AI assistant for doctors introduced in the Sverdlovsk region
Voice AI assistant for doctors introduced in the Sverdlovsk region
Voice AI assistant for doctors is being introduced in the Sverdlovsk region. The system automates routine processes: based on a dialogue with the patient, it forms records in the medical record, and also helps doctors quickly find answers to professional questions. This is reported by the Department of information policy of the Sverdlovsk region.
The pilot project was implemented in two large hospitals of Yekaterinburg - CGKB No. 6 and CGB No. 7. Within a month, the technology was tested by five doctors – therapists and cardiologists. During this time, the system helped to process 253 reception protocols and 106 medical requests. The results showed that the technology significantly reduces the burden on specialists: the volume of text in the protocols increased by 15%, while the documents themselves became fuller – they did not use abbreviations.
The function of the assistant includes not only documentation, but also reference support. Most often, doctors asked for information about drugs - this topic accounted for half of all applications. Also, doctors were interested in information about diseases, formulations of diagnoses, general clinical recommendations, standards and parameters of diagnostic methods. On average, 1.5 requests were required to receive a full response.
The project is planned to extend to more than a thousand doctors and paramedics, who conduct about 18 thousand appointments per shift. 981 rooms are equipped with microphones. According to estimates, this coverage will save up to 270 hours of working time daily – these hours doctors will be able to devote to direct work with patients.
Earlier, “MK-Ural” wrote that in the framework of the project “Polyclinic”. Reboot registrars in the Sverdlovsk region received audio badges. A small device is attached to the employee’s clothes and records the dialogue with the visitor. At the end of the workday, the file is automatically uploaded to the server, where artificial intelligence decrypts the speech, analyzes it and searches for forbidden phrases.
