Hospital cuts costly falls by 39% due to predictive analytics

El Camino Hospital linked EHR, bed alarm and nurse call light data to analytics to identify patients at high risk of falls.
By Bill Siwicki
01:05 PM
Share
Cheryl Reinking

Chief nursing officer Cheryl Reinking helped spearhead the effort to transform the organization’s fall prevention program using prescriptive analytics.

The 420-bed El Camino Hospital in California has seen a dramatic turnaround with is hospital fall rates, and advanced predictive analytics is getting a lot of the credit.

“We were having a lot of difficulty being able to get our falls under control and at the level we wanted them, to be in the top quartile in the nation. It seemed like a lot of the efforts we tried were not getting us to where we wanted to be,” said chief nursing officer Cheryl Reinking. “We had heard through different connections in Silicon Valley about this company, Qventus, that was able to do predictive analytics. They had done some work with us, in the operating room.”

Reinking helped spearhead the effort to transform the organization’s fall prevention program using prescriptive analytics to ensure patients were being proactively and optimally managed. And the successful effort involved going beyond traditional “predictive analytics” into what Reinking calls action-focused insights that allow providers to immediately respond and impact patient safety.

[Also: How analytics roadmaps can break organizations from vicious data cycles]

At the beginning, Reinking said El Camino staff and Qventus staff sat down for formal talks to go over the details of their problem.

“They were able to help us come up with a methodology where we could in real-time know and understand which patients are at higher risk for falls so we could focus our efforts clinically in real-time on those patients,” she said. “We know which patients through screening at admission are at high risk for falls. We put that in their medical records, and we make sure they have yellow slippers and yellow bands and a sign on their door so clinicians know they are at high risk of falls. And that is a lot of patients in the hospital.”


 Learn more at the Big Data & Healthcare Analytics Forum in San Francisco, May 15-16, 2017. Register here.​


Qventus pulls data from El Camino’s electronic health record and looks for patients at high risk, and then looks at nurse call light data and bed alarm data.

“So this patient sets off the bed alarm the most, uses the call light the most, Qventus does machine learning in the background and once certain triggers are met, they send an alert to the nurse through Vocera with a report saying your patient in room X is at high risk for a fall,” Reinking said. “That tells the nurse they need to go look at that patient to see if there is something else they need to do.”

This is going beyond predictive analytics, Reinking said, to action-focused insights or prescriptive analytics.

“When certain data elements line up, you take action in the moment so that your prediction does not come true,” she said. “You can change the outcome in the moment. So if a patient is predicted to fall, let’s take action now so that a patient does not actually fall.”

Nearly 1 million patients in the United States. fall in the hospital each year, according to the Agency for Healthcare Research and Quality. Further, patients who sustain an injury from a fall add 6.3 days to a hospital stay and cost an additional $14,000, according to the Joint Commission Center for Transforming Healthcare.

Through analytics, El Camino’s care team was able to predict exactly which patients were at risk for an imminent fall and alerted nurses and case managers of at-risk patients in real-time, which was one factor that helped result in a 39 percent reduction in falls within 6 months, Reinking said.

“It was this effort with the use of predictive analytics, but many other things, too,” she said. “We also were working on initiatives like making sure the risk assessments were really appropriate for our patients when they first came in. We also did a lot of training on the bed alarms. We got more chair alarms in place. We also did a lot of training on bathroom falls. Understanding patients fall from a chair and the bathroom more than from a bed. Those were some of the other actions that were occurring during this time along with the work that was occurring with Qventus.”

Reinking will deliver an address on predictive analytics at the HIMSS and Healthcare IT News Big Data & Healthcare Analytics Forum on May 15 and 16 in San Francisco during a session entitled “Going Beyond Predictive Analytics to Impact Care.”

Twitter: @SiwickiHealthIT


 Read more of our preview coverage of the  Big Data & Healthcare Analytics Forum in San Francisco, May 15-16, 2017.
⇒ How analytics roadmaps can break organizations from vicious data cycles
⇒ Machine learning 101: The healthcare opportunities are endless
⇒ Big data is actually 'Bigger Data' for healthcare
⇒ HIMSS Big Data and Healthcare Analytics Forum to focus on machine learning, value-based care and population health