Clinical information processing block of the biotechnological system for latent Diabetes mellitus type 2 detecting, based on the diagnosis of an experienced endocrinologist

摘要

In the article it is proposed firstly a block of a biotechnological system (BTS) for processing clinical information for the early diagnostics of latent Diabetes mellitus type 2 (DM2). It was based on mathematical models of the intellectual activity of an expert-endocrinologist in identifying latent DM2 (Prediabetes), based on the glycemic data of the oral glucose tolerance test (OGTT) and on examples of these data with expert diagnoses. At present this expert method is recognized as the most trustworthy and precision, but unsuitable for mass screening of the population for latent DM2, which is widespread and dangerous due to its late complications. It was analyzed the causes of the failure of our predecessors' attempts to introduce objective methods for diagnosing Prediabetes by non-intersecting intervals of glycemic level (glucose concentration in the blood) in the OGTT clinical data, as well as by non-intersecting intervals of values of various parameters, supposed to be diagnostic. It is shown that proposed by the authors of the article the mathematical models of the intellectual activity of the expert- endocrinologist in the form of an artificial neural network (ANN) and in the form of a logical formula of the propositional algebra allow the diagnosis of the Prediabetes to be carried out at an equally high level. The clarified diagnostic capabilities of these models allow their separate and combined use as part of the corresponding BTS in mass screening investigations of the population of Prediabetes.