Cognitive Computing

How Cognitive is Helping One Hospital Curb ‘Code Blues’

Share this post:

Preventable medical errors remain the third leading cause of death in the United States, despite tremendous investments in healthcare technology to improve care, standardize practices, and make choosing “the next best action” easier for healthcare professionals.

In Canada, Code Blue, the term used for emergency medical intervention, is the second costliest type of acute care claim for hospital insurers. According to Blair Bigham who reported on this topic for CBC news, insurers end up paying many of these claims because it is often determined that clinical staff could have recognized that a patient was at risk. Because of such results, Code Blue is often referred to as a “failure to rescue.”

Dr. Alison Fox-Robichaud, Critical Care Physician at Hamilton Health Sciences and Associate Professor of Medicine at McMaster University, heard Code Blue calls many times a day while working at the Hamilton General Hospital. It was her belief that “failure to rescue” implied the possibility of success, and furthermore a possibility to predict and prevent Code Blues from happening in the first place.

Imagining a world with no Code Blues was Dr. Fox-Robichaud’s vision.

In the spring of 2016, Hamilton Health Sciences teamed with IBM on an integration platform that combined IBM’s Watson Explorer natural language processing and analytics capabilities with Thoughtwire’s Ambiant technology.

Four high-priority projects and sponsors were initially identified, but the most impassioned sponsor was Dr. Fox-Robichaud. Her project and vision was to bring technology to bear on the Hamilton Early Warning Score research (HEWS) in order to eliminate Code Blues.

The commitment, experience and expertise in the Health Information Technology Services and Clinical Informatics groups added to our confidence at IBM. Despite the fact there were easier projects we might have selected, HEWS was the most profound.

The project involved many facets, such as using the leading edge research conducted by Dr. Fox-Robichaud to better understand the key vital sign metrics and indicators that could predict and prevent Code Blue. We then automated vital sign capturing via integration to electronic medical records (EMR) to eliminate delays in identifying patterns of decline.

We embedded the research into an algorithm for action and paired the technology and research with changes in HHS policy, culture and practice to support the right action by the right person at the right time.

We then made all of this accessible via mobile devices with “smart agent” technology that provided timely action notifications.

The results have been tangible. In 2006, Hamilton General responded to 400 Code Blues. In 2016, it logged just 54.

I recently had the privilege of hosting a webcast with Dr. Fox-Robichaud, Christine Probst and Leslie Cicero of Hamilton Health Sciences to share this exciting work: http://ibm.biz/earlywarningscore

What’s next?

But we’re not stopping there. The next projects on tap are Pediatric HEWS, Sepsis prediction and prevention, and we’re adding natural language processing capabilities to the platform to be able to access the rich clinical documentation that is text based.

_________________________________________________

A longer version of this story first appeared on Linkedin.

Add Comment
No Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More Cognitive Computing Stories

Bringing Cognitive to Bear in the Fight Against Poaching in South Africa

I’ve always wanted to be a large-animal vet. As a child, I was caretaker to Russell, the Dutch rabbit, two guinea pigs named peanut butter and jelly, more than 100 rats, mice and more blue-belly lizards than my long-suffering, but supportive mother could count. I’ve always felt a connection to animals and though my path […]

Continue reading

Auto Industry Turns to Cognitive to Redefine Vehicle Safety

My first car was a ’65 Chevy Impala, which I bought from my uncle. The engine was in bad shape so within the first week, I swapped it out with one I bought from the local junk yard. It was a pretty simple car – manual transmission, manual windows and locks, no air and an […]

Continue reading