Speaker
F. Perry Wilson, MD MSCE
Associate Professor of Medicine and Public Health
Director, Clinical and Translational Research Accelerator
Yale University
New Haven, CT
Keywords
Electronic Alerts; Acute Kidney Injury; Action Team
Key Points
- Acute Kidney Injury (AKI), or abrupt decline in kidney function, is common, affecting about 15% of hospitalized patients. A hospitalized patient with AKI has an inpatient mortality rate 8.5% higher than average. However, early recognition and nephrologist involvement can improve clinical outcomes.
- First, the research team conducted a multicenter, parallel-group, randomized controlled trial (RCT) to test the effect of an electronic alert system on best practice utilization and three clinical outcomes (progression of AKI, dialysis, and death). The pragmatic trial, Electronic Alerts for Acute Kidney Injury Amelioration AKA ELAIA-1, instituted “best practice alerts” in the electronic medical record.
- The Grand Unified Theory of Electronic Alerts states that alerts can’t work if 1) the provider already knows what’s wrong with the patient; 2) they don’t care about what’s wrong with the patient; 3) they have no specific action to take in response; or 4) the action doesn’t matter – i.e., it doesn’t change outcomes.
- For ELAIA-2, the second iteration of the trial, the researchers focused on one tenet of this theory: alerts should be tied to actions. This open-label, parallel group RCT used alerts to encourage cessation of kidney-relevant medications – NSAIDs, RAASi, and PPI. They looked at rates of cessation of one of those medications and clinical outcomes (progression of AKI, dialysis, and death).
- Though they found that automated alerts for AKI can increase medication cessation, there was limited evidence that these alerts would change clinical outcomes. The rate of discontinuation was highest for PPIs, which are an under-recognized contributor to AKI. Alerts may be beneficial in physicians whose patients are receiving PPIs – a population that tends to be sicker.
- These findings led to a new hypothesis: AKI is heterogenous, caused by many factors. The research team sought to customize recommendations given to providers. To this end, they created a Kidney Action Team (KAT) with the goal of improving in-hospital mortality and AKI progression.
- The KAT-AKI trial, a multicenter RCT administered across 2 hospital systems, looked at the proportion of recommendations implemented in 24 hours and clinical outcomes (progression of AKI, dialysis, and death). 34% of KAT recommendations were implemented within 24 hours in the intervention arm compared to 24% in usual care. However, this also did not have an effect on clinical outcomes.
- The research team concluded that more personalized AKI alerts could potentially lead to better outcomes.
Discussion Themes
An optimized user interface – e.g., delivering information at the right place in a provider’s work flow, catching them at the right time – may further increase the adoption of best practices.
There’s some doubt about the nephrotoxicity of RAAs. In the ELAIA-2 study, though there were higher rates of RAA discontinuation in the intervention group, the clinical outcomes were nearly identical.
The heterogeneity of AKI patients interfered with the intervention’s effectiveness. To improve clinical outcomes, researchers may want to devote resources to prevention rather than response.