In this world, we focus on the meta-analysis of I diagnostic test studies. Each study is usually reported in a 2×2 table assuming a perfect reference test,
True Positive(TP) | False Positive(FP) |
False Negative(FN) | True Negative(TN) |
Full Bayesian inference using integrated nested Laplace approximations (INLA) is performed on a hierarchical model with 3 levels, for example, the model for analysing sensitivity and specificity jointly with a logit link function is
1. Level
2. Level
Arends et al.(2008) "Bivariate random effects meta-analysis of ROC curves." Medical Decision Making, 28(5), 621-638.
Step 1: Upload your data or choose one example dataset.
Step 2: Choose the priors you want or use the default priors.
Step 3: Press RunINLA button to run the model.
Step 4: Check result.
Note: Please press the RunINLA button to update results if any settings changed in Prior Control Panel or Model Control Panel.
Meta4diag is an R-INLA project. © 2016 INLA group.