A random-effects model was used to pool relapse proportions and the AUCs, weighted by their inverse variance. Summary AUCs were used to estimate the pooled risk differences (RDs) and the numbers-needed-to-treat (NNT) (formulae given).
A random-effects model was used to pool relapse proportions and the AUCs, weighted by their inverse variance.Summary AUCs were used to estimate the pooled risk differences (RDs) and the numbers-needed-to-treat (NNT)(formulae given).