Technologies to reduce EHR usability burden

Technology innovations have long promised to relieve burden by freeing physicians from manual EHR tasks.

Today there is clear evidence that technologies and technology-enabled services can relieve some EHR usability burden by inserting technology or technology- assisted services into the EHR workflow and eliminating or streamlining tasks.

Diagram showing tech tools plotted across two axes: scalability/affordability, and tech level (human based to AI based). Transcription services and scribes are lower tech, human based, high cost and not scalable. Virtual scribes and ambient systems use advanced voice AI and are high cost and not scalable. Dictation systems and AI assistants are scalable, affordable, and use advanced voice AI.

Medical scribes

$2,500 to $4,500 per month

Medical scribes are professionals who record information during clinical visits in real time and can perform other EHR tasks under physician supervision.

Medical scribe companies

  • Pro
    • Delegating documentation and other administrative tasks to an in-person scribe takes burden off the physician.
  • Cons
    • Not broadly used in family medicine or primary care due to costs and investment required
  • Impact
    • Increased productivity
    • 57% increase in patient face time
    • 27% decrease in EHR time

Virtual scribes

$1,000 to $1,200 per month

Virtual scribes reduce potential intrusiveness by removing the third party from the exam room and reduce costs by scaling and offshoring the scribe’s work.

Virtual scribe companies

  • Pro
    • Cheaper than medical scribes and no retention issues
  • Cons
    • Performs only patient charting
    • Not real-time; up to 24-hour delay
    • May incur added hardware costs
  • Impact
    • Up to 85% burnout reduction
    • 1.1 hour per day EHR time reduction
    • 1 hour per day documentation reduction

Medical speech recognition

Enterprise: $25 to $75 per month
Single physician: $200 per month

Medical speech recognition allows physicians to use their voices to dictate into their EHRs more easily. Their accuracy and integration have evolved to provide many physicians relief from documentation burden. However, physicians still must navigate the EHR to enter and edit information.

Medical speech recognition: Dragon Medical KLAS Research

  • Pro
    • Established technology option with greatly improved accuracy
  • Con
    • Similar to dictation but physician must navigate EHR and edit
  • Impact
    • 79% of users were satisfied with speech recognition
    • 77% reported that it improved efficiency

Ambient speech recognition

$1,800 per month

Ambient speech recognition technology “listens” to the caregivers’ and patient’s conversation with the physician throughout the encounter, including history and exam details and works with an external staffed service to create a visit note.

  • Pro
    • Shows promise in decreasing burden and enhancing care and visit documentation efficiency
  • Cons
    • Ambient discussion in exam room is personal
    • Medical decision-making is purposeful and may not be done in the exam room
  • Impact
    • 79% of users reported better documentation quality
    • 70% saw reduced burnout and fatigue
    • 81% of patients saw greater physician focus

AI assistant

$150 to $200 per month

AI-powered, voice-enabled digital assistants can help physicians complete documentation and other administrative tasks. Although AI assistants still have room to develop, they are the only documentation solutions on the market that show promise in totally automating the EHR documentation process.

  • Pros
    • Mobile assistant allows physician to step away from EHR
    • Consumer-friendly features and cost
    • Some AI assistants use a "human in the loop" to edit and correct the transcription while the AI assistant learns.
  • Con
    • EHR integration is essential and vendors have been slow to support, but progress is being made.
  • Impact
    • 72% reduction in documentation time
    • 40% decrease in after-hours work, including weekend work
    • 20% increase in practice satisfaction

Differences between AI and medical speech recognition

Unlike AI, legacy speech recognition systems did not leverage deep learning needed to perform voice-to-text conversion or have natural language understanding. 

Although newer speech recognition solutions have progressed in these areas, AI assistants better understand natural language and can detect intent (i.e., commands from the user). 

Whereas voice recognition solutions require the user to specify where dictation should be placed in the EHR note, AI assistants’ built-in model of documentation recognizes embedded commands and knows where to place text. This ability enables the physician to generate a note without having to navigate or edit in the EHR.