Table 10.

Estimation of overall and rater-specific sensitivity and specificity as well as prevalence for digital radiography using partial verification designs

Rater	Estimator	r = .10		r = .20		r = .50		r = .80
		FM	GRE	FM	GRE	FM	GRE	FM	GRE
Overall	Pd	.21 (.05)	.22 (.07)	.22 (.05)	.23 (.06)	.23 (.04)	.24 (.05)	.24 (.04)	.24 (.04)
	SENS	.84 (.08)	.80 (.14)	.82 (.08)	.78 (.11)	.78 (.07)	.76 (.07)	.76 (.06)	.76 (.06)
	SPEC	.90 (.02)	.89 (.02)	.90 (.02)	.90 (.02)	.91 (.02)	.90 (.02)	.91 (.01)	.91 (.01)
	Pd	.20 (.05)	.23 (.06)	.21 (.05)	.22 (.05)	.23 (.04)	.23 (.04)	.24 (.04)	.24 (.04)
1	SENS	.81 (.13)	.75 (.16)	.76 (.13)	.72 (.15)	.71 (.11)	.69 (.11)	.68 (.10)	.67 (.10)
	SPEC	.99 (.01)	1.00 (.01)	.99 (.01)	1.00 (.01)	.99 (.01)	.99 (.01)	.99 (.01)	.99 (.01)
2	SENS	.85 (.09)	.76 (.13)	.83 (.09)	.76 (.12)	.80 (.09)	.76 (.10)	.79 (.08)	.77 (.09)
	SPEC	.86 (.04)	.86 (.05)	.87 (.04)	.86 (.05)	.87 (.04)	.87 (.04)	.97 (.04)	.87 (.04)
3	SENS	.62 (.13)	.55 (.14)	.57 (.14)	.53 (.14)	.54 (.11)	.52 (.11)	.52 (.11)	.51 (.11)
	SPEC	.99 (.01)	.99 (.01)	.99 (.01)	.99 (.01)	.99 (.011	.99 (.01)	.99 (.01)	.99 (.01)
4	SENS	.96 (.07)	.90 (.14)	.93 (.08)	.88 (.13)	.87 (.09)	.85 (.10)	.83 (.08)	.83 (.08)
	SPEC	.97 (.03)	.98 (.03)	.97 (.03)	.98 (.03)	.97 (.02)	.97 (.02)	.97 (.02)	.97 (.02)
5	SENS	.95 (.07)	.87 (.12)	.92 (.08)	.86 (.11)	.88 (.08)	.86 (.09)	.86 (.06)	.85 (.08)
	SPEC	.72 (.05)	.72 (.06)	.72 (.05)	.72 (.05)	.72 (.05)	.72 (.05)	.72 (.05)	.72 (.05)
6	SENS	.91 (.08)	.82 (.13)	.90 (.08)	.83 (.11)	.88 (.07)	.85 (.09)	.89 (.06)	.87 (.07)
	SPEC	.86 (.04)	.86 (.05)	.87 (.04)	.87 (.05)	.88 (.04)	.88 (.04)	.89 (.03)	.88 (.04)

NOTE: Individuals were resampled with replacement to obtain a resampled dataset of 113 patients. Verification was done completely at random with probability r. One thousand resampled datasets were obtained and means (SE) across these datasets are presented. Both the GRE and the FM models were used for estimation.