77 lives: That’s how many patients Houston Methodist Hospital saved over a nine-month period in 2014 and 2015 after deploying an early warning system that continuously scans the electronic health record (EHR) to identify patients whose condition is deteriorating.
The 11 nursing units that initially used the tool saw their risk-adjusted mortality decline by 32 percent, which includes an 8 percent reduction in sepsis deaths.
“This tool is saving lives that at one point we would have considered unavoidable deaths,” says Katherine E. Walsh, DrPH, RN, vice president of operations and chief nursing officer, Houston Methodist St. John Hospital. “It’s an easy-to-use tool that helps us intervene sooner to further improve outcomes.”
The 1,119-bed Houston Methodist Hospital now uses the tool on all its nursing units, and the other six hospitals in the Houston Methodist system are starting to use it. It’s integrated to pull data from the EHR and makes use of other information technology to deliver its results.
The early warning system, developed by PeraHealth, incorporates the Rothman Index, which assigns patients a score of 0 to 100 (with 0 being the poorest health). The score is determined via an algorithm that assesses a patient’s vital signs, lab values and functional status as determined by nursing assessments.
When a patient’s status deteriorates, nurses at Houston Methodist follow protocols established by nursing leaders. Depending on the patient’s condition, a nurse might consult a physician, take the patient’s vital signs more frequently, provide specific help to the patient (for example, provide additional oxygen), or call the hospital’s rapid response team to come immediately.
The Rothman Index is unique in that it identifies gradual changes in a patient’s condition—from assessment to assessment, day to day, or visit to visit—that physicians and nurses often miss because the information that would clue the clinicians to these downward trends is buried in the EHR, explains PeraHealth’s Michael Rothman, PhD, chief science officer.
Rothman developed the index with his brother Steven after their mother died from an undetected complication following a routine, low-risk surgery. “My brother and I sat down and said, ‘What went wrong? This was a completely avoidable death,” remembers Rothman.
Despite their grief, the Rothmans were able to step back and evaluate the situation logically, drawing on their dual experience as scientists and data analysts. They concluded that their mother’s death was due to an EHR failure. “Although the electronic health record had all the data in it, it was impossible to see a trend. And that’s what the clinicians missed in my mother. They could not see the downward trend in her condition.”
The other unique aspect of the Rothman index is that it interprets nursing data, which is rarely used in as part of predictive analytics. As part of standard nursing practice, nurses conduct head-to-toe evaluations of hospital patients, assessing each of the patient’s physiological systems, Rothman explains. Nurses then document whether or not the patient meets a minimum standard for each physiological system. For example, patients who meet the minimum standard for the gastrointestinal system have no nausea, are bowel continent and meet other criteria.
The nursing data have been key to the tool’s success to date, Rothman says. Because nurses conduct their assessments frequently, the nursing data capture a patient’s functional status over time, providing an early indication when something is going wrong. “Nurses start to pick up problems before patients get to the crisis stage. The nurses note when patients stop eating or are confused or are having trouble walking or start to build up fluid in their extremities. These are all the things picked up by nursing assessments, and they are picked up before you see changes in vital signs.”
The Rothman Index has been in the making for more than 10 years. The Rothman brothers began by collaborating with the hospital where their mother died, Florida’s Sarasota Memorial Hospital. “My brother and I spent several weeks at the hospital observing, and we worked with a nurse informaticist to develop a relatively crude model that looked positive enough to get people excited.”
After testing that model on various patient populations, the Rothmans continuously refined the algorithm into a more rigorous statistical model, publishing 18 peer-reviewed journal articles on their research.
Houston Methodist participated in some of the research studies to help validate the tool. Then in 2014, nursing leaders decided to pilot the tool as an early warning system on 11 nursing units.
The early warning system does not require additional data entry because it pulls information directly from the EHR as clinicians document care. However, adopting the tool still represented a change in practice for nurses. Walsh credits three strategies that Houston Methodist used for the hospital’s success.
First, the roll out was driven by nursing leaders. “The managers, directors and myself, the VP, were actually on the units driving utilization,” Walsh says.
Second, the tool was made visible via large TV monitors on the nursing units. Color-coded graphs from the tool make it easy to tell which patients have taken a turn from stable (blue) to deteriorating (yellow) to high risk (red). “Anybody walking onto the unit can look at the big monitor and see the index,” Walsh says. “You can see the colors change from across the room, ‘Oh, there are three red patients,’ and those are patients that you need to attend to.”
The third key strategy was taking time to educate nursing staff on the tool and then encourage nurses to share stories of patients saved by the tool. Walsh recalls one story of a charge nurse who saw from the Rothman Index that a patient was deteriorating and quickly enlisted a resident’s help to get the patient transferred to ICU. “When you can say, ‘This is what happened to our patient Miss Jones and this is what we did,’ that is very meaningful to the nurses.”
Houston Methodist is finding additional ways to use the tool, such as preventing readmissions and ICU bouncebacks. Recalling one patient who had had three admissions in six months, Walsh says a nurse used the Rothman Index to identify that the patient’s condition had worsened from one admission to the next, primarily related to a renal condition.
Register or login for access to this item and much more
All Health Data Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access