Мать и дитя в Кузбассе (May 2018)
FORECASTING OF SURVIVAL OF CHILDREN WITH THE PRENATALLY DIAGNOSED PATHOLOGY OF THE CARDIOVASCULAR SYSTEM
Abstract
The development of effective methods for the analysis and prognosis of the survival of newborns with prenatally diagnosed congenital malformations of the cardiovascular system are the urgent task of modern medicine. Objective – a neural network model for predicting the survival of children with prenatally diagnosed congenital malformations of the cardiovascular system was developed. Materials and methods. To create the artificial neural networks, the method of constructing multifactor mathematical prediction models in the software package Statistica 6.0 was used. The significance level of the factors influencing the survival of children with prenatally diagnosed congenital malformations of the cardiovascular system was determined using Wald statistics. When checking statistical hypotheses, the critical level of significance was assumed to be 0,05. Results. A neural network model for the determination of the probability of survival of a child with prenatally diagnosed congenital malformations of the cardiovascular system, which has a high prognostic ability of 0,88, sensitivity of the model was 77,6 %, specificity 86,4 %. The value of prognostic survival probability is in the range from 0 to 100 %. With an indicator value of more than 80 %, the probability of survival of a child with prenatally diagnosed congenital malformations of the cardiovascular system is estimated as high, ranging from 20 % to 80 % – as an average and less than 20 % – as low. Conclusion. In the algorithm for predicting the survival of children with prenatally diagnosed congenital malformations of the cardiovascular system it is necessary to include a combination with other pathology of cardiovascular system, with other organs and systems, with chromosomal abnormalities, with microdeletion and monogenic syndromes.
