Automatic documentation of professional health interactions: A systematic review

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Electronic health records are prevalent but can detract from the doctor-patient dynamic.
• Digital scribes aim to automate clinical documentation, freeing physicians to focus on patient interaction.

Approach:

• A systematic literature review examined intelligent solutions for automatic speech recognition (ASR) with automatic documentation during medical interviews.
• Included original research on systems simultaneously transcribing and structuring speech during patient-physician interactions, excluding speech-to-text-only technologies.

Key Points:

• Eight articles met inclusion criteria, detailing ASR systems with natural language processing, medical lexicons, and structured text output.
• No commercially available products were identified; real-world experience and prospective validation in large-scale clinical studies are lacking.

Conclusions:

• Automatic speech recognition shows potential for faster, more reliable medical registration, improving transparency, accuracy, and empathy.
• Significant gaps exist in clinical data regarding the usability and benefits of these digital scribe applications, necessitating future research.