy to practice and may be created  for the most common pathologic conditions  frequently encountered in clinical practice.
ISSN:2223-9022
2541-8017