Back to all Webinars
Safe Foundations for a GPT-Powered Digital Front Door

Safe Foundations for a GPT-Powered Digital Front Door

Watch the full recording of Hyro’s July 2023 fireside chat on the responsible application of generative AI and large language models for the digital front door. Our panel of experts included URMC’s Michael Hasselberg, Ballad Health’s Taylor Hamilton, and Microsoft’s Timothy Gray.

Watch the full webinar to gain insight into:

 

  • How leading health systems are currently leveraging generative AI / LLMs such as GPT
  • Where LLMs such as GPT can transform the digital front door and enrich patient-facing use cases
  • What crucial safeguards are required to ensure the security and reliability of LLM-powered tools

Featuring:

 

  • Michael Hasselberg, Chief Digital Health Officer, University of Rochester Medical Center
  • Taylor Hamilton, Senior Vice President & Chief Consumer Officer, Ballad Health
  • Timothy Gray, Chief Health Innovation Officer, Microsoft
  • Moderated by Aaron Bours, VP Marketing, Hyro

Webinar Highlights

According to a Hyro survey of 200 patient access leaders, 46% of US health systems plan to deploy GPT-powered solutions in the next 12 months. With 47% of healthcare workers expected to exit the industry by 2025, GPT’s potential to streamline administrative tasks and enhance patient experience presents a compelling proposition for healthcare leaders struggling to bridge worsening workforce gaps.

 

While this is an encouraging sign for a sector once considered slow to adopt new technologies, health IT leaders are now voicing their concerns over the security, explainability and compliance surrounding large language models (LLMs), especially in the context of patient-facing communications.

Use Cases and Implementation of GPT in Healthcare

The panelists discussed GPT’s potential in healthcare, including automating administrative tasks like appointment scheduling, bill payments, and directions.

 

Tim Gray highlighted GPT’s benefits in healthcare call centers, such as managing call summaries and reducing wait times.

 

Michael Hasselberg shared the University of Rochester Medical Center’s innovative AI applications, including GPT. Overall, GPT can enhance healthcare efficiency by automating tasks, reducing administrative burdens, and improving response times for better patient experiences and streamlined operations.

AI in Healthcare: Past Pessimism to Present Possibilities

Responsible adoption of GPT and LLMs was emphasized for patient safety and data security. Michael discussed the challenge of triaging patient messages in healthcare, a common issue for many health systems, and the past difficulties in creating accurate NLP models for this purpose.

 

Michael also expressed skepticism about AI vendors claiming healthcare problem-solving and stressed the importance of assessing data suitability, emphasizing data structure’s role in AI success. He then highlighted their recent success with GPT4 in sorting patient messages accurately after fine-tuning the model using patient data.

 

This success story underscored AI’s immense potential in healthcare, particularly in areas like ambient documentation, speech-to-text, revenue cycle automation, and prior authorizations. Ballad Health is in the early stages of adopting AI technologies, aiming to enhance various healthcare aspects.

 

Tim Gray from Microsoft discussed how AI streamlines knowledge workers’ tasks, such as email summarization and presentation creation, improving efficiency. The panelists agreed on the profound impact of scaling AI across organizations, both in healthcare and beyond, simplifying complex tasks and enhancing productivity.

Challenges and Safeguards

The panel highlighted the significance of aligning GPT-powered AI tools with healthcare’s unique challenges. Patient expectations were a central theme, with a growing trend towards patients seeking quick, convenient access to healthcare services, similar to digital experiences in other industries.

 

Tim Gray stressed the need to address patient privacy concerns, emphasizing healthcare organizations’ responsibility to protect patient data and maintain trust. Taylor Hamilton shared Ballad Health’s focus on transparency and communication to meet patient expectations while ensuring privacy.

 

Michael Hasselberg discussed AI’s potential in improving health equity through language translation but cautioned against overreliance on AI without human oversight. The panel emphasized the importance of safeguards, continuous monitoring, and ethical considerations to maintain AI trustworthiness.

 

They also discussed the need for specific regulations governing AI in healthcare, ensuring responsible and ethical implementation. Transparency emerged as crucial, with healthcare organizations disclosing AI use and limitations to patients. Collaboration with reputable technology partners was highlighted, especially for smaller healthcare organizations looking to implement secure and reliable AI solutions.

Patient Expectations for Digital Healthcare

The experts recognized the growing trend of consumers seeking digital and tech-driven services in various industries, with expectations for AI-driven healthcare solutions likely to increase as patients become more familiar with AI technologies.

Implementing AI in healthcare may face resistance from specific departments or individuals who are cautious about potential risks. Effective communication, transparency, and education are emphasized as key strategies to gain support for AI initiatives, as highlighted by Taylor Hamilton.

 

Healthcare organizations must engage patients across different generations, each with varying levels of comfort with digital tools. A mix of services should be offered to cater to diverse patient needs.

 

Balancing patient privacy and efficiency is crucial when implementing AI technologies. AI can streamline processes, improve access to care, and provide faster responses to patient inquiries while ensuring data protection.

 

Transparency is a fundamental aspect of AI implementation in healthcare. Patients should be well-informed about AI usage, limitations, and data handling to build trust and manage expectations.

 

Addressing security concerns, the experts recommend educating employees about AI-related risks and the importance of protecting patient data with simple measures like informative letters to raise awareness among staff.

Start your 30-day trial with Spot™, Hyro’s GPT-powered site search solution for health systems.

Watch more webinars

Watch more webinars
See all Webinars
See all Webinars

Unlock your digital potential with the #1
adaptive communications platform.

Book a Demo