A decision support tool developed at Vanderbilt University Medical Center could help predict the pediatric risk of acute kidney injury (AKI). Embedded in the EHR, the tool applies statistical models to calculate AKI risk, and generates an alert to notify providers of a child’s risk status.
The tool was recently put to the test in a clinical trial, led by Sara Van Driest, M.D., assistant professor of pediatrics at Vanderbilt. “Our goal was to see if notifications in the tool can increase appropriate screening for AKI and reduce disease severity,” Van Driest said.
A Common Problem
Hospitalized children are particularly vulnerable to AKI, according to the latest research. Incidence hovers around 25 percent for critically ill children, and five percent for those outside the ICU. A diagnosis also increases the risk of chronic kidney disease while skyrocketing care costs.
Said Van Driest, “Pediatric acute kidney injury is an eventual common pathway with many causes, from the acute problems that send children to the hospital to kidney-related side effects of various drugs we use in the hospital to treat these diseases and conditions.”
“Our goal was to see if notifications in the tool can increase appropriate screening for AKI and reduce disease severity.”
ICU vs. Ward
For nearly a year, Van Driest and 13 Vanderbilt colleagues tested the tool in real-time, publishing their results in Pediatric Research. The tool was applied to over 12,000 pediatric inpatients (ages 28 days to 21 years) at Monroe Carell Jr. Children’s Hospital at Vanderbilt. Half were randomized to trigger an EHR alert – a screening suggestion to providers if calculated AKI risk exceeded 50 percent (and no serum creatinine was ordered within 24 hours). The other half received usual clinical care with no AKI alerts.
For ICU patients, the tool increased the rate of serum creatinine testing by nine percent. In the inpatient wards, there was no increase in AKI screening as a result of the tool. Incidence and severity of AKI did not vary based on whether or not the tool was used.
Alerting systems for AKI are increasingly common. They can enable early detection and change the course of care. But lessons learned from other kinds of decision support tools suggest providers may encounter over 100 clinically insignificant alerts to prevent one adverse event, and that they override the vast majority of them.
“On the pediatric wards this alert only contributed to alert fatigue.”
Identifying ineffective alerts in the new study was just as important as finding those that worked, Van Driest said. “On the pediatric wards, this alert did not work. This negative finding is equally important as the finding that it did work in the pediatric ICU, because that means that on the pediatric wards this alert only contributed to alert fatigue. This told us that this is not a good approach for the pediatric wards: we need to find a better way to let clinicians know when a pediatric patient is at risk for acute kidney injury.”