P-values below 0.05 were considered statistically significant. SCID diagnoses were used as external criterion for the calculation of the sensitivity, specificity, positive likelihood ratio (LR+), negative likelihood ratio (LR−), positive predictive value (PPV), and negative predicted value (NPV) of the MDQ. In order to take into account the different proportion of MDQ positives (111/161 = 0.689) and MDQ negatives (59/214 = 0.276)

who were assessed with the SCID (Fig. 1), estimates for sensitivity and specificity were buy DAPT weighted according to these sampling fractions (Whitmore et al., 1999). In order to compare the MDQ performance using different external criteria and different MDQ versions (using only section A or sections A plus B) in a SUD population, receiver operating curve (ROC) analyses were conducted taking into account differences in sampling fractions between MDQ screen positives and MDQ screen negatives. As hypomanic episodes in DSM-IV are (by definition) not associated

with marked impairment in social or occupational functioning as required for a positive MDQ score, there might be under-detection of BD II. Therefore, analyses were repeated without the impairment criterion (section C). Finally, since substance use can mimic manic symptoms, all analyses were repeated taking into account sections D and E. After baseline (T0), 28 of the 403 included patients were excluded due to inadequate scoring of the MDQ (Fig. 1). Of the 375 remaining patients, Selleckchem Regorafenib 161 (43%) patients were MDQ positive and 214 (57%) were MDQ negative. All MDQ positives (N = 161, 43%)

and a random sample of the MDQ negatives (N = 60, 28%) were approached Idoxuridine for the second assessment (T1). A total of 50 MDQ positives (31%) were lost to follow-up due to relapse, drop-out or inability to be traced after discharge from the inpatient department. The data of one MDQ negative patient were excluded from further analyses due to a score of less than 23 on the MMSE at T1. As a result, the analyses of the operating characteristics of the MDQ included data of 111 of all 161 MDQ positives (68.9%) and 59 of all 214 MDQ negatives (27.6%). These fractions (0.689 and 0.276) were used as weighting factors in the calculations. Because the MDQ is a screening instrument that is likely to be used early in the diagnostic process, in the primary analyses MDQ data at T0 were used in the comparison with the SCID at T1. In a secondary analysis we also compared MDQ data at T1 with SCID data at T1. It should be noted, however, that the test–retest correlation of the total sum scores of the MDQ section A scores (all cases N = 170) between T0 and T1 was rather high (R = .604, p < .0001 [correlation is significant at the 0.01 level], R2 = .36): test–retest correlation of the MDQ positive cases was .455 (R2 = .21) and for MDQ negative cases .608 (R2 = .37). Mean age of all 375 eligible patients was 40.4 years (SD ± 11.