ECONOMIC EVALUATION AND DIAGNOSTIC ACCURACY OF COMPUTERIZED DETECTION ON LUNG RADIOGRAPHS IN CORONAVIRUS INFECTION (COVID-19) COMPLICATED BY PNEUMONIA

To strengthen COVID-19 containment strategies, an addition to the polymerase chain reaction, recognized as the gold standard for diagnosis, but limited by high cost and complex operation, is needed. Chest X-ray is widely used for screening due to its accessibility, but interpretation difficulties due to weak radiological features limit its accuracy. Artificial intelligence offers a solution for automated, accurate and fast X-ray analysis. The economic and diagnostic aspects of using artificial intelligence to detect pneumonia in COVID-19 are being evaluated. The study was conducted in X-ray rooms using specially designed X-ray description templates. The study compared the accuracy of the diagnosis of pneumonia in COVID-19 among three radiologists and artificial intelligence: doctors showed an accuracy of 92-94%, artificial intelligence – 91%. Artificial intelligence accelerates the process by analyzing the image in 7-12 seconds, whereas doctors' time ranged from 302 to 840 seconds (median 452 seconds). A survey of patients showed that 42% support the use of artificial intelligence, 30% disagree, and 28% found it difficult to answer. In a study with 167 participants, the use of artificial intelligence reduced diagnostic costs from 48,096 to 11,506 soms, providing savings of 36,589 soms. The technology improves diagnostic accuracy, reduces staff workload, and improves access to medical care. The successful implementation of artificial intelligence requires trained specialists, affordable equipment, information campaigns, and building trust among patients and medical organizations.