The University of Murcia together with the Murcian company Vocali Intelligent Systems has developed a new tool that assists doctors during consultations, with which physicians can provide more personal and exhaustive care, by not having to pay attention to the manual collection of data on the ailment.
This computer application, which is based on a deep learning system, encourages conversation and humane treatment during the consultation, with the aim that medical assistance becomes a closer and more satisfactory experience.
What does this application offer to improve doctor-patient interaction?
“The new platform is capable of identifying different speakers and recording the audio independently. This is possible thanks to a custom device developed within the framework of the project, which is based on a microphone array– a device similar to a smart speaker – and a intelligent software that enables audio transcription and clinical knowledge extraction”, points out Rafael Valencia, one of the UMU researchers responsible for this work.
Transcripts collected by this system are further processed using knowledge extraction and natural language processing technologies., achieving in this way capture and structure different knowledge useful for the doctor such as symptoms, diseases, previous episodes, medications to be prescribed or diagnostic tests to be performed. “Thus, the doctor can operate on the patient during the consultation and semi-automatically write a report with the patient’s visit and it is reliably functional to be able to prescribe medications or generate flyers for diagnostic tests”, adds Professor Valencia.
On behalf of the University of Murcia, in addition to Rafael Valencia, professors Francisco García Sánchez and José Antonio García Díaz, belonging to the Knowledge Modeling, Processing and Management Technologies (TECNOMOD) research group, have participated in this project.
These researchers have developed the intelligent extraction module of medical knowledge that allows extracting clinical information in natural language text and they have carried out different prototypes based on the recognition of named entities and semantic annotation systems from domain ontologies that allow the identification of these concepts and the relationship between them.
All the prototypes have been developed autonomously and independently, so the company can apply these models to the solution of other problems, by adapting them to programs that can be applied in real time.
Vócali is a technology-based company that develops natural language processing and speech recognition solutions. Currently, it focuses its activity in the field of health and medicine and has different solutions for issuing clinical reports in Spanish, such as the INVOX Medical platform, which is implemented in many hospitals throughout Spain, Portugal and Latin America. .
This UMU-Vócali collaboration is part of the project entitled ‘System based on deep learning for the transcription and extraction of knowledge in doctor-patient interviews’ with reference IDI-20200407, which has been financed by the Center for Technological and Industrial Development, agency under the Ministry of Science and Innovation.