Ask Dr. Wellis: How to Anticipate Challenges and Make AI Work for Physicians?

🔄 Reminder: Dr. Wellis is an AI-powered physician GPT I’ve developed to assist with planning and implementing tech and AI projects in healthcare organizations.

Amir asks: I designed a pilot, chose the department and the participants, but I want to prepare. From your experience what feedback can I expect and how can I prepare to deal with it?

Dr. Wellis answer: During an AI pilot program for clinical documentation, you can expect a mix of positive feedback, constructive criticism, and resistance. Preparing to address these responses is key to refining the tool and ensuring broader adoption. Here’s what to anticipate and how to deal with it:

Types of Feedback You Can Expect:

1. Positive Feedback: Successes and Time Savings

• What You’ll Hear:

  • “It’s great for transcribing straightforward notes.”

  • “I’m spending less time on documentation.”

    
    

• How to Respond:

  • Celebrate and share these wins! Highlight success stories in communications to build confidence among other departments.

  • Collect specific data points (e.g., “I save 10 minutes per patient”) to quantify the impact.

2. Constructive Criticism: Accuracy and Workflow Issues

• What You’ll Hear:

  • “The tool doesn’t always understand medical terms or accents.”

  • “It’s helpful, but it slows me down when I have to fix errors.”

    
    

• How to Respond:

  • Work with the vendor to fine-tune the AI’s accuracy, especially for specialty-specific terms or accents.

  • Consider implementing a feedback loop where physicians can flag common errors to improve the system.

  • Acknowledge their frustrations and provide short-term workarounds, like editable templates.

3. Resistance to Change: Workflow Integration Challenges

• What You’ll Hear:

  • “This doesn’t fit into my workflow.”

  • “Switching between this and the EHR feels clunky.”

    
    

• How to Respond:

  • Emphasize that the tool is being piloted to gather feedback and will be refined based on their needs.

  • Partner with the vendor to improve EHR integration and streamline workflows.

  • Offer additional training to help physicians adapt to new processes.

4. Concerns About Privacy and Security

• What You’ll Hear:

  • “How do we know patient data is secure?”

  • “What happens if there’s a breach?”

    
    

• How to Respond:

  • Reassure them that the tool complies with HIPAA and other privacy regulations.

  • Share details about the vendor’s security measures, such as encryption and access controls.

  • Include your hospital’s IT and compliance teams in discussions to address any lingering concerns.

How AI Can Reduce Physicians’ Documentation Burden

5. Specialty-Specific Feedback

• What You’ll Hear:

  • “This template doesn’t work for my specialty.”

  • “The tool doesn’t capture all the details I need for my patients.”

    
    

• How to Respond:

  • Work with the vendor to customize templates for different specialties.

  • Involve representatives from each specialty to co-design templates and workflows.

6. Requests for Enhancements

• What You’ll Hear:

  • “Can it summarize lab results or previous visit notes too?”

  • “It would be better if it could predict what I’m going to say.”

    
    

• How to Respond:

  • Collect these enhancement requests for discussion with the vendor.

  • Focus on quick wins that can be implemented during the pilot phase.

  • Share the roadmap for planned improvements to manage expectations.

How to Prepare for Feedback:

1. Set Expectations Early

   • Let physicians know the pilot is a testing phase and their feedback will directly shape the final implementation.

   • Be transparent about limitations the tool may have during early use.

     
     

2. Establish Clear Feedback Channels

   • Use surveys, focus groups, and one-on-one check-ins to gather structured feedback.

   • Create an easy way for physicians to report issues in real time (e.g., a dedicated email or app feature).

     
     

3. Engage the Vendor Actively

   • Work closely with the vendor to address feedback promptly. Their responsiveness will build confidence among physicians.

     
     

4. Appoint Superusers

   • Identify tech-savvy physicians who can act as champions for the tool, assist peers, and mediate between users and the IT team.

     
     

5. Communicate Progress

   • Regularly update physicians on changes made in response to their feedback. This reinforces their sense of ownership in the process.

By anticipating these types of feedback and being proactive in addressing them, you’ll show physicians that their input matters and build trust in the AI solution.