Table 3.
SDQ AUC for whole sample (a) and SDQ AUC restricted to those with a positive psychiatric diagnosis–sensitivity analyses (b)
| AUC [95 % CI] | |||
|---|---|---|---|
| SDQ scales parent version | Age < 10 | 11–13 | 14–16 |
| (a) | |||
| Male any ADHD/non-ADHD, n | 184/4725 | 96/2417 | 55/1762 |
| Hyperactivity/inattention | 0.92 [0.91–0.94]a,b | 0.92 [0.91–0.94]a | 0.91 [0.89–0.94] |
| Conduct problems | 0.81 [0.77–0.84]c | 0.85 [0.81–0.89]c | 0.88 [0.83–0.92] |
| Total difficulties | 0.90 [0.89–0.92] | 0.92 [0.90–0.94] | 0.91 [0.88–0.94] |
| Female any ADHD/non-ADHD, n | 39/4827 | 21/2317 | 12/1769 |
| Hyperactivity/inattention | 0.92 [0.88–0.97]a,d | 0.93 [0.87–0.99]a | 0.96 [0.94–0.99]a |
| Conduct problems | 0.82 [0.75–0.90]c | 0.83 [0.73–0.93]c | 0.86 [0.77–0.95]c |
| Total difficulties | 0.93 [0.91–0.95] | 0.92 [0.85–0.98] | 0.94 [0.89–0.99] |
| (b) | |||
| Male any ADHD/non-ADHD, n | 184/330 | 96/209 | 55/184 |
| Hyperactivity/inattention | 0.76 [0.72–0.80]a,b,¥ | 0.78 [0.73–0.83]a,b | 0.75 [0.68–0.82]a,b |
| Conduct problems | 0.57 [0.52–0.62]c | 0.63 [0.56–0.70] | 0.65 [0.57–0.73] |
| Total difficulties | 0.63 [0.58–0.68] | 0.65 [0.59–0.72] | 0.68 [0.60–0.75] |
| Female any ADHD/non-ADHD, n | 39/233 | 21/170 | 12/198 |
| Hyperactivity/inattention | 0.78 [0.70–0.85]a,b | 0.80 [0.70–0.90]a | 0.88 [0.80–0.95]a |
| Conduct problems | 0.61 [0.52–0.71] | 0.63 [0.49–0.76] | 0.66 [0.49–0.83] |
| Total difficulties | 0.64 [0.56–0.72] | 0.70 [0.57–0.83] | 0.77 [0.61–0.93] |
Venkatraman’s test for two paired ROC curves (n bootstrap replications = 2000). Superscript letters represent pairwise significantly different ROC comparisons (p < 0.05). Note that Venkatraman’s test compares the entire curves, so it can detect overall differences in performance even when confidence intervals for point estimates overlap
¥Significant (p < 0.05) Venkatraman’s test for two unpaired ROC curves (H/I subscale male vs. female)
aH/I > CD
bH/I > TD
cCP < TD
dH/I < TD