. Author manuscript; available in PMC: 2013 Feb 1.

Published in final edited form as: Kidney Int. 2012 Aug;82(4):445–453. doi: 10.1038/ki.2012.169

Table 2.

Precision, goodness of fit, and agreement of estimated glomerular filtration rate (eGFR) derived from coefficients of indicated regression model; N=643 (i.e., 2/3 of 965) training set children-visits of the Chronic Kidney Disease in Children (CKiD) study.

eGFR = a[height/Scr]^b [1.8/Cystatin C]^c [30/BUN]^d [e^male] [height/1.4]^f

Model

\sqrt{MSE}

R²

% of eGFR within 30% of iGFR

% of eGFR within 10% of iGFR

Univariate

42.3 ± 0.3

0.780 ± 0.016

0.184

78.5%

84.3%

40.4%

40.9 ± 0.3

0.931 ± 0.020

0.190

77.1%

84.9%

42.3%

III

41.0 ± 0.5

0.613 ± 0.024

0.280

50.2%

67.8%

26.1%

Bivariate

I & II

41.6 ± 0.3

0.443 ± 0.026

0.479 ± 0.031

0.157

84.3%

90.1%

46.5%

I & III

41.9 ± 0.3

0.662 ± 0.021

0.171 ± 0.021

0.175

80.6%

86.5%

42.8%

II & III

40.8 ± 0.3

0.796 ± 0.027

0.157 ± 0.022

0.183

78.7%

84.9%

42.8%

Multivariate I & II & III

41.5 ± 0.3

0.417 ± 0.026

0.431 ± 0.032

0.088 ± 0.019

0.155

84.8%

89.4%

47.1%

Final

39.8 ± 0.4

0.456 ± 0.026

0.418 ± 0.031

0.079 ± 0.018

1.076 ± 0.013

0.179 ± 0.032

0.147

86.3%

91.3%

48.8%

eGFR and iGFR (ml/min per 1.73m²); height (m); Scr, serum creatinine (mg/dl); cystatin C (mg/l); BUN, blood urea nitrogen (mg/dl); MSE, mean square error; entries for a – f are regression coefficient ± standard error