Penn Medicine shortens ICU stays with real-time data

Many patients in hospital intensive care units (ICUs) depend upon a breathing machine for survival. Getting critically ill patients breathing on their own is an essential step towards getting them safely out of the ICU and freeing beds for other patients.

In an effort to shorten the amount of time patients spend on mechanical ventilation, the non-profit Penn Medicine (consisting of the Raymond and Ruth Perelman School of Medicine at the University of Pennsylvania and the University of Pennsylvania Health System) has built a dashboard that leverages its electronic health record (EHR) vendor’s real-time data streams to alert respiratory and nursing staff when interventions are needed and when patients may be ready to be weaned from ventilators. Penn Medicine has received a Digital Edge 50 Award for digital innovation for the project, dubbed the ABC application (for Awakening and Breathing Coordination).

[ Learn the essential skills and traits of elite data scientists and the secrets of highly successful data analytics teams. | Prove your data science chops by earning one of these data science certifications. | Get the insights by signing up for our newsletters. ]

“Many of the critically ill patients we take care of here in the ICU require a breathing machine to survive. That’s a mechanical ventilator,” says Dr. Barry Fuchs, medical director of the Medical Intensive Care Unit and Respiratory Care Department at Penn Medicine. “Although these ventilators save the lives of patients, they’re associated with risks and complications. And the longer that patients remain on a breathing machine, the longer they stay in the ICU.”

Longer stays in the ICU aren’t just associated with risks and complications. They also drive up costs. Penn Medicine’s ICUs serve just five percent of the patient population of the organization, but account for roughly 15 percent of the organization’s costs, Fuchs explains.

The amount of time patients spend in the ICU is a key metric for Penn Medicine, and reducing it has been a goal for years. It’s a metric that is very difficult to move the needle on, but benchmark data pinpointed an opportunity, Fuchs says.

“A few years ago, we learned that at Penn Medicine, patients stay on the ventilator longer than expected based on the benchmark data,” Fuchs says. “So, we knew there was an opportunity to reduce the amount of time that patients stayed on the breathing machine. We knew that if we could solve this problem, it would improve both the care for the patient individually and reduce the costs in the ICU as well as hospital costs.”

A solution forged in collaboration

Penn Medicine discovered two major factors that contributed to its ICU patients staying on mechanical ventilators for 12 to 24 hours longer than expected. The first were delays in reducing ventilator support by respiratory therapists and sedative medications by nurses. The second was that the process respiratory therapists follow to assess patients for eligibility for ventilator liberation trials was inefficient and complicated: It was a manual process performed once a day in the morning. If the patient wasn’t screened at that time or failed the screening in the morning, they wouldn’t be screened again for another 24 hours.

To address these issues, Penn Medicine built the ABC application, an electronic dashboard and alert system that gathers sedation and ventilation protocol data in real-time and applies clinical decision support (CDS) rules based on ICU treatment guidelines and input from experts. Patients must awaken before they can be taken off ventilation, so the ABC application sends text alerts to bedside nursing staff when a patient is over-sedated, allowing them to scale back or stop sedation. Additionally, rather than waiting for the once-a-day assessments previously used to determine whether a patient was ready to come off ventilation, the application sends respiratory therapists alerts when a patient’s vital signs meet certain criteria. The respiratory therapists can then conduct a trial to determine whether the patient is ready to breathe on their own.

Implementing the ABC application required Penn Medicine to bring together its Penn Center for Health Care Innovation with the information services (IS) and data science departments. The innovation center and IS built the dashboard based on the CDS engines and alerts created by Penn Medicine’s clinical experts and programmed by the data science department.

“This was the first project that really drove the deep collaboration with these three departments,” says Michael Draugelis, chief data scientist, Penn Medicine. “That project, that interaction, has adjusted the way that we work. We’re organizationally shifting in how we manage our projects.”

Finding strength in setbacks

The project was not without its difficulties. Months after the pilot started in December 2016, Penn Medicine changed its EHR vendor. That delayed the whole project by months as they had to get the new EHR up and running and then reconnect the application to it.

“That was one of the biggest delays in getting this to implementation,” Fuchs says. “It was a huge hurdle that the teams were able to get around.” And they gained valuable experience that helped create a better application.

“Building out our first deployment on our first EHR definitely drove the requirements in terms of how we wanted to interface with our new EHR and that process generalized our interfaces with those two different systems,” says Draugelis. “It made it more robust and helped us have a more foundational build with our current system. It was painful, but it paid back.”

“I think it validated our vendor-neutral strategy,” adds Kevin Pollock, senior software development manager in Penn Medicine’s IS department. “Data is data. We should not be thinking of vendors, but how we can use data.”

To date, the ABC application has helped Penn Medicine reduce the time ICU patients spend on a mechanical ventilator by more than 24 hours. The organization has since added fields to facilitate delirium management to the platform and plans to add components for tracking exercise and family engagement.

More on data science:

• What is data science? A method for turning data into value
• Essential skills and traits of elite data scientists
• Developing data science skills in-house: Real-world lessons
• The age of the citizen data scientist has arrived
• Top data science certifications that will pay off
• Best data science bootcamps for boosting your career
• 10 hot data analytics trends — and 5 going cold
• 19 big data certifications that will pay off
• 12 myths of data analytics debunked
• 7 sure-fire ways to fail at data analytics
• The secrets of highly successful data analytics teams
• How to assemble a highly effective analytics team
• IT career roadmap: How to become a data scientist
• What does a data scientist do? 7 of these in-demand professionals offer their insights