Researchers at Cincinnati Children’s have created a software application and clinical system called NINJA that taps into the power of electronic medical records to protect hospitalized children from a potentially serious safety threat – acute kidney injury (AKI).
After three-plus years of use at Cincinnati Children’s, the approach achieved a 38% reduction in nephrotoxic medication exposures (600 fewer cases) and a 62% reduction in AKI rates (almost 400 fewer cases). Additional hospitals are now testing versions of the system and are reporting an overall 30-percent reduction in AKI rates.
Built on research at Cincinnati Children’s and powered by electronic health record (EHR) data, NINJA—which stands for Nephrotoxic Injury Negated by Just-in-time Action—is now working to prevent pediatric acute kidney injuries through a 14-hospital collaborative. And, plans are taking shape to roll out the process at up to 110+ children’s hospitals across the United States and Canada over the next few years through a partnership with the Solutions for Patient Safety (SPS) network.
Three Cincinnati Children’s physician-researchers are behind the NINJA effort. Stuart Goldstein, MD, a nephrologist and director of the Center for Acute Care Nephrology, laid the groundwork for the system with his research into the causes and impacts of AKI on hospitalized children. He worked with Eric Kirkendall, MD, MBI, the hospital’s Associate Chief Medical Information Officer and a researcher in Biomedical Informatics, to build an EHR-based system to gather and report the data needed. Together with Steve Muething, MD, co-director of Cincinnati Children’s Anderson Center for Health Systems Excellence, they worked to apply quality improvement tools and processes to the problem of keeping kids safe from AKI.
Early Research Builds the Foundation
Nephrotoxic medication exposure is one of the most common causes of AKI in hospitalized children.
Goldstein’s research has shown that one of every four children outside the ICU exposed to nephrotoxic medications will eventually develop AKI. (His research defines an “exposed” patient as one who has received three or more nephrotoxic medications or an IV aminoglycoside for more than 3 days). He has found that, even after those kids recover from an AKI episode, they have a 50% chance of progressing to chronic kidney disease, a serious condition that could lead to kidney failure as well as the complications of decreased kidney function such as cardiovascular disease, anemia, and bone disease.
“I began this line of research because exposure to nephrotoxic medications is one of the most common and preventable causes of AKI in non-critically ill hospitalized children,” he says. “After frequently being called for consults, I became tired of writing ‘AKI secondary to nephrotoxic medications, will follow up with you’ in the charts. I wanted to do something to prevent the problem from occurring at all.”
Early on, little data on accurate AKI rates were available. In trying to develop that initial data, Goldstein found that the standard, inexpensive test that measures kidney function – serum creatinine (SCr) readings – was performed inconsistently, typically only about half the time kids were exposed to nephrotoxic medications. An early research effort was simply to test the impact of conducting routine SCr tests on all patients exposed to such medications.
“The research started as an economic exercise: What is the cost v. benefit of frequent SCr tests?” he says. “Eventually, it became clear that several medications in various combinations increased risk of AKI. Our vision became to ensure that children only get the nephrotoxic medications they need for the duration they need them.”
EHR Data Proves Key
Goldstein turned to Kirkendall to develop an EHR-based screening intervention to reliably detect nephrotoxic medication exposure in order to direct daily SCr assessment in non-critically ill patients. The NINJA team used those data to assess the rate of high nephrotoxic medication exposure and associated AKI as well as ongoing changes in AKI prevalence or duration as they worked to reduce those exposures and injuries.
By reviewing prescribed medications and daily SCr test results for patients who eventually developed AKI v. those who did not, they developed a list of 45 nephrotoxic medications to be closely monitored.
“Even though they can be nephrotoxic, these medications are often necessary. They include everything from ibuprofen to chemotherapy agents and certain classes of antibiotics,” says Kirkendall. Many of these medications are benign when used in isolation but can cause damage when used over an extended time or in combination with other medications.
