Table 2 Sensitivity and specificity for selected thresholds, based on model *DICS and compared to results by Roberts et al. [2] (for thresholds greater or equal to 500 ng/L, Roberts et al. performed no meta-analysis)
From: Modelling multiple thresholds in meta-analysis of diagnostic test accuracy studies
New model | Roberts et al. [2] | |||
|---|---|---|---|---|
point estimate [95 % confidence interval] | point estimate [95 % confidence interval] | |||
Threshold [ng/L] | Sensitivity | Specificity | Sensitivity | Specificity |
100 | 0.94 [0.92, 0.95] | 0.63 [0.53, 0.72] | 0.95 [0.93, 0.96] | 0.63 [0.52, 0.73] |
226 | 0.84 [0.80, 0.87] | 0.84 [0.77, 0.90] | 0.85 [0.81, 0.88] | 0.86 [0.79, 0.91] |
500 | 0.64 [0.56, 0.71] | 0.94 [0.90, 0.97] | - | - |
1000 | 0.41 [0.31, 0.52] | 0.98 [0.96, 0.99] | - | - |