The Effectiveness Evaluation of Artificial Intelligence Assisted Diagnosis System for Chest Diseases in the Diagnosis of General Practitioners in Primary Healthcare Institutions
摘要
Background In primary healthcare institutions, due to insufficient staffing of radiologists, it is impossible to ensure that they are on duty 24/7, and general practitioners have limited interpretation ability of chest images, which affects the quality and efficiency of medical services. Objective To explore the application effect of artificial intelligence assisted diagnosis system for chest diseases in primary healthcare institutions, with a focus on evaluating its role in improving the diagnostic efficiency, accuracy, and patient satisfaction of general practitioners. Methods During the vacation period of radiologists in April 2024, 16 general practitioners from Donghua Community Center were selected as the research subjects. They were randomly divided into a general practitioner+AI group of 8 and a control group of 8. At the same time, a total of 100 respiratory system disease patients who required X-ray imaging examination were included in the two groups of doctors, with 50 patients in each group. The general practice+AI group uses the assisted diagnosis system for chest diseases (ADSC) to identify X-ray images and assist general practitioners in making disease diagnoses. The control group receives diagnosis and treatment according to routine procedures, records the time spent by the two groups of doctors for comparison, and uploads the X-ray images of the two groups of patients to the regional imaging center for review and evaluation of the diagnostic accuracy of the two groups of doctors; Conduct a satisfaction survey on patients and evaluate the ADSC usage experience of two groups of doctors. Results The gender, age, education, and professional title of the two groups of doctors were compared, and there was no statistically significant difference (P>0.05) . The gender and age of the two groups of patients were compared, and the difference was not statistically significant (P>0.05) . The time required for diagnosis in the general practice+AI group was s