. 2014 Dec 8;2014:580620. doi: 10.1155/2014/580620

Table 3.

Interobserver validation studies in lupus nephritis.

	Number of cases	Methodology and results
		Comparison of pathologic diagnoses between two pathologists
Yokoyama et al., 2004 [18]	60	Fisher exact test:
		WHO 1995 versus ISN/RPS: 83% versus 98%, P: 0.084

		κ values^*:
Furness and Taub, 2006 [19]	20	WHO 1995 versus ISN/RPS: 0.44 versus 0.53, P: 0.002
Furness and Taub, 2006 [19]	20	Acute changes: 0.39
		Chronic changes: 0.35

		ICC^**:
		WHO 1995 versus ISN/RPS: 0.182 versus 0.414
Grootscholten et al., 2008 [20]	126	Glomerular lesions: 0.439–0.950
Grootscholten et al., 2008 [20]	126	Tubulointerstitial lesions: 0.418–0.514
		Activity index: 0.716
		Chronicity index: 0.494

Wilhelmus et al., 2014 [21]	30	κ values/ICC^***:
Wilhelmus et al., 2014 [21]	(microphotographs)	presence of class III/IV lesion: 0.39

^* κ values with 95% confidence interval were calculated to represent the level of interobserver agreement (0 = no agreement and 1 = perfect agreement).

^**ICC (intraclass correlation coefficient) is an index of concordance that indicates the degree of agreement: >0.8: excellent; 0.6–0.8: good; 0.4–0.6: moderate; <0.4: poor concordance.

^***Interobserver agreement among nephropathologists was studied. Glomeruli pictures were shared with 360 members of Renal Pathology Society and they were asked whether glomerular lesions were present and compatible with class III or IV.