When I lived in Los Angeles, neither I nor anyone I knew watched the weather. I often wonder how local meteorologists keep focusing on providing “Today will be hot and sunny” or “Today will be hot and sunny”.
This is not always the case: I grew up in Athens, Ohio, and later lived in Boston and Cleveland. In all three places, I spend time wondering and talking about what the weather will be like, how this will affect my activities, and how we plan our expectations. I will be prepared for any adjustments. Accurate weather forecasts have helped in this process and are very helpful.
Now imagine you were hit by a hurricane. You have a real-time forecasting weather analysis service that grabs your soaked woolen sweater and screams: “It’s going to rain, you need to be prepared.”
With the onslaught of monsoons, it’s hard to imagine that you will appreciate the update, nor will you be willing to follow the service when it asks you to acknowledge its very insightful and useful warnings. As clinicians, this describes our daily interaction with clinical decision support.
With my academic background in the decision-making process of experts (especially doctors), the promise of computing power in clinical medicine to provide information for expert decision-making, just like in meteorology, is very exciting.
We dream that one day, a caring and compassionate doctor is equipped with the latest artificial intelligence computer system, which can find out the exact cause, treatment method and guidance. Not only can this help anxious patients avoid disasters, but it can also help them thrive in ways previously unimaginable.
Unfortunately, our daily clinical practice does not match reality. The clinical decision support system provides 1 part useful insights and 99 parts frustrating distractions. Many of them fall into the category of “if you need to tell me, everything is lost”. We were told it was raining and we were soaked.
For example: I am standing in the emergency department, just received the code 3, high sensitivity ambulance, 47 years old male, fever 102, heart rate 160, blood pressure 70, blood oxygen saturation 70%. I intubated the patient, quickly placed the centerline, and then ran to my computer to enter orders for antibiotics and other sepsis packages. Then I have to tell me through the computer that I should think this patient might be sick and explain why this patient might be sick, and explain why I should consider sepsis. If the clinician sees the patient and does not realize that the patient is in septic shock, the computer pop-up window will never save the patient.
Another major issue? The information provided by our clinical decision support system will not lead to any action, intervention or other trajectory changes. In a world where our attention is the most precious resource, this information just needs more attention and distraction.
Constantly requesting to document the rationale for my evidence-based care, a pop-up window detailing drug interactions based on case reports of alternative dimensions, and side effects warnings that I’m already familiar with: I need to intubate the paralyzed patient and the patient may stop breathing and cause them to stop breathing. Need airway management.
As a clinician, I don’t want to send out sepsis alarms all the time, causing people to run around and create more chaos in a chaotic environment. Likewise, I don’t like driving and being instructed to go out of the exit I just passed.
I want others to say that the patient in front of me has some mild infections, but it is likely to be infected within one or two days or even within three days. As Dan Heath detailed in his book Upstream, I want to “solve problems before they happen.” I would call it insight. Then I can prevent sepsis alarms, shock alarms, and other confusions. I can also develop a clear path for each patient so that those who will become septicemia receive some special attention and close monitoring in a calm, planned and thoughtful manner.
The clinical decision support system does not need to be distracted. We can design them to stop pointing out patients in disasters and start helping us prevent them from getting there. We can predict the weather in a way that we can predict-if not in Los Angeles-then in Cleveland or Boston. I want to tell me something, “It looks sunny outside now, but there is a 20% chance it will rain… so bring an umbrella this morning.”
Photo: Metamorworks, Getty Images
