Validation of the Center for Medicare and Medicaid Services algorithm for eligibility for dialysis

Abstract

The Center for Medicaid and Medicare Services (CMS) has recently revised their end-stage renal disease (ESRD) Medical Evidence Report, Medicare Entitlement, and Patient Registration CMS 2728 Form. The modified algorithm calls for the use of formulae to estimate glomerular filtration rate (GFR). The new criterion is defined as estimated GFR of less than 20 ml/min per 1.73 m(2). GFR is either estimated by Schwartz formula (C(SCH)) in children or Modification of Diet in Renal Disease formula (C(MDRD)) in adults. The purpose of this communication is to test the validity of the new CMS GFR algorithm in detecting children who need renal replacement therapy. We evaluated two cohorts of children. Group I included single-center data from 626 (125)I-iothalamate clearance studies (C(IO)) that were compared with the simultaneous estimation of GFR by C(SCH). Group II included data on 659 children from the patient incidence registry obtained from the ESRD Network of Texas between February 1996 and October 2003. In group I there were 76 children (76 C(IO)) with C(SCH) less than 20 ml/min per 1.73 m(2) of whom 50 (67%) had C(IO) less than 15 ml/min per 1.73 m(2). Of children with C(IO) less than 15 ml/min per 1.73 m(2), 62% had a C(SCH) less than 20 ml/min per 1.73 m(2). The ability of C(SCH) greater than 20 ml/min per 1.73 m(2 ) to predict C(IO) greater than 15 ml/min per 1.73 m(2 )(negative predictive value) is 0.95. The number of children who were started on dialysis in Texas within the study period was 659 (group II). The mean C(SCH)+/-SD was 10.8+/-7.7 ml/min per 1.73 m(2). Of the patients who were initiated on dialysis, 94% had C(SCH) less than 20 ml/min per 1.73 m(2). The results were sustained when race, gender, age range, and type of diagnosis were considered. The new CMS algorithm provides a good negative predictive estimate of GFR less than 15 ml/min per 1.73 m(2).

Validation of child serum creatinine-based prediction equations for glomerular filtration rate

“…Thus, we concur with other authors that it is more valid to evaluate equation models rather than the exact equations as published [12,17]. In combination with laboratory reporting of estimated GFR, as recommended by the K/DOQI guidelines [1], this practice could have positive effects in epidemiologic and multicenter studies, and provide more accurate GFR estimates to medical benefit and insurance companies that rely on these estimates for decisions of dialysis eligibility and compensation [24]. It is not clear what the sample size requirements are for calculating an accurate value for k. Given that the range of individual subject values of k in previous pediatric populations ranged from about 26.5 to 76.0 [9,12], we propose that a realistic goal would be to have at least 30 patients for each CKD stage (1 to 5), in order to estimate a k value applicable to groups with a wide range of GFR.…”

Serum neutrophil gelatinase-associated lipocalin as a marker of renal function in children with chronic kidney disease

“…This was especially evident for eGFR as a marker in the lower ranges of measured GFR, at which both measured serum creatinine and eGFR performed poorly. These results are expected since it is well known that the Schwartz formula can overestimate kidney function in subjects with advanced kidney failure [30]. In our subjects, the best overall agreement with measured GFR was found for serum NGAL.…”

Patterns and time of initiation of dialysis in US children

“…The largest category of missing data was for serum albumin (14.8%) followed by height, with 6.3% missing. Using gender and age-specific percentile data provided by the Center for Disease Control [7], 14.8% had a reported height above the 99.9th or lower than the 0.1st percentiles, 13.2% had a reported weight outside those percentiles, and 8.3% had a BMI outside of those percentiles. Of those with an unacceptable height, 15.0% were less than 1 year of age and 6.0% were less than 2 years old.…”

“…Centers for Medicare and Medicaid Services has recently revised their dialysis algorithm. The new criterion for eligibility for renal replacement therapy is defined as an estimated GFR of less than 20 mL min 1 per 1.73 m 2 [7]. Hence there needs to be concerted effort to initiate dialysis early.…”

“…An entry was also excluded if it had an estimated GFR value greater than 30 mL min 1 per 1.73 m 2 , or if the data needed to estimate GFR were missing. The upper limit of 30 on GFR was chosen because it captured 98% of children who were started on dialysis in the state of Texas [7]. After all exclusions, data suitable for complete analysis were available for 4,808 (67.8%) children and 570,808 (88.1%) adults.…”