Abstract: Objective: To explore the effectiveness and safety of an intelligent infusion system based on Convolutional Neural Network (CNN) for visual recognition in clinical practice, in order to provide references and insights for the intelligent management of infusion systems in general wards. Methods: The study selected 280 patients (140 cases each before and after the application of the intelligent infusion system) and 16 nursing staff from a general surgery ward of a tertiary grade A hospital in Hangzhou by convenience sampling method at October 2022 and March 2023 as the subjects of the study. Then the workload of nursing staff, patient satisfaction, and noise levels in the ward before and after the intervention were compared. Results: Due to the usage of the intelligent infusion system, nurses spent less time going back and forth between the nurse station and the wards (P<0.05); patient satisfaction with the intelligent infusion system increased (P<0.05); the frequency of nursing staff responding to calls decreased (P<0.05); noise levels in the ward except at nighttime were lower than before (P<0.05). Conclusion: The intelligent infusion system based on CNN for visual recognition can reduce the workload of nursing staff, improve patient satisfaction, and is of significant value in improving the quality of infusion nursing and the ward environment.

Key words: Convolutional Neural Network; visual recognition; intelligent infusion system; nursing quality