Kidney Week

Abstract: FR-PO924

Estimating Underdetection of CKD in Real-World Health Systems

Session Information

• CKD: Epidemiology, Risk Factors, Prevention - II
  November 04, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
  Abstract Time: 10:00 AM - 12:00 PM

Category: CKD (Non-Dialysis)

• 2201 CKD (Non-Dialysis): Epidemiology‚ Risk Factors‚ and Prevention

Authors

• Chu, Chi D., University of California San Francisco, San Francisco, California, United States

Chi D. Chu,

• Xia, Fang, Bayer Corporation, Whippany, New Jersey, United States

Fang Xia,

• Du, Yuxian, Bayer Corporation, Whippany, New Jersey, United States

Yuxian Du,

• Tuot, Delphine S., University of California San Francisco, San Francisco, California, United States

Delphine S. Tuot,

• Shlipak, Michael, University of California San Francisco, San Francisco, California, United States

Michael Shlipak,

• Lamprea, Julio, University of California San Francisco, San Francisco, California, United States

Julio Lamprea,

• Singh, Rakesh, Bayer Corporation, Whippany, New Jersey, United States

Rakesh Singh,

• Gualtieri, Ralph, Bayer Corporation, Whippany, New Jersey, United States

Ralph Gualtieri,

• Kong, Sheldon X., Bayer Corporation, Whippany, New Jersey, United States

Sheldon X. Kong,

• Williamson, Todd E., Bayer Corporation, Whippany, New Jersey, United States

Todd E. Williamson,

• Estrella, Michelle M., University of California San Francisco, San Francisco, California, United States

Michelle M. Estrella,

Background

Albuminuria testing is widely underused in persons at risk for CKD but is crucial to guide implementation of evidence-based treatments to prevent CKD progression and reduce cardiovascular morbidity. We aimed to estimate the extent of albuminuria underdetection due to lack of testing in a large real world US cohort of patients with hypertension or diabetes.

Methods

We used National Health and Nutrition Examination Survey (NHANES) 2007-2018 data and the Optum EHR 5PCT Database, which includes EHR data from diverse US healthcare organizations. We included persons aged ≥18 years with hypertension, diabetes, or both. Using NHANES, we developed a logistic regression model to predict albuminuria (urine albumin/creatinine ratio ≥30 mg/g) using age, sex, race-ethnicity, systolic blood pressure, diabetes, heart failure, coronary artery disease, and eGFR. Our Optum EHR study population included patients with ≥2 outpatient visits from January 1, 2017 to December 31, 2018. Among those who did not have albuminuria testing during this period, we applied the prediction model from NHANES to estimate the prevalence of albuminuria.

Results

The albuminuria prediction model had c-statistics of 0.73 for NHANES and 0.68 when applied to the subset of Optum patients with albuminuria testing. The Optum study population included 192,108 patients (mean age 60±15 years; 26% with diabetes; mean eGFR 84±21 ml/min/1.73m²). 18% had albuminuria testing (n=33,629), of whom 34% had albuminuria (n=11,525), representing 6.0% of the total study population. Among patients who had not been tested (n=158,479), the predicted prevalence of albuminuria was 15% (n=23,369). Thus, the projected proportion of patients with albuminuria who had been detected was only 11,525/34,894 (33%). In the top quintile of predicted risk, only 37% had been tested (14,033/38,421).

Conclusion

In a real-world patient population with hypertension or diabetes, we estimated that approximately 2/3 of patients with albuminuria are undetected due to lack of testing. Improving detection of CKD represents a significant missed opportunity to optimize care delivery for reducing CKD progression and its cardiovascular complications.

Funding

• Commercial Support –