AI Center of the China Medical University Hospital


Nursing voice assistant system helps nurses reduce tedious work


The shortage of nursing staff and the high turnover rate are issues that the current medical system attaches great importance to. Nursing staff have a large workload and a high turnover rate. Whether they are senior or new nurses, they spend the most time at work for handovers and writing nursing records。


Nursing Voice Assistant System Development Goal


 1.Simplify the work flow of nurses


Assist nursing staff to improve the problem of handover and nursing record writing, so that clinical nurses can record the required information completely, reduce the requiring time to write nursing records, and simplify the work process. This allows nursing staff to quickly integrate important shift information, and reduce the pressure on nursing staff.


2.Let the voice recognition system truly be implemented in the medical field


Due to nursing-related records, there are often problems with medical terms and mixed Chinese and English, as well as those meaningless auxiliary words, such as spontaneous speech, such as pauses, redundant words, repetitions... etc. Although there are many voice recognition software in the market , the accuracy rate is lower than that of ordinary daily language when applied to voice recognition in nursing care. This will cause difficulty in identification and cannot be easily judged correctly.


3.Continue to increase and optimize the corpus to improve the accuracy of speech recognition


The original corpus included nearly 2,200 general voice data in Chinese and English. With the assistance of various nursing units in the hospital, 3,300 actual nursing handover corpora were added. The current version has reached 80% word accuracy with free dictation whereas 90% with standard texts. It is predicted that the transferring time of each patient will be saved 3 to 5 minutes. In the future, we will continue to invest in nursing handover corpus to further improve the accuracy of voice recognition, and create a professional nursing handover speech recognition system.






 Booth No. M804