Comparison of reference distributions acquired by direct and indirect sampling techniques: exemplified with the Pediatric Reference Interval in China (PRINCE) study

Table 5 Bias ratio of reference intervals of creatinine and urea establihsed by direct and two indirect sampling techniques

Age Groups	GMM Partitioned LIS		SOM Partitioned LIS		GMM Partitioned LIS with radical outlier detection strategy		GMM Partitioned LIS with radical outlier detection strategy
Age Groups	LL	UL	LL	UL	LL	UL	LL	UL
Creatinine (μmol/L)
1 to < 6 years	0.06	0.40	-0.21	0.06	0.10	0.46	-0.27	1.92
6 to < 12 years	-0.35	0.52	-0.51	1.95	-0.28	0.56	-0.63	2.39
12 to < 17 years boys	-0.07	-0.71	0.09	0.03	-0.20	-1.16	-0.07	-0.28
12 to < 17 years girls	-0.20	0.13	1.47	1.08	-0.07	0.07	-0.64	1.78
Urea (mmol/L)
1 to < 17 years	-0.09	-0.44	-0.08	-0.13	-0.07	-0.45	-0.17	-0.58

LIS Laboratory Information System, GMM Gaussian Mixture Model, SOM Self-Organizing Map, LL lower limit, UL upper limit

ISSN: 1471-2288