Dan Buckland, Medical Director of Duke University Health System and assistant professor at Duke University, discusses Duke University's unique approach to AI in healthcare, using tested and proven models. He emphasizes the importance of virtual care as a last resort and the role of technology in the patient experience. The podcast also explores generative AI for interoperability in healthcare IT systems and advancements in AI for patient communication.
Read more
AI Summary
Highlights
AI Chapters
Episode notes
auto_awesome
Podcast summary created with Snipd AI
Quick takeaways
Duke University prioritizes the use of mature AI models with requisite training to optimize patient care.
The successful implementation of AI in healthcare depends on the end user's understanding and effective use of the technology.
Virtual care is not always the best solution and should be used selectively based on the appropriateness of the situation.
Technology can cause friction in the patient experience, and human factors play a significant role in achieving efficient healthcare processes.
Generative AI systems like language models have the potential to improve interoperability in healthcare by translating specialized vocabulary.
Deep dives
Duke University's Approach to AI: Prioritizing Maturity and Expert Training
Duke University Health System takes a unique approach to AI technologies by using models that are at least two years old and have undergone requisite expert training. This strategy allows them to ensure that the subtleties and limitations of the technology have been worked out before applying it to patient care. Additionally, Duke emphasizes the importance of virtual care as a last resort, relying on in-person experiences when data suggests it is necessary. By prioritizing maturity and expert training, Duke aims to optimize the application of AI in healthcare.
Importance of User Implementation in Healthcare Technology
According to Dan Buckland, the Medical Director of Duke University Health System, the implementation of technology in healthcare is not solely dependent on the AI or technology itself. Rather, the end user's understanding and effective implementation of the technology account for 80-85% of the success in solving healthcare problems. By ensuring that the technology is rock solid, Duke can identify implementation errors rather than attributing them to technological limitations. This approach simplifies troubleshooting and problem-solving processes in healthcare.
Virtual Care as a Strategic Tool and the Limitations of Telemedicine
Virtual care, including telemedicine, is not viewed as a last resort at Duke University Health System, but rather as a tool with specific use cases. The prioritization of virtual care depends on the appropriateness of the situation and the level of interaction required. While virtual care can be effective for phone calls and asynchronous messaging, it may not always be necessary for video consultations, especially in acute care settings. The pandemic has revealed that the use cases for telemedicine with video interactions are not as wide-ranging as initially anticipated, highlighting the importance of finding the right balance between in-person and virtual care.
Friction Caused by Technology in the Patient Experience
In healthcare, technology can sometimes cause friction in the patient experience. Issues such as IT difficulties during telemedicine appointments can escalate a simple phone call to a lengthy IT troubleshooting session, resulting in patient dissatisfaction. Furthermore, whenever new technology is implemented in healthcare workflows, it requires training and may disrupt the efficiency of the existing processes. The turnover of healthcare employees can exacerbate this friction, as new staff members need to adapt to the technology and its integration into the workflow. Although technology itself may not be the cause of friction, human factors play a significant role in achieving efficient healthcare processes.
The Role of Generative AI in Healthcare Interoperability and Communication
Generative AI systems, including large language models, have potential in addressing the challenge of interoperability among healthcare IT systems. Specifically, these systems can assist in translating and summarizing medical information across different specialties or professional domains, where terminology and language may differ. By understanding and translating specialized vocabulary, generative AI can facilitate clearer communication between healthcare professionals who come from different backgrounds. However, considering the importance of HIPAA compliance and patient privacy, the adoption and implementation of generative AI must be approached in a safe and cohesive manner.
The Value and Role of Fax Machines in Healthcare Communication
Surprisingly, fax machines retain their value and integral role in healthcare communication. They serve as a common modality for sharing medical information between healthcare systems that do not share compatible electronic medical records. Faxing is often used when there is a need for communication across different systems, ensuring information exchange between clinics or institutions. While other industries have shifted away from faxing, healthcare continues to rely on this technology for the reliable and immediate transfer of patient information.
Technology and Tech Debt in Healthcare
The healthcare industry faces significant challenges related to tech debt, particularly in ensuring compatibility and interoperability among different systems. While generative AI shows promise in decreasing tech debt by improving interoperability, healthcare professionals still require training and tuning to effectively use the technology. Moreover, by fine-tuning language models to accommodate different educational backgrounds or language barriers, generative AI can support clearer and more efficient communication across the healthcare enterprise. However, it is crucial to balance advancements in technology with the need to maintain backward compatibility with existing systems, such as fax machines, to ensure seamless communication.
Future Prospects and Implications of Generative AI in Healthcare
Generative AI has the potential to revolutionize healthcare by enabling effective communication, reducing language barriers, and enhancing the patient and provider experience. As these technologies continue to develop, the healthcare industry must stay attentive to patient privacy and adhere to regulatory requirements. By leveraging generative AI appropriately, healthcare professionals can better understand and meet the unique needs of individual patients, leading to more effective care and improved outcomes.
Conclusion
The podcast conversation with Dan Buckland, Medical Director of Duke University Health System, sheds light on Duke's approach to AI technologies, their emphasis on maturity and expert training, and the strategic use of virtual care. The discussion also highlights the challenges and friction caused by technology in healthcare, as well as the role of generative AI in communication and interoperability. The enduring significance of fax machines in healthcare communication and the potential of generative AI to address tech debt are also addressed. Overall, the conversation provides valuable insights into the intersection of healthcare and technology.
Today’s guest is Dan Buckland, Medical Director of Duke University Health System. Dan is also an assistant professor at Duke University as well as a deputy human system risk manager at NASA. Obviously, he’s no stranger to the latest and greatest technology. However, Dan joins us on the program today to talk about Duke University’s unique approach to AI technologies - often using models that are at least two years old and have requisite training - and using virtual care as a last resort whenever data is showing in-person experiences are holding back patients and caregivers. To access Emerj’s frameworks for AI readiness, ROI, and strategy, visit Emerj Plus at emerj.com/p1.
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
