Effective communication between healthcare professionals and patients is fundamental to quality medical care. However, barriers can obstruct this communication, and lead to decreased patient satisfaction, as well as clinical burnout. Patient examinations only comprise 52.9% of the time spent by professionals, and 37% of the time is spent completing paperwork. This inefficient use of time contributes to clinician burnout, poor patient satisfaction, and ultimately diminished quality of care. The proposed solution is the AI-powered device (DigiQuill), to seamlessly digitize handwritten notes and documents, transforming the way corporations and hospitals handle paperwork. Equipped with a microphone, the device captures and digitizes every stroke in real time, and the audio is sent to the cloud for processing. Additionally, OpenAI’s Whisper model is used to transcribe different languages; the initial languages used will be Hindi and English for proof of concept. An LLM (Phi3 3.8 billion) will then be used to extract insights from the transcriptions and then format it according to the patient data forms.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

DigiQuill: Smart Pen for Automated Healthcare Documentation

  • Abhinav Pathak,
  • Sarah Fernandes,
  • Alina Josephin Biju,
  • Jagadish Nayak

摘要

Effective communication between healthcare professionals and patients is fundamental to quality medical care. However, barriers can obstruct this communication, and lead to decreased patient satisfaction, as well as clinical burnout. Patient examinations only comprise 52.9% of the time spent by professionals, and 37% of the time is spent completing paperwork. This inefficient use of time contributes to clinician burnout, poor patient satisfaction, and ultimately diminished quality of care. The proposed solution is the AI-powered device (DigiQuill), to seamlessly digitize handwritten notes and documents, transforming the way corporations and hospitals handle paperwork. Equipped with a microphone, the device captures and digitizes every stroke in real time, and the audio is sent to the cloud for processing. Additionally, OpenAI’s Whisper model is used to transcribe different languages; the initial languages used will be Hindi and English for proof of concept. An LLM (Phi3 3.8 billion) will then be used to extract insights from the transcriptions and then format it according to the patient data forms.