Latent Class Analysis of Noninvasive Methods and Liver Biopsy in Chronic Hepatitis C: An Approach without a Gold Standard

Table 4

Observed and estimated frequencies and standardized residual for 16 combinations estimated by the Latent Class Analysis (LCA) model that better fits data (2LC) for diagnosis of significant fibrosis ()and cirrhosis ().

TE	APRI	ELF	LB	Significant fibrosis ()			Cirrhosis ()
				Observed	Estimated	Standardized residual	Observed	Estimated
1	1	1	1	20	17.158		4	3.973
1	1	1	0	2	2.790		10	9.516
1	1	0	1	1	3.969		0	0.236
1	1	0	0	2	0.706		0	0.688
1	0	1	1	18	19.618		3	2.956
1	0	1	0	4	5.626		7	7.642
1	0	0	1	8	5.347		1	0.176
1	0	0	0	10	9.787		5	4.813
0	1	1	1	1	1.429		0	0.356
0	1	1	0	1	0.297		2	1.119
0	1	0	0	0	0.352		0	0.021
0	1	0	1	0	0.299		2	2.090
0	0	1	1	3	2.503		0	0.265
0	0	1	0	10	9.579		10	10.173
0	0	0	1	3	3.625		0	0.016
0	0	0	0	34	33.915		73	72.960

2LC, two latent class; TE, transient elastography; APRI, aspartate-to-platelet-ratio-index; ELF, enhanced liver fibrosis; LB, liver biopsy. 0, negative; 1, positive. Latent Class Analysis considered models that data better fitted for diagnosis of significant fibrosis [ of 9.9504 ( value = 0.1268)/Bayesian information criteria = −18.6226] and cirrhosis [ of 5.6494 ( value = 0.4636)/Bayesian information criteria = −22.9237].