Development and Validation of a Pragmatic Electronic Phenotype for CKD

Area of Science:

• Nephrology
• Biomedical Informatics
• Health Services Research

Background:

• Poor identification of chronic kidney disease (CKD) hinders research and clinical care.
• Electronic health records (EHRs) offer potential for improved patient identification.
• A pragmatic electronic phenotype is needed to identify individuals with CKD.

Purpose of the Study:

• To develop and validate a pragmatic electronic (e-) phenotype for identifying patients with CKD.
• To establish urine albumin (UA) dipstick cutoffs for CKD detection in the absence of urine albumin-to-creatinine ratio (UACR).

Main Methods:

• The CKD e-phenotype was developed by an expert group and implemented across five healthcare organizations.
• Urine albumin-to-creatinine ratio (UACR) and urine albumin (UA) dipstick results were compared to determine UA cutoffs.
• Validation involved a random sample of patients undergoing blinded chart review.

Main Results:

• The CKD e-phenotype demonstrated high accuracy: 99% sensitivity, 99% specificity, and 98% diagnostic accuracy.
• The phenotype accurately identified patients on dialysis (94% sensitivity, 89% specificity) and transplant recipients (97% sensitivity, 91% specificity).
• Established UA dipstick cutoffs allow CKD identification when UACR is unavailable.

Conclusions:

• The developed CKD e-phenotype is a pragmatic and accurate tool for EHR-based patient identification.
• This method facilitates improved research and clinical management of CKD.
• The e-phenotype enhances the ability to detect CKD, including end-stage renal disease requiring dialysis or transplant.