Table 2.

Estimation of prevalence and rater-specific sensitivity and specificity for digital radiography with no gold standard (GS) and with the consensus rating as the gold standard

	Estimator		CI		GRE		FM
Rater			GS	No GS	GS	No GS	GS	No GS
	Pd	Est	.24	.18	.24	.22	.24	.22
		SE	(.04)b	(.04)	(.04)	(.07)	(.04)	(.07)
1	SENS	Est	.67	.88	.66	.77	.67	.78
		SE	(.09)	(.11)	(.09)	(.16)	(.09)	(.11)
	SPEC	Est	.99	.99	.99	1.00	.99	1.00
		SE	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
2	SENS	Est	.78	.89	.77	.80	.78	.81
		SE	(.08)	(.07)	(.08)	(.14)	(.08)	(.08)
	SPEC	Est	.87	.85	.87	.86	.87	.86
		SE	(.04)	(.04)	(.04)	(.05)	(.04)	(.04)
3	SENS	Est	.52	.68	.51	.57	.52	.57
		SE	(.10)	(.13)	(.10)	(.14)	(.10)	(.13)
	SPEC	Est	.99	.99	.99	.97	.99	.99
		SE	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
4	SENS	Est	.81	1.00	.82	.92	.81	1.00
		SE	(.07)	(.05)	(.07)	(.14)	(.08)	(.01)
	SPEC	Est	.97	.95	.97	1.00	.97	.99
		SE	(.02)	(.03)	(.02)	(.03)	(.02)	(.03)
5	SENS	Est	.85	1.0	.86	.92	.85	.92
		SE	(.07)	(.05)	(.07)	(.14)	(.07)	(.04)
	SPEC	Est	.72	.71	.72	.72	.72	.72
		SE	(.05)	(.05)	(.05)	(.06)	(.05)	(.05)
6	SENS	Est	.89	.94	.88	.84	.89	.85
		SE	(.06)	(.07)	(.07)	(.09)	(.06)	(.06)
	SPEC	Est	.90	.85	.89	.86	.90	.86
		SE	(.03)	(.04)	(.03)	(.05)	(.03)	(.04)

NOTE: Models were fit under the conditional independence (CI), finite mixture (FM), and Gaussian random effects (GRE) models^a using Iinuma et al.'s data.

^aThere are 13 (2J + 1) parameters for the CI model and 15 (2J + 3) parameters for the FM and GRE models.

^bStandard errors were estimated using a bootstrap with 1,000 bootstrap samples.