Abstract
The reliability of the diagnosis of bipolar‐II disorder (BP‐II) is still a problem. Semi‐structured interviews by clinicians might partly overcome this problem. The aims of this study were to find the degree of agreement in the diagnosis of BP‐II between the Structured Clinical Interview for DSM‐IV (SCID) and a semi‐structured interview based on Angst's hypomania checklist (Angst et al., 2003), and to assess the priority among hypomanic symptoms for the diagnosis of BP‐II. Remitted depression outpatients (N = 102) were interviewed during a follow‐up visit using the Structured Clinical Interview for DSM‐IV (SCID), and then with Angst's semi‐structured interview, following DSMIV criteria. Bipolar I (BP‐I) patients were excluded. Using the SCID, 29 patients were diagnosed BP‐II, 26 BP‐I, and 47 major depressive disorder (MDD). By the semi‐structured interview 69 patients were diagnosed BP‐II, 33 MDD, and none BP‐I. Agreement for the diagnosis of BP‐II between the two interviews was 53.9% (k = 0.18). Re‐analysis, after deleting the SCID question on the impact on functioning (DSM‐IV unclear boundary between BP‐I and BP‐II), increased agreement to 78.4% (k = 0.55). Elevated mood and overactivity (increased goal‐directed activity) had the lowest agreement (k = 0.46–0.49). For predicting BP‐II, overactivity had the highest sensitivity (94.2%), while elevated mood had a sensitivity of 84.0%. Multivariate analysis for predicting BP‐II (diagnosed by semi‐structured interview), including all DSM‐IV hypomanic symptoms, found that mood change and overactivity were the only independent predictors. Overactivity plus at least three symptoms (as suggested by Angst and Gamma, 2002) were present in 71 patients, of whom 91.5% also met DSM‐IV criteria for hypomania. Overactivity and elevated mood were strongly associated (but not overactivity and irritability). Findings may support a diagnosis of BP‐II based on Angst's semi‐structured interview versus the fully structured SCID interview. While DSM‐IV always requires mood change for the diagnosis of hypomania, the present findings may suggest that overactivity could have the same priority, as suggested by Angst et al. (2003) and by Akiskal et al. (1977, 2001, 2003). Copyright © 2004 Whurr Publishers Ltd.