Endocrine Abstracts

JOINT1482

Relevance: Diabetes Mellitus (DM) ranks among the most prevalent endocrine disorders, requiring accurate differential diagnosis to determine the optimal treatment strategy. Variability in clinical manifestations, the presence of atypical forms (LADA, MODY, pancreatogenic diabetes), and limited resources complicate this process. A critical aspect also involves the timely detection of complications, notably diabetic retinopathy (DR), a leading cause of blindness among DM patients. The adoption of digital technologies, including Telegram Bots, can significantly enhance the efficiency of diagnosis and patient monitoring.

Materials and Methods: The Telegram Bot "DiaPredict" was developed to implement an algorithm for differential diagnosis of DM types and complication risk assessment. This algorithm is based on an analysis of international guidelines (ADA, WHO, IDF) and clinical data. The Bot automates the collection of anamnestic information, evaluates laboratory indicators, analyzes DR risk factors, and generates personalized recommendations for further examinations.

Results: The functionalities of the "DiaPredict" Telegram Bot include:

- Interactive collection of anamnesis, symptoms, and laboratory indicators.

- Patient stratification based on phenotypic characteristics.

- Risk assessment for the development of DR based on a multifactorial analysis (glycemic control, duration of DM, comorbidities).

- Formation of personalized recommendations for additional examinations and ophthalmological monitoring. "DiaPredict" approbation demonstrated high accuracy in the differential diagnosis of DM types and stratification of DR risks. Using the Bot allowed:

- Minimization of cognitive biases in doctors when diagnosing;

- Automation of clinical data processing, reducing decision-making time;

- Increased awareness among primary care physicians regarding atypical forms of DM and DR;

- Optimization of referrals for ophthalmological examination.

Conclusions: The "DiaPredict" Telegram Bot is an effective tool for optimizing the diagnostic process in DM, and the timely detection of DR. Its implementation can enhance the quality of medical care, reduce the burden on doctors, and improve patient management outcomes. Further research is directed towards integrating the Bot with electronic medical systems and evaluating its long-term effectiveness.