Open Mics with Dr. Stites: “Show Me the Science” How Artificial Intelligence Helps Patients

The University of Kansas Health System is treating a total of 35 COVID patients today, versus 30 yesterday. Other significant numbers:
19 with the active virus today, 16 yesterday
3 in ICU, same as yesterday
1 on a ventilator, 2 yesterday
16 hospitalized but out of acute infection phase, 14 yesterday
Key points from today’s guests:

Chris Harper, vice president and chief informatics officer, The University of Kansas Health System
The health system is looking at data to turn it into intelligence to help medical staff improve patient care.
The AI Center of Excellence puts the human element in the center of the technology
Siri and Alexa are considered part of AI with machine learning, along with Tesla as they collect data.
AI is another smart tool to help us take care of patients better.
We have over 3.5 petabytes worth of data in our healthcare and what the team has done is really create a data system that is able to really work with researchers here to use different modeling techniques, different data centric AI capabilities to really make sense of it by using things like predictive modeling.
The exciting thing is we're at a new age of medicine and science that we're able to apply more of these AI techniques that we've been creating for the last 20 years and really making intelligence out of those data and the systems.
Dr. Alan Yu, nephrologist, The University of Kansas Health System
Transfer learning is actually a new field of learning in machine learning. It's aimed to improve learning on a new task based on what the machine has learned from us related tasks.
And it's important for personalized learning because the conventional way of building and prediction model is to train your model using a particular population data from particular population, and then the model can be applied to the new patients.
But for personalized predictive modeling, the difference is that actually you're training a model dynamically for each individual patient coming into the hospital.
It is thrilling to use science like this to improve patient care.
Mei Lu, Ph. D., associate professor, Medical Information, The University of Kansas Medical Center
There are kinds of kidney AKI detection systems implemented in the hospital where you can look at looks at the serum craning values and then can produce early warning signals in the electronic health record systems.
Currently, we are working on how to predict risk before onset.
There are pros and cons to AI, but I am looking forward to using it to improve outcomes.
Dr. Steve Stites, chief medical officer, The University of Kansas Health System
The whole goal with all major areas of critical care and hospital care now is to make sure we keep the kidneys healthy, because we know that when we don't the risk of death and the risk of bad outcomes or prolonged recoveries, it goes up remarkably, and we try to do those things are going to help take care of the kidney.
We are the drivers of new things like AI.
It is amazing to think about what the future of AI in medicine.
Dr. Dana Hawkinson, medical director, Infection Prevention and Control, The University of Kansas Health System
Right now, anybody 12 and over who received their last booster, two months ago or their last dose two months ago, is eligible at this point in time for the new booster.
We may have a yearly COVID vaccine.
Currently, there are no other variants of concern on the horizon.
We know that the virus is constantly changing.
We haven't seen that huge surge in hospitalizations like we did in that early Delta and early Omicron.
Thursday, September 8 at 8:00 a.m. is the next Morning Medical Update. You don't think of puppy kisses and wagging tails as medicine, but it can be! Studies show Fido is not only a distraction for patients in pain, but a way to ease stress and lower blood pressure. We'll take you inside the hospital's pet therapy program.