Kidney Week

Abstract: FR-PO491

Identification and Prevalence of Pediatric CKD in a Large National Insurance Database

Session Information

• CKD: Epidemiology, Outcomes - Non-Cardiovascular - I
  November 03, 2017 | Location: Hall H, Morial Convention Center
  Abstract Time: 10:00 AM - 10:00 AM

Category: Chronic Kidney Disease (Non-Dialysis)

• 304 CKD: Epidemiology, Outcomes - Non-Cardiovascular

Authors

• Modi, Zubin J., University of Michigan, Ann Arbor, Michigan, United States

Zubin J. Modi,

• Robinson, Ian T, University of Michigan, Ann Arbor, Michigan, United States

Ian T Robinson,

• Banerjee, Tanushree, University of California, San Francisco, San Francisco, California, United States

Tanushree Banerjee,

• Powe, Neil R., University of California, San Francisco, San Francisco, California, United States

Neil R. Powe,

• Saydah, Sharon, Centers for Disease Control and Prevention, Atlanta, Georgia, United States

Sharon Saydah,

• Rolka, Deborah, Centers for Disease Control and Prevention, Atlanta, Georgia, United States

Deborah Rolka,

• Saran, Rajiv, University of Michigan, Ann Arbor, Michigan, United States

Rajiv Saran,

• Gipson, Debbie S., University of Michigan, Ann Arbor, Michigan, United States

Debbie S. Gipson,

Background

Population-level surveillance of chronic kidney disease (CKD) in pediatrics is undeveloped. Children and adolescents with CKD are currently poorly identified in large databases, as existing surveillance systems focus primarily on adults. We examined 3 different claims-based algorithms to identify children and adolescents with CKD in a large, national, single-payer insurance database.

Methods

Using the Clinformatics™ data from 2014, children and adolescents < 21 years with potential CKD were identified by 3 ICD-9 code algorithms: 585-CKD codes alone, adult CKD stage-specific algorithm and a novel pediatric-specific CKD algorithm derived from chart review (N=110) at a large academic center. Patients were included if code was used at least once during the study period. Demographics were compared between patients identified via each method and concordance between methods was evaluated.

Results

860, 8637, and 4294 children were identified via the 585-CKD, adult, and pediatric algorithms, respectively. Of all patients identified by at least 1 of the 3 methods, 44.6% were identified by both the adult and pediatric algorithms. All 860 patients identified by the 585-CKD algorithm were also identified by both adult and pediatric algorithms. Some code differences included prostatic obstruction, abnormal creatinine testing and acute renal failure in adult algorithm and specific genetic, autoimmune, and urologic disorders in pediatric algorithm. 144 patients were uniquely identified by the pediatric algorithm and 5861 were unique to the adult algorithm. Of the uniquely identified patients, those in the pediatric group were older, compared to the adult group (11 y vs 8 y). A higher prevalence of anemia was observed among patients uniquely identified by the pediatric algorithm compared to adult algorithm (43% vs 23%).

Conclusion

Adult and pediatric code algorithms examined identified substantially more potential patients with CKD than the 585-CKD algorithm. Refinement and validation of code algorithms for pediatric CKD as well as expansion to use of ICD-10 codes may be necessary to allow for current and historic case identification and pediatric CKD surveillance efforts.

Funding

• Other U.S. Government Support