TABLE 3.
Individual‐level concordance of mania/hypomania diagnoses based on the original and recalibrated CIDI compared to diagnoses based on the SCID in the clinical calibration sample (n = 485). a
Original CIDI | Recalibrated CIDI | ||||
---|---|---|---|---|---|
Est | (SE) | Est | (SE) | ||
Prevalence b | 3.8 | (0.5) | 4.7 | (0.7) | |
SN | 9.8 | (5.4) | 38.3 | (13.2) | |
SP | 96.3 | (0.5) | 96.2 | (0.6) | |
PPV | 6.6 | (3.3) | 21.4 | (6.6) | |
NPV | 97.6 | (0.8) | 98.3 | (0.7) | |
TCA | 94.1 | (0.9) | 94.7 | (0.9) | |
AUC | 0.53 | 0.67 | |||
|
6.6 c | 20.9 c |
Abbreviations: AUC, area under the receiver operator characteristic curve; NPV, negative predictive value; PPV, positive predictive value; SE, standard error of Est; SN, sensitivity; SP, specificity; TCA, total classification accuracy.
The calculations are based on a weighted version of the WMHQ clinical calibration sample that adjusted for the over‐sampling of respondents with CIDI diagnoses.
SCID prevalence is 2.6% (0.8).
Significant at the 0.05 level, two‐sided test.