Voice Based Prescription Generation Using Natural Language Processing

Abstract

Abstract— Since prescriptions are still written by hand in hospitals, poor handwriting could result in serious issues like patients taking the incorrect medication or insufficient amounts of it. This scenario has the potential to significantly affect the patient's overall health status or lead to mortality. To address this problem, a voice-based prescription generation system was developed, which converts spoken prescriptions into written ones using the Web Speech API. In order to extract medical entities from text, Name Entity Recognition (NER) was used to obtain this text transcript. The digital prescription will be created in this manner. The NER task is carried out using en_core_med7_lg model. Next JS, MongoDB was used to build this system.