Statistical Significance between pooled sensitivities of two tests

[Nitesh Gautam]

Good evening!

I am trying to do a meta analysis of two diagnostic tests, using MIDAS command. I have the pooled sensitivity and specificity of the tests A and B. I am confused as to how to compute the statistical significance between the two, i.e p value.

Here is the data for the two index tests:


Test 1:
ROC Area, AUROC = 0.78 [0.75 - 0.82]


Heterogeneity (Chi-square): LRT_Q = 0.412, df =2.00, LRT_p =0.407

Inconsistency (I-square): LRT_I2 = 0, 95% CI = [ 0-100]



Parameter Estimate 95% CI

Sensitivity 0.86 [ 0.83, 0.89]

Specificity 0.36 [ 0.31, 0.41]

Positive Likelihood Ratio 1.3 [ 1.2, 1.5]

Negative Likelihood Ratio 0.38 [ 0.29, 0.49]

Diagnostic Odds Ratio 4 [ 3, 5]

Test 2:
ROC Area, AUROC = 0.85 [0.82 - 0.88]


Heterogeneity (Chi-square): LRT_Q = 6.684, df =2.00, LRT_p =0.018

Inconsistency (I-square): LRT_I2 = 70, 95% CI = [ 33-100]



Parameter Estimate 95% CI

Sensitivity 0.82 [ 0.78, 0.86]

Specificity 0.85 [ 0.77, 0.91]

Positive Likelihood Ratio 5.5 [ 3.4, 8.8]

Negative Likelihood Ratio 0.21 [ 0.16, 0.27]

Diagnostic Odds Ratio 26 [ 13, 51]


Any help will be highly appreciated! Thanks!