Hallym University Hangang Sacred Heart Hospital

Hallym University Dongtan Sacred Heart Hospital has recently adopted the voice-to-text AI solution in EMR. The new solution enables medical staff to reduce the overall time spent in medical record by up to 75 percent.

The system powered by AI automatically converts the voice into text through the bluetooth microphone that is connected to the application. It writes the operation record and progress note through the voice. The AI recognize whether it's English or Korean, and then it makes documentation that fits the context, even in the condition where two languages are used simultaneously. It also perceives the frequently used terminologies in medical field, including the abbreviations.

As for the ward round documentation, it can record the consultation of a doctor, and make documentation afterwards using the recorded voice. The document is then transmitted to the EMR system and added to the existing patient data.

In order to make the system work, Dongtan Hospital prepared for about four months and selected six doctors of three different departments. Then the hospital recorded the doctors’ 12,000 example sentences and taught the AI to recognize the recorded voice.

After a month of use, the recognition rate reached about 90 percent. The documentation time spent for medical record has decreased by a fourth of what it used to be. That means it enabled each doctor to save 25 minutes a day, which is 500 minutes saved per month.

"Thanks to the introduction of voice recognition EMR system powered by AI, we could have more time to see more patients and conduct research, and that is a dramatic increase in efficiency," said Prof. Dong Woo Shin of General Surgery. "It will certainly enhance the quality of care as well as patent satisfaction," he added.

By Chul Kwon, Int’l Cooperation Team, HUMC (chris@hallym.or.kr)