Alert System Reduces Odds of Overlooked AKI Events among Hospitalized Patients

Decline in kidney function and increased risk for mortality are closely associated with acute kidney injury (AKI). To date, there is no universal treatment for AKI, although several therapeutic and preventive interventions have been developed. It is widely accepted that early detection of AKI events is key to improving outcomes; the importance of nephrologist care is also well known in this patient population.

AKI is primarily defined by criteria based on serum creatinine (Scr). There have been efforts to develop an efficient surveillance system for AKI (AKI alert); however, negative results were reported from a published randomized trial of AKI alerts. Other studies have demonstrated promising preliminary outcomes, but have carried several limitations to the system.

Researchers in South Korea, led by Sehoon Park, MD, reported results of an AKI alert system implemented in their hospital in 2014 [American Journal of Kidney Diseases; 2018;71(1):9-19]. In the system reported on, the attending clinicians could easily generate automatic direct consultation with the nephrology division. The current report assessed the impact of the system using comparisons of outcomes of patients with AKI events prior to and following implementation of the alert system.

The study was conducted at a tertiary referral hospital in Korea with >1000 general admission beds. The AKI alert system was launched on June 1, 2014. There were no other major changes in activities related to AKI or in laboratory procedures related to AKI.

The study cohort included index admission cases of adult patients who had measurements of Scr during hospitalization in the first year of the alert system. The historical cohort included index admission cases with the same criteria who had been admitted from January 1, 2013, to December 31, 2013, prior to introduction of the system. Patients in the study group who developed AKI were included in the alert group, and patients in the historical cohort with AKI events were included in the usual-care group.

The primary outcomes of interest were overlooked AKI events (defined as not measuring the follow-up creatinine), and the consultation pattern of clinicians. Secondary outcomes were severe AKI events, AKI recovery (defined based on the creatinine-based criterion), and patient mortality.

Exclusion criteria included ongoing renal replacement therapy, impending end-stage renal disease (defined as baseline estimated glomerular filtration rate <15 mL/min/1.73 m2), and admission to the nephrology unit; death events on the day of AKI development were also excluded, and patients who were already enrolled in the historical cohort were excluded from the alert group.

The baseline Scr concentration was the minimum Scr concentration within 2 weeks of admission. In cases with no earlier laboratory value available, the baseline value was the first concentration measured during hospitalization. The system screened AKI events every midnight. AKI events were defined as Scr concentration elevation of at least 1.5-fold or 0.3 mg/dL from baseline. When physicians opened the patient’s electronic medical record (EMR) the following morning, a pop-up window displayed with this message: (Mild/Moderate/Severe) acute kidney injury (stage X). Do you want to send a consultation request to nephrology division?

If the physician clicked yes on the display, a nephrology consultation was automatically generated. Options if the physician clicked no on the display included a request for a consultation with a more detailed description, a request for a later consultation, or no request for a consultation. Nephrologists at the center were encouraged to respond to all requests within 1 day.

Following application of exclusion criteria, the final overall study cohort included 1844 in the usual-care group and 1309 in the alert group. At baseline, there were significant differences between the two groups. Patients in the alert group were older and more likely to be male and to have had surgery during hospitalization (all P<.001). Baseline kidney function was also poorer in the alert group compared with the usual-care group (P<.001), and patients in the alert group had lower concentrations of hemoglobin (P=.002) and albumin (P<.001).

There were also differences in medical histories. Those in alert group more commonly had history of hypertension (P<.001), diabetes mellitus (P<.001), ischemic heart disease (P=.001), and heart failure (P<.001) compared with the usual-care group. Patients in the alert group also more commonly used a diuretic (P<.001). Nonsterodial anti-inflammatory drug use was more common in the usual-care group. Patients in the alert group developed AKI more quickly and had longer hospitalization stay following the AKI event compared with the usual-care group (P<.001 for both).

Following implementation of the alert system, the odds of overlooked AKI events were lowered significantly (adjusted odds ratio [OR], 0.49; 95% confidence interval [CI], 0.30-0.52). In addition, the odds of a consultation with a nephrologist were significantly higher (OR, 6.13; 95% CI, 4.80-7.82). The odds of a severe AKI event were reduced following alert system implementation (adjusted OR, 0.75; 95% CI, 0.64-0.89), and the likelihood of recovery from the AKI event was improved in the alert group compared with the usual-care group (adjusted hazard ratio [HR], 1.70; 95% CI, 1.53-1.88). There was no effect on mortality in the alert group compared with the usual-care group (adjusted HR, 1.07; 95% CI, 0.68-1.68).

The researchers cited some limitations to the study, including (1) the study design and lack of randomization, (2) the inability of the AKI alert system to report AKI events in real time, (3) the lack of criteria for baseline Scr concentrations, (4) the possibility of selection bias because the alert group was admitted to the hospital for longer periods and had worse baseline characteristics, and (5) the possibility that the results may not be transferable to other hospitals, depending on size and location.

“In conclusion, our EMR-based AKI alert system altered the behavior of clinicians, increased the involvement of specialists, and improved AKI outcomes. Therefore, adoption of an AKI alert system linked to early nephrology intervention could be considered in hospitals to improve patient prognosis,” the researchers said.

Takeaway Points

  1. Researchers in South Korea developed and tested a system utilizing electronic medical records to provide early alerts of acute kidney injury (AKI) in hospitalized patients.
  2. The system alerted physicians to AKI based on concentrations of serum creatinine and provided an option to call for a consultation with the nephrology unit of the tertiary care center.
  3. The alert system resulted in lower odds of overlooking AKI events, increased odds of an early consultation with a nephrologist, and a reduction in the odds of a severe AKI event among patients in the alert system group.