In this episode

Tina Shah, MD, MPH has extensive experience as a practicing physician, the Chief Clinical Officer at Abridge, a federal advisor on physician burnout for both the Obama and Trump administrations, and the VA’s first director of clinician well-being. In this episode, she discusses care team burnout, opportunities for using AI, and large-scale technology and provider experience trends.

Shah notes that burnout is contributing to the growing provider shortage in the United States, which will continue be exacerbated by an aging population in coming years. She believes that addressing burnout is possible, but requires investment in technology and change management to implement evidence-backed interventions.

“Can we return humanism back to medicine? I think AI has massive potential to do that. If we use this tool to help people focus on each other and have conversation, I believe burnout will improve.”

– Tina Shah, MD, MPH

 

Key takeaways

Dr. Shah discussed the contributing factors to clinician burnout and how AI could be used as one intervention to address them. She calls burnout “an occupational condition, literally caused by the toxic workplace culture in the United States and around the globe.” Dr. Shah argues that “if more than fifty percent of US physicians are burnt out, it’s clearly not a ‘me thing.’ It’s a ‘we thing.’ But somehow, we’ve stigmatized it.” She outlines some of the primary causes of clinician burnout as:

• “Pajama time” or after-hours work, especially charting.
• The cognitive load of providing care, which persists even if “pajama time” is addressed.
• Physician shortages and physicians exiting the profession.
• Lack of change management and other resources to implement evidence-backed interventions for burnout.
• Direct managers without the “full capability to help [physicians] thrive” or address challenging work conditions.

Reducing cognitive load, not just workload

Dr. Shah believes that AI can help solve not just direct tasks that outpace physicians’ capacity, but reduce the cognitive load on physicians.

“We all generally understand burnout. It’s like you’re all at the point of no return. You go on a vacation, but you come back to the same place, and you need dramatic changes…otherwise you’ll be in the same situation again. That’s burnout, and we have surveys that can measure it, the degree of it, and they’re validated,” she said.

“But there’s an intermediate metric called cognitive load, which is literally the amount of working memory you have, and have you exceeded it. And that has a validated survey that you can measure too. There are papers that show that [exceeding cognitive load] is directly linked to the odds of burning out and the odds of making a medical mistake on a patient.”

Dr. Shah cites promising research showing that AI applied to repetitive tasks, such as responding to basic messages, might not meaningfully reduce physicians’ workload but does reduce cognitive load.

“With Abridge’s AI for SOAP note documentation, we’re seeing crazy reductions in cognitive load, somewhere in the order of 40-80% percent reduced cognitive load,” said Dr. Shah. “I’m really, really excited because as a burnout expert, I’ve never seen any intervention move the needle like this.”

Allowing providers to be more present 

Dr. Shah sees promise for AI to help providers be more present with their patients due to reduced cognitive load, which could also improve patient care.

“We’re asking a question to clinicians about ‘distracted doctoring,” Shah said. “It’s essentially a question that says, ‘How present do you feel with your patients? [With AI support], we’re seeing something like 30-40%  more more physicians and say ‘I am more present with my patient.'”

“Just speaking about myself, perhaps if I was more present, I would write a nicer message back to my patient. There’s something about when you have just so much work to do, you aren’t as thoughtful. Also, you miss cues.”

Dr. Shah expresses optimism that with applications of AI, such as ambient listening, providers will have more complete documentation and fewer administrative tasks that allow them to provide better patient care.

Dr. Shah’s recommendations for implementing AI

Dr. Shah has three recommendations for health systems evaluating potential AI solutions:

• “Try it, and fail fast. You could take the perspective of, ‘I’m going wait. We’re not an early adopter organization. We’re running on razor thin margins. We just can’t afford it.’ But it costs between $500K-1M every time you lose a physician to burnout…Use AI for a low-stakes use case, somewhere  like clerical work, whether it’s for the front desk staff or your physician,” to help you get started and see wins quickly.
• “Expect your vendor to prove it,” showing data on the improvements provided by their solutions. “You should ask, ‘Can you show me your data on what you did with cognitive load?’
• “Focus on transparency because there is a lot of concern about AI coming into our clinical workspace and it hurting somebody. I think the stakes are way too high…it’s an appropriate question to say, ‘Help me understand, without being a computer scientist, how does your technology work?'”