Abstract
Introduction. Cataract - pathological condition of the organ of vision, associated with eye lens opacity. There are currently around 37 million blind and 124 million visually impaired people worldwide. 47% of them - patients with cataract. From the point of view of economic feasibility statistically substantiated technique is an important tool for a medical institution to predict operative complications, which allows to significantly reduce their number. Purpose. The research task was to assess type, number and frequency of operative complications of surgical treatment of cataracts, to develop and describe statistically based technique for predicting operative complications. Material and methods. The study material was a structured database of the clinic medical control department, containing information on all treatment cases of patients with lens pathology for 5 years. The technique was developed using classification tree method. The inputs to the model were factors, significantly associated with the frequency of operative complications. The qualitative predictor was the patient’s diagnosis, the quantitative predictors were the patient’s age and the number of surgeries performed by the surgeon for cataracts during the last 5 years. Model output - the presence or absence of operative complications. When constructing the model, the branching type “Complete enumeration for one-dimensional branches using C&RT method” was used; goodness measure is Gini coefficient. The stop parameter is direct stop (FACT) with fraction of unclassified objects of 5%. Results and discussion. Among the predictors, patient’s age and number of cataract operations previously performed by the surgeon had greater impact on the outcome. The patient’s diagnosis was less important. Though a bit specific, the model is still applicable, because rate of complications is significant in this case. Conclusion. The technique for predicting operative complications depending on controllable factors helps to significantly reduce the frequency of complications, which is relevant for medical organizations.
