Transcript: The importance of nursing assessments in AI modeling to predict patient deterioration

Watch the video here.

TRANSCRIPT
Tuesday, March 25, 2025 (length: 3 minutes 36 seconds)

[Michael Rothman, PhD]

My mother, she was a patient at a hospital in Florida and had an operation. After the operation, she started to deteriorate slowly. She was discharged from the hospital, came back four days later to the emergency room, and passed away.

After this happened, my brother and I really understood that this was an avoidable death, and asked the question, Why? We started a project that became known as the Rothman Index in memory of my mother.

Turns out, when a nurse goes in to see a patient, he or she does a very structured evaluation of that patient’s condition. They are grading each of about a dozen physiological systems, and that data is critical in what we call the Rothman Index. It shows us when things are going wrong and there’s still time for the doctor and nurse to come in and intervene and prevent that crisis.

The idea is to take the data, extract it, and compute a single number which is a measure of how sick or how well you are.

We have users that love using our tool because it helps them do their job. I was just speaking to a physician who organizes his day by our score. He’s added a column into his patient list, which is the Rothman Index. He sorts the list and starts by seeing the sickest patient first.

The Rothman Index model was developed using SAS analytical software. We’ve spent years studying the impact patient care; on outcomes. Using SAS to do the analysis. So it has been an essential tool as we have developed our system over time.

Under controlled conditions, we see a statistically significant decline in risk-adjusted mortality rates. You know, we expected two percent of patients were going to die, and 1.3% of patients died. That doesn’t sound like much, but that’s a 30% reduction in mortality.

What we’re doing is helping doctors and nurses do their jobs more effectively, more efficiently, and to the extent that we do that, we improve outcomes. We’ve reduced mortality rates, we’ve reduced unplanned transfers to the ICU.

We entered into this whole enterprise with the idea that, if we could prevent what happened to our mother from happening to one other person, it would be worthwhile. And that’s why we’ve been working on this.