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: FR-PO188

Development and Validation of an Electronic (e) Phenotype for CKD

Session Information

Category: CKD (Non-Dialysis)

  • 1901 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention

Authors

  • Norton, Jenna M., National Institute of Diabetes and Digestive and Kidney Diseases, BETHESDA, Maryland, United States
  • Jurkovitz, Claudine T., Christiana Care Health System, Newark, Delaware, United States
  • Kiryluk, Krzysztof, Columbia University, New York, New York, United States
  • Park, Meyeon, UCSF, San Francisco, California, United States
  • Kawamoto, Kensaku, University of Utah, Sandy, Utah, United States
  • Shang, Ning, Columbia University, New York, New York, United States
  • Ali, Kaltun H., National Institute of Diabetes and Digestive and Kidney Diseases, BETHESDA, Maryland, United States
  • Narva, Andrew S., National Institute of Diabetes and Digestive and Kidney Diseases, BETHESDA, Maryland, United States
  • Drawz, Paul E., University of Minnesota, Minneapolis, Minnesota, United States
Background

Identifying CKD patients is an essential step for surveillance, research recruitment and quality improvement (QI). Using diagnostic codes to identify CKD patients is challenged by widespread under-diagnosis of CKD. An electronic CKD phenotype based on data widely available in the electronic health record could facilitate identification of patients likely to have CKD.

Methods

A working group of patients, nephrologists, primary care providers and informaticists developed the ePhenotype. Five clinical sites implemented the ePhenotype. Each site collected study population demographics (age, race, sex), labs (eGFR, UACR, UPCR, UA) and dialysis/transplant status. The ePhenotype determined CKD status (eGFR <60 ml/min/1.73m2, UACR ≥30 mg/g), CKD stage and chronicity (prior lab value indicative of CKD 90+ days prior). Four sites conducted a blinded, manual validation on a random subsample of the population across CKD status. Diagnostic accuracy (proportion of patients with correctly identified CKD stage) and sensitivity/specificity of the ePhenotype for identifying CKD, dialysis and transplant were calculated for each site and overall.

Results

The validation population included 1,680,334 patients across 4 sites with average age 49.8±18.5. Of these, 58.7% were female, 10.4% were black, 59.3% had at least 1 eGFR and 40.6% had any proteinuria measurement. The proportion of patients with any proteinuria measurement varied across sites (20.2% to 52.1%) and increased as eGFR decreased. The ePhenotype was successfully implemented at all sites. Diagnostic accuracy for identifying CKD stage was 98%. Sensitivity and specificity, respectively, across the validation sites were 99.3% and 98.5% for CKD, 94.5% and 90.7% for dialysis and 97.3% and 91.1% for transplant (Table 1).

Conclusion

The ePhenotype was successfully implemented at multiple sites with a high degree of accuracy and has the potential to facilitate identification of patients with CKD for surveillance, research and QI.

ePhenotype Performance
SiteCKDDialysisTransplant
Population sizeSensitivity
(95% CI)
Specificity
(95% CI)
Population sizeSensitivity
(95% CI)
Specificity
(95% CI)
Population sizeSensitivity
(95% CI)
Specificity
(95% CI)
Total20699.3%
(96%, 100%)
98.5%
(92.1%, 100%)
23194.5%
(87.6%, 98.2%)
90.7%
(84.6%, 95.0%)
16097.3%
(85.8%, 99.9%)
91.1%
(84.6%, 95.5%)
Minnesota58100%
(89.7%, 100%)
95.8%
(78.9%, 99.9%)
5897.8%
(88.2%, 99.9%)
92.3%
(64%, 99.8%)
58100%
(69.2%, 100%)
97.9%
(88.9%, 99.9%)
Christiana Care70100%
(92.9%, 100%)
100%
(83.2%, 100%)
7292.3%
(64%0, 99.8%)
98.3%
(90.9%, 100%)
72100%
(54.1%, 100%)
98.5%
(91.8%, 100%)
Columbia60100%
(91.2%, 100%)
100%
(83.2%, 100%)
8086.4%
(65.1%, 97.1%)
91.4%
(81.0%, 97.1%)
8094.1%
(71.3%, 99.9%)
90.5%
(80.4%, 96.4%)
UCSF1892.9%
(66.1%, 99.8%)
100%
(39.8%, 100%)
21100%
(71.5%, 100%)
40%
(12.2%, 73.8%)
10100%
(39.8%, 100%)
50%
(11.8%, 88.2%)

Funding

  • NIDDK Support