New Voice2Med product release: surgeons and cardiologists can interact with EHRs through voice

Speech Technology Center Group, a Sber ecosystem resident, has presented a new product release of Voice2Med, an AI-powered voice recognition solution for data entry in Electronic Health Records.

Voice2Med allows users to fill in EHRs in real time, converting a doctor's voice into text. Two new dictionaries of medical terms have been added to the previously available dictionaries for general practitioners, pediatricians, CT, MRI, PET (positron emission tomography) radiologists, pathomorphologists, ultrasound doctors and endoscopists: a surgeon's dictionary for filling out patient diaries and surgery protocols, as well as for use in emergency rooms, and a cardiologist's dictionary for use during cardiac examinations. Over 500,000 impersonal protocols have been processed to create the new dictionaries, which resulted in the term recognition accuracy of more than 93%.

Voice2Med now also supports dictionaries of medical terms for surgeons and cardiologists, which allows MDs to record details using voice, without being distracted from a surgery or examination, and to devote more time to the patient, rather than routine paperwork. Furthermore, to simplify the implementation process, we have designed packaged solutions for outpatient clinics and hospitals: Voice2Med.Hospital and Voice2Med.Polyclinic. Apart from software and training webinars, the workstation includes a professional headset: this comprehensive solution removes the need to pick an optimal headset - now everything is available in a single solution.

Dmitry Dyrmovsky

CEO, STC Group

Radiologists already use Voice2Med successfully in 80 healthcare facilities in Moscow, 35 in Mordovia, as well as in Udmurtia, Murmansk, Kemerovo, and Volgograd. Voice2Med keeps progressing. During the examination, the doctor dictates data using a special microphone. The data are then decrypted and sent to the open protocol of the medical information system. The high quality of speech recognition results in a fully completed EHR that only requires minor adjustments. The Diagnostic and Telemedicine Center has recorded a 22% reduction in the time MDs spend on EHRs in the medical facilities of the Moscow Healthcare Department: before the adoption of voice input, it used to take 5-6 minutes to fill out EHRs, after the adoption it takes 3-4 minutes.

Continuous work is being put into broadening the list of Russian operating systems that Voice2Med supports. Currently it supports various versions of Astra Linux and Alt.