Leveraging AI to reduce low-value tasks can create decision-making capacity.
It was late in the afternoon on a busy Monday, my inbox overflowing with messages and patients lined up in rooms waiting for me, when my medical assistant asked me a routine question about an FMLA form. I stopped her mid-sentence and said, “I can't deal with this right now. I'm mentally underwater and have 10 other priorities to manage first.”
Months later, when my colleague, Doug Fullington, MD, boasted about being intellectually fresh at 4 p.m., I stopped to listen closely. We were at a leadership meeting, and the topic of artificial intelligence (AI) had been interwoven throughout the day. Thanks to judicious use of AI tools, he said, “When that 4 p.m. chest pain patient gets added to my schedule, I'm still ready to go!”
He went on to explain how he can save his intellectual reserve for the entire day by using AI tools to help with the low-value but never-ending grind: inbox management, chart prep, documentation, and chart review.
COGNITIVE CAPACITY PRESERVED
We are at an inflection point in health care. Similar to when the stethoscope became a tool in every doctor's bag,1 AI is becoming an indispensable tool today. But the real power of AI isn't necessarily in the time saved but rather the cognitive capacity preserved, as my colleague discovered. It's an intellectual pressure relief valve, helping preserve our cognitive reserve.2
That's not an insignificant thing. An Australian study two years ago showed how physicians were more likely to prescribe antibiotics toward the end of their workday.3 Interestingly but not surprisingly, they were less likely to prescribe statins and treat osteoporosis at the end of the workday versus earlier. If we think about what's different from the beginning of the workday to the end, it's how much charge remains in our cognitive battery. Having a conversation with a patient about antibiotic stewardship takes time. Talking about the benefits of statins in reducing ASCVD risk requires shared decision making that can be mentally taxing.
This phenomenon is not limited to medication prescribing. As the day goes on, we are less likely to engage our patients in discussions around cancer screening.4 Think about that: A patient at 4 p.m. is less likely to get a mammogram ordered than a patient at 8 a.m. While not defensible, it's understandable. These shared decision-making conversations require a mental investment that can be challenging at times.
To be fair, there is some debate about how much decision fatigue really exists in health care settings.5 Yet while the impact might not be large, it's subjectively widespread and the underlying argument still applies. Every physician recognizes the experience, and the documentation burden driving it is not in dispute.
We talk so often about “pajama time,” the after-hours time spent on documentation, and AI tools have been proposed as a solution. Turns out, AI scribes might not reduce pajama time,6 and I'm OK with that. I want help throughout the day managing the rote tasks.
TASKS VS. DECISIONS
I've talked before about “tasks for staff, decisions for physicians,”7 the concept of offloading non-clinical responsibilities from physicians so we can be free to focus on what we've been trained to do: complex medical decision making. Leveraging AI tools to help with those tasks keeps us mentally charged up to focus on what matters most: our patients. No matter what time of day.