Prediction of the progression of endometrial hyperplasia in women of premenopausal and menopausal age based on an analysis of clinical and anamnestic indicators using multiparametric neural network clustering

Background
A number of studies are aimed at solving the problems of implementing innovative medical information technologies, but the is-sues of family medicine informatisation have not been fully resolved. This is especially important to optimise the diagnosis of the most common diseases.

Objectives
The aim of our study was to develop a methodology for predicting the progression of endometrial hyperplasia at the primary healthcare level based on an analysis of clinical and anamnestic indicators using multiparameter neural network clustering.

Material and methods
Clinical and anamnestic data was obtained based on the results of a retrospective analysis of the inpatient charts of 52 patients with non-atypical endometrial hyperplasia. For a deeper analysis and clustering, the neural network approach was used with the NeuroXL Classifier add-in application for Microsoft Excel.

Results
The results of the cluster analysis showed that when predicting the progression of endometrial hyperplasia based on an analysis of clinical and anamnestic indicators, it is important to take into account the combination of the use of intrauterine contraception, as well as infertility, obesity and dis-eases of the gastrointestinal tract in patients. At the same time, the probability of progression of endometrial hyperplasia also increases with an increase in the number of operative obstetric and gynaecological interventions.

Conclusions
In order to effectively and objectively assign patients to the risk group for the progression of endometrial hyperplasia according to the indica-tors obtained during observation, neural network clustering was used, which allows one to determine the value of combined changes of certain parameters for the prognosis of the progression of the disease.