“This isn’t just physicians prescribing medications without considering consequences,” he says. “In clinical care, the physician must choose the best medication to get the job done. But our system, with an eye to the heightened risk of AKI using certain medications, asks doctors to be vigilant when prescribing them and to take compensatory measures where needed.”
This is where the beauty of a computer-based system comes in. It is difficult to impossible for a physician or pharmacist to constantly analyze a crush of data and keep multiple, shifting variables in mind at every point in a patient’s care, across multiple patients. Luckily, that task can be executed by computers.
It was up to Kirkendall to develop a computer-based, risk-stratifying system based on electronic triggers—pieces of data or elements from the EHR that signal underlying events of interest. Triggers can efficiently survey an entire patient population for adverse events, with near-real-time identification. Those triggers operate in the context of an algorithm or computer-based decision tree, diving into EHR and electronic medication administration record (eMAR) data and sorting through it to identify patients exposed to target medications. It detects and flags patients at risk of AKI due to medication choices before injury occurs.
The system generates a daily report detailing those patient exposures and delivers it to the clinical pharmacists caring for the patients. The pharmacists share those results with the treating physicians as well as the AKI team and leadership.
“The system offers pharmacists a communications tool they can use to discuss medication risks and benefits with treating physicians, and grounds those conversations in data and evidence, ” says Kirkendall. “The system is based on supporting the relationship between the pharmacist and the clinicians rather than inaccurate, often irritating EHR alerts.”
The results have already been impressive. Simply by prompting this ongoing conversation between pharmacists and physicians, the NINJA system has succeeded in significantly reducing both the use of nephrotoxic medications and the incidence of acute kidney injuries.
Continuous Road to Safety
When Goldstein first brought this line of research to Cincinnati Children’s from his previous job at Texas Children’s Hospital, he consulted with Steve Muething, who also serves as the SPS strategic advisor. Muething and his quality improvement experts worked with Goldstein and Kirkendall to initially design and then constantly improve a process to integrate their electronic surveillance system into the clinicians’ workflow. They applied quality improvement tools early and often throughout the NINJA system’s development.
Based on early NINJA successes, Goldstein won a $100,000 grant from the Casey Lee Ball Foundation in 2014 to roll out the system at additional hospitals. That win was followed by a 3-year, $1.5 million grant from the Agency for Healthcare Research and Quality. Today, 14 hospitals are using the system, including the original nine hospitals funded by the grants plus five more that are funding the system’s costs on their own.
Once the AHRQ grant expires in 2018, the team plans to roll out the process within the SPS network of hospitals, initially to 15 more hospitals. Eventually, the program will be available to the entire SPS network of 110+ hospitals nationwide (including Canada).
SPS is a national learning network of hospitals working to achieve specific goals to reduce harm in pediatric hospitals through the transparent sharing of data, successes, and learnings. It currently oversees the use of 12 different “harm initiatives” or “safety bundles” in its hospitals. Muething is looking forward to adding NINJA to that list.
“Safety is a forever journey,” says Muething. “When we roll out a new safety initiative, our initial goal is typically to reduce harm across target hospitals by 40 percent within two years. But we don’t stop there. We will continue until we’ve reached the level of zero harm.”
Muething points to the EHR-based data as one important key to success. “All our hospitals say that reliable data is the lifeblood of their safety efforts,” he says. By constantly analyzing results and feeding lessons learned back into their processes, the SPS process aims for continuous quality improvement. “Without that data, these initiatives would not exist.”
Ongoing research into AKI and nephrotoxic medications is continuing behind the scenes. The list of potentially nephrotoxic medications was recently expanded to include 61 drugs based on experience from multiple hospitals.
Future research goals include combining the expanded data sets available from the additional hospitals to further study potentially harmful medications and combinations; developing the data tool into a platform-independent application; conducting pharmacogenomics studies; and developing/validating additional AKI biomarkers as well as personalized AKI detection and reduction strategies. Like the NINJA system, the NINJA team does not plan to rest until hospitalized children are no longer at risk from nephrotoxic medications.