ASN's Mission

To create a world without kidney diseases, the ASN Alliance for Kidney Health elevates care by educating and informing, driving breakthroughs and innovation, and advocating for policies that create transformative changes in kidney medicine throughout the world.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on Twitter

Kidney Week

Abstract: TH-OR09

AKI Flagger: A Standardized AKI Definition Tool for Electronic Health Record (EHR) Research

Session Information

Category: Acute Kidney Injury

  • 102 AKI: Clinical‚ Outcomes‚ and Trials

Authors

  • Aklilu, Abinet Mathias, Yale School of Medicine, New Haven, Connecticut, United States
  • Saran, Ishan, Yale School of Medicine, New Haven, Connecticut, United States
  • Wilson, Francis Perry, Yale School of Medicine, New Haven, Connecticut, United States

Group or Team Name

  • CTRA (Clinical and Translational Research Accelerator)
Background

EHR data is increasingly utilized in large studies investigating in-hospital AKI. Although such studies often attempt to use guideline criteria to diagnose AKI, there is often variability in the methodologies. We have hence developed the AKI Flagger-an open-source tool to help standardize the definition of AKI in large database studies.

Methods

The AKI Flagger tool employs three techniques to operationalize the definition of AKI using the KDIGO serum creatinine(SCr) criteria for AKI: Rolling Minimum Window(RMW, using 48hour and 7day windows), Historical Baseline Trumping(HBT, median SCr within 7-365days pre-AKI) and Baseline Creatinine Imputation(BCI, imputed from eGFR 75mL/min/1.73m2). The flagger was internally validated in a cohort of 40,106 adult inpatients at 6 hospitals. We determined the sensitivity and specificity of various AKI definitions in predicting stage 2 AKI, dialysis, and death.

Results

The incidence of AKI was higher using BCI(40.0%) compared to the HBT(31.1%) and RMW(26.1%) techniques. AKI was detected earlier with BCI(median 0.9d, IQR:0.0-5.7), followed by HBT(median 2.7d, IQR:0.0-8.1), and RMW(median 4.2d, IQR:1.6-10.3). AKI diagnosed using RMW was associated with the highest percentage of dialysis(3.2 vs 2.7 vs 2.1%) and mortality(8.0 vs 7.3 vs 6.2%), and the longest hospital stay(9.01 vs 8.31 vs 8.03 days) compared to HBT and BCI, respectively. BCI had the highest sensitivity(0.93) and lowest specificity(0.69) for AKI progression with an AUC of 0.81. RMW had the lowest sensitivity(0.87) and highest specificity(0.83) for predicting AKI progression with an AUC of 0.85.

Conclusion

The incidence of AKI and related outcomes vary according to the definition and timeframe used. The AKI Flagger can be a useful tool to promote consistent reporting and reproducibility in AKI research. It is available for use without permission in R and Python.

Performance of the AKI Flagger tool in predicting AKI outcomes
TechniqueSensitivity/SpecificityAUC
AKI progression*
BCI
HBT
RMW
0.93 / 0.67
0.89 / 0.77
0.87 / 0.83
0.81
0.83
0.85
Dialysis
BCI
HBT
RMW
1.00 / 0.61
0.99 / 0.70
0.98 / 0.75
0.80
0.84
0.86

* doubling of admission SCr