Statistical methods to correct for verification bias in diagnostic studies are inadequate when there are few false negatives: a simulation study

Table 1 Examples of data subject to verification bias and with a low number of false negatives

Example	Number with normal imaging result and biopsied			Corrected Sensitivity* (%)
	Diseased (False negatives)	Nondiseased (True negatives)	Total
1	5	35	40	57
2	4	36	40	63
3	3	37	40	69
4	2	38	40	77
5	1	39	40	87
6	0	40	40	100

ISSN: 1471-2288