Nowadays, administrative tasks can consume up to 30% of clinicians' working hours, diverting attention from patient care and contributing to physician burnout. The MediSpeech project aims to reduce administrative waste in healthcare by creating an open digital healthcare ecosystem for automated medical reporting. The proposed solution uses AI-powered speech recognition, data interoperability and harmonisation, and technology-advanced clinical decision support to transform the healthcare model into a patient/doctor-centered approach.
The MediSpeech project emerges as a transformative solution to this pervasive challenge. At its heart, MediSpeech leverages the burgeoning field of Automatic Speech Recognition (ASR) technology, particularly tailored for the healthcare sector. This technology promises to revolutionize patient-healthcare provider interactions by accurately capturing and processing clinical conversations. By utilizing clinical LLM’s to understand and document the nuances of clinical speech—including medical terminologies such as MesH, Snomed, Loinc, ICD—it aims to automate the generation of clinical reports. This not only streamlines administrative workflows but also enables healthcare professionals to refocus on patient care, thereby enhancing the quality and timeliness of healthcare delivery.
The project proposes a multi-faceted digital IT ecosystem that not only prioritizes the automation of medical reporting but also champions a shift towards a patient/doctor-centred healthcare model. This is achieved through the integration of several state-of-the-art (SotA) technologies, including AI-powered speech recognition for enhanced accuracy, data interoperability for seamless service integration, advanced clinical decision support systems for improved patient outcomes, and comprehensive medical reporting for data-driven decision-making.
Contact:
Khiet Truong k.p.truong@utwente.nl,
Roeland Ordelman roeland.ordelman@utwente.nl
Website: https://itea4.org/project/medispeech.html
LinkedIn: https://www.linkedin.com/company/medispeech/