Table 3 |
||||
|
Spearman correlation coefficient Logit (sensitivity) vs Logit (1- specificity) |
||||
|
Test |
Rifampicin |
Isoniazid |
||
|
|
||||
|
Spearman correlation coefficient |
p-value |
Spearman correlation coefficient |
p-value |
|
|
|
||||
|
Nitrate Reductase Assay |
0.400 |
0.600 |
1.000 |
0.000 |
|
MODS |
0.086 |
0.872 |
0.143 |
0.787 |
|
Genotype® MTBDR |
0.500 |
0.667 |
1.000 |
0.000 |
|
Genotype® MTBDRplus |
-0.200 |
0.747 |
-0.100 |
0.873 |
|
|
||||
|
MODS = Microscopic Observation Drug Susceptibility. Note: A strong positive Spearman correlation coefficient between the logit of sensitivity and logit of 1-specificity suggests threshold/cutoff effect [27]. Since pooling sensitivity and specificity is more reliable in the absence of a threshold effect the NRA and the Genotype® MTBDR assays should be primarily judged based on their areas under sROC, while the other tests can be reliably judged based on their pooled values. |
||||
|
Bwanga et al. BMC Infectious Diseases 2009 9:67 doi:10.1186/1471-2334-9-67 |
||||