Possibilities for forecasting and preventing early postoperative complications

Objective. The article provides information on how to improve the forecast of the early postoperative period by additional individualization of anesthetic management of patients during emergency surgical interventions on the gallbladder using artificial neural network technologies. 

Materials and methods. The course of combined anesthesia and the features of the postoperative period were analyzed in 92 patients with an endoscopic cholecystectomy performed for urgent indications. The prediction of the variant of the postoperative stage of hospitalization was realized using the analysis of the significance of 20 different-modal variables selected for the description of patients using fuzzy logic technologies. The possibility of changing the forecast to a more favorable one was achieved on the basis of the developed algorithm for evaluating the results of training neural networks on the Neuro Pro 0.2 neuroimitator. 

Results. According to the generally accepted criteria, all patients had endoscopic cholecystectomy and anesthesia wit out complications. At the postoperative stage, 2 groups of persons were identified - with the expected short hospitalization (72 cases - 6.7±2.1 days) and with the clinic, which led to its reliable prolongation (20 cases - 12.2±3.5 days). It has been shown that the use of a neural network approach makes it possible with a confidence of more than 80% to assume cases with a high probability of postoperative disorders and in half of such patients to improve the prognosis within the framework of neural network technology and the developed algorithm for selecting the severity of the selected 5 variable factors related to the method of conducting anesthesia. 

Conclusion. Neural network technology makes it possible to predict cases with individual “unpredictable” responses to surgical trauma. Assessing the significance of the factors used and varying their severity create the basis for the individualization of anesthetic management of patients, prevention of postoperative reactions and a reduction in the period of hospitalization.

1. Воскресенский О.В., Смоляр А.Н., Дамиров М.М., Александрович Ю.С., Гордеев В.И. Оценочные и прогностические шкалы в медицине критических состояний. Санкт-Петербург: Сотис, 2007. 140 c.

2. Басова Л.А., Карякина О.Е., Мартынова Н.А., Кочорова Л.В. Прогнозирование послеоперационных осложнений на основе нейросетевых технологий. Вестник новых медицинских технологий. 2015;22(4):117-121. DOI: 10.12737/17035

3. Вараксин А.Н. Статистические модели регрессионного типа в экологии и медицине. Екатеринбург: Гощицкий, 2006. 255 с.

4. Иванов Н.В. Нейронные сети в медицине. Сложные системы. 2018;4(29):46-70

5. Койчубеков Б.К., Сорокина М.А., Мхитарян К.Э. Математические методы прогнозирования в медицине. Успехи современного естествознания. 2014;4:29-36. URL: http://www.natural-sciences.ru/ru/article/view?id=33316

6. Лебедев Н.В. Система объективной оценки состояния больных и пострадавших. Москва: БИНОМ, 2015. 160 с.

7. Милова К.А. Интеллектуальная система прогнозирования развития осложнений у хирургических больных. Нейрокомпьютеры: разработка, применение. 2010;11:20-22

8. Точило С.А. Предикторы развития синдрома полиорганной дисфункции у пациентов после абдоминальных хирургических вмешательств. Новости хирургии. 2017;25(5):494-502. DOI: 10.18484/2305-0047.2017.5.494

9. Basit A., Sarim M., Raffat K., Kamran A., Nadeem A., Muhammad S. Artificial Neural Network: A Tool for Diagnosing Osteoporosis. Res J Recent Sci. 2014;3(2):87-91.

10. Brooks G.P., Barcikowski R.S. The PEAR method for sample sizes in multiple linear regression. Multiple Linear Regression Viewpoints. 2012;38(2):1-16.

11. Díaz-Hierro J., Martín Martín J.J., Vilches Arenas A., López del Amo González M.P., Patón Arévalo J.M., Varo González C. Evaluation of time-series models for forecasting demand for emergency health care services. Emergencias. 2012;24(3):181-188.

12. Moriña D., Puig P., Ríos J., Vilella A., Trilla A. A statistical model for hospital admissions caused by seasonal diseases. Stat Med. 2011;30(26):3125-3136. DOI: 10.1002/sim.4336

13. Naqvi I.H., Mahmood K., Ziaullaha S., Kashif S.M., Sharif A. Better prognostic marker in ICU - APACHE II, SOFA or SAP II!. Pak J Med Sci. 2016;32(5): 1146-1151. DOI: 10.12669/pjms.325.10080

14. Siettos C.I., Russo L. Mathematical modeling of infectious disease dynamics. Virulence. 2013;4(4):295-306. DOI: 10.4161/viru.24041

15. Unkel S., Farrington C., Garthwaite P., Robertson C., Andrews N. Statistical methods for the prospective detection of infectious disease outbreaks: A review. Journal of the Royal Statistical Society. Series A (Statistics in Society). 2012;175(1):49-82. URL: http://www.jstor.org/stable/41409708

16. Wick J.P., Turin T.C., Faris P.D., MacRae J.M., Weaver R.G., Tonelli M., Manns B.J., Hemmelgarn B.R. A Clinical Risk Prediction Tool for 6-Month Mortality After Dialysis Initiation Among Older Adults. Am J Kidney Dis. 2017;69(5):568-575. DOI: 10.1053/j.ajkd.2016.08.035