Skip to main content

Table 6 Different estimates [with 95% CI] of FNAC test performance according to the different methods

From: Statistical methods for evaluating the fine needle aspiration cytology procedure in breast cancer diagnosis

Handling verification bias

Without

With (Begg and Greenes)

Cell matrix

  

2×2

 

3×2

3×2

Methods

Worst case

Conventional

Best case

MICE

Conditional

Conditional

Se/ Sec

0.848

0.970

0.974

0.946

0.970

0.863

 

[0.82-0.87]

[0.96-0.98]

[0.96-0.98]

[0.93-0.96]

[0.96-0.98]

[0.84-0.884]

Sp/ Spec

0.946

0.998

0.998

0.986

0.998

0.998

 

[0.93-0.96]

[0.99-0.99]

[0.99-0.99]

[0.98-0.99]

[0.99-0.99]

[0.991-0.999]

LR+

16

570

603

67.5

570

552

 

[11.4-22.2]

[80.4-4,037]

[85-4,273]

[-]

[80.4-4,037]

[78-3911.5]

LR−

0.160

0.029

0.026

0.055

0.029

0.137

 

[0.14-0.19]

[0.02-0.04]

[0.02-0.04]

[-]

[0.02-0.04]

[0.117-0.160]

Accuracy

0.887

0.982

0.984

0.962

0.887

0.837

 

[0.87-0.90]

[0.97-0.99]

[0.98-0.99]

[0.95-0.97]

[0.871-0.902]

[0.819-0.854]

prev

0.608

0.588

0.608

0.608

0.608

0.620

 

[0.583-0.632]

[0.561-0.612]

[0.583-0.632]

[0.583-0.632]

[0.583-0.632]

[0.597-0.643]

Y+

-

-

-

-

0.875

0.889

     

[0.852-0.895]

[0.868-0.907]

Y−

-

-

-

-

0.948

0.952

     

[0.928-0.964]

[0.933-0.967]

LR ±

-

-

-

-

2.4

2.5

     

[1.7-3.5]

[1.5-3.5]

  1. Se sensitivity, Sp specificity, Sec conditional sensitivity, Spec conditional specificity, LR likelihood ratio, prev disease prevalence, Y test yield
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us