Using Electronic Health Records to Identify Patients With Chronic Kidney Disease
An estimated 37 million American adults—15% of the US population—have chronic kidney disease (CKD), yet 9 in 10 adults do not know they have it. People with diabetes and high blood pressure are at high risk of developing CKD; however, almost 60% of Medicare beneficiaries with both diabetes and high blood pressure are not tested for kidney disease. Many people with CKD are not diagnosed until the disease is advanced and they need dialysis (a treatment that filters the blood) or a kidney transplant to survive.
In electronic health records, individuals with CKD cannot be correctly identified using diagnosis codes alone. Thus, an expert working group of the National Institutes of Health developed and validated a set of laboratory tests and values, called e-phenotype, using data that are widely available in the electronic health record to help identify patients likely to have CKD.
The e-phenotype was implemented among adults receiving inpatient or outpatient care at five healthcare organizations (about 2 million people). A sample of about 600 patients was randomly selected at four sites for validation via blinded chart review.
The CKD e-phenotype was defined as:
- The most recent estimated glomerular filtration rate (eGFR)—a measure of kidney function—of less than 60 mL/min/1.73 m2 with another value of less than 60 mL/min/1.73 m2 at least 3 months earlier.
- A urine albumin-to-creatinine ratio (UACR)—a measure of protein in the urine that if elevated signals kidney damage—of 30 mg/g or greater in the most recent test with another value of 30 mg/g or greater at least 3 months earlier.
Diagnostic and procedure codes in electronic health records helped identify patients who received dialysis or a transplant for end-stage kidney disease (ESKD).
The validation results of the e-phenotype were:
- For CKD, 99% sensitivity, 99% specificity, and 99% positive predictive value.
- For dialysis, 94% sensitivity, 89% specificity, and 85% positive predictive value.
- For transplant, 97% sensitivity, 91% specificity, and 77% positive predictive value.
This information may be useful to health care providers to promote early detection of CKD in people, help manage CKD, and help reduce CKD-related complications such as heart disease and kidney failure.
- Centers for Disease Control and Prevention. Chronic Kidney Disease in the United States, 2019. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2019.
- United States Renal Data System. 2019 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United Statesexternal icon. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2019.
- Norton JM, Ali K, Jurkovitz CT, et al. Development and validation of a pragmatic electronic phenotype for CKD.external icon Clin J Am Soc Nephrol. 2019;14(9):1306–1314.