76% of the primary diagnoses given by the expert were in the affective spectrum. Agreement concerning affective disorder (F30-F39) was moderate both concerning the whole spectrum, major depression and bipolar disorder. The only exception was for current mania, where clinicians correctly identified seven out of eight patients. As shown in Table , 16 patients (28%) of the 58 patients with a bipolar disorder did not receive an affective diagnosis (F30-F39) at all by the clinicians. This finding indicates that previous studies of patients initially presenting with depression, referred to in the Introduction, may have underestimated the problem concerning misdiagnosis of bipolar disorder. In our study, as many as 40% received a diagnosis of unipolar depression (F32-F33) instead of bipolar disorder. The most striking feature was the misdiagnosis of bipolar depression as unipolar depression (F32-F33) by the clinicians, altogether 21 patients (50%) out of 42, a finding in accordance with others [34
Primary diagnosesf given by the clinicians when the expert diagnosed bipolar disorder (N = 58)
The clinical consequences of underdiagnosing bipolar disorder were briefly accounted for in the introduction. Secondly, there are some potential administrative consequences of underdiagnosing bipolar disorder. Misdiagnosis can represent an undercommunication of the burden these patients constitute for the health care system and consequently give wrong indications concerning developmental strategies. Misleading medical statistics may cause spurious conclusions in planning and evaluation of treatment for patients [35
Third, our findings indicate that register diagnoses are dubious for research purposes and this pertains especially to affective disorders, a finding which is in accordance with the two studies reviewed by Byrne et al. [12
], both of which were blinded [13
]. Further, this is in accordance with the investigations of Baca-Garcia et al. [36
] who found diagnostic instability of psychiatric disorders in clinical practice. McConville et al. [14
] conclude that the case register was not acceptable even as a screening instrument, for the diagnoses of neurotic or affective disorders.
The discrepancies found may be due to several unresolved controversies regarding the identification and classification of bipolar disorder, supposedly due to its heterogeneity [35
]. There is an ongoing debate on the validity of the bipolar spectrum which could hamper both the adherence to and knowledge of bipolar disorders. Not asking for manic symptoms could also be due to the general phenomenon that clinicians rely on a limited number of heuristic principles that in some instances may lead to severe and systematic errors [38
]. We believe clinicians are more apt to use a heuristic top-down approach when they diagnose patients, i. e. not asking for other symptoms when the patient presents with depression. The expert who uses data from a structured clinical interview, however, employs a bottom-up approach in the diagnostic process, i. e. asking questions which at first seem irrelevant. The risk of misclassification is supposedly higher using the top-down diagnostic approach in that it relies on the diagnostic manual to confirm a clinical impression rather than to openly screen for alternative or additional diagnoses. Lack of relevant information in the patients' records is shown to be a general phenomenon affecting all diagnostic groups [40
On the other hand, diagnosing bipolar disorder is not easily ascertained due to the following reasons [25
]: (1) the typical presentation of bipolar disorder, when help is sought, is usually a depressive episode; (2) the diagnostic criteria for the depressive phase of bipolar disorder and for unipolar depression are identical in ICD-10; (3) it is not easy to ascertain previous episodes of (hypo) mania by recording patient histories because subjects often consider their manic symptoms to be normal and hence do not report hypomanic episodes as symptoms. Irritable mood may be misclassified as a depressive symptom. Manic symptoms during depressive episodes are reported to be indicative of a bipolar disorder and should be given more attention [41
]. There are several features of a depressive episode that could indicate that it belongs within the bipolar spectrum [42
]. A probabilistic approach to develop criteria for bipolar depression has been proposed [43
]. The International Society for Bipolar Disorders Diagnostic Guidelines Task Force Report proposes to distinguish between unipolar and bipolar depression in the revised versions of the DSM and ICD manuals [44
]. That could raise the awareness of bipolarity in affective disorders. Further, it is shown that diagnostic irrelevant information can affect the likelihood of a diagnosis of bipolar disorder [45
]. Mantere et al. [46
] found in their study that no previous hospitalization, lack of psychotic symptoms and the presence of rapid cycling predicted lack of bipolar I diagnosis, while no psychotic symptoms, female gender and shorter time in treatment predicted lack of bipolar II disorder. In our study the presence or absence of psychosis did influence whether a bipolar diagnosis was given in the clinic opposed to the findings of Mantere et al. [46
], but the number of patients is small (Table ).
Our study has some advantages that strengthen the validity of the results. First, a structured diagnostic interview was performed, with additional information extracted from patients' records when necessary, and second, the clinical diagnoses were blind to the expert. On the other hand, the expert never actually saw the patient such that signs and symptoms may have been missed or misinterpreted. However, the expert only scored a symptom as present if there was given a description of overt behaviour or citations from the patient in either the interview protocol or in the hospital records. Furthermore, there is always a risk that an interview that screens for all psychiatric symptoms may be overinclusive. This possible bias may result both from a "yes-saying" response style of the patient, and from a tendency of the interviewer to put weight on positive answers about signs and symptoms that are not clinically significant. Thus, there is a risk that the high number of diagnoses given by the expert is a result of response bias and scoring bias. On the other hand, the possibility that comorbidity is not diagnosed in the clinic seems more reasonable to assume. However, we do not believe that this possible bias will disturb the main findings. Our results are in accordance with those of Pinninti et al. [47
] where MINI--diagnoses were compared with clinical ones. Structured interviews are shown to be better than unstructured traditional diagnostic assessment [40
], and combining structured interviewing with a review of the medical records appears to produce more accurate primary diagnoses and to identify more secondary diagnoses than routine clinical methods or a structured interview alone [49
]. The studies reviewed by Byrne et al. [12
], where only case notes were checked and no new information added, should be regarded more as reliability studies than validity studies. Additionally, the clinicians' diagnoses were blind to the expert thus avoiding bias in either direction. The interviews were made through collaboration between different professions and among them one psychiatrist. This could be a weakness. On the other hand, it reflects clinical practice in the hospitals where not all diagnoses are set by psychiatrists. Interrater reliability can be low even if diagnoses are determined by researchers as found by Cheniaux et al. [51
]. However, to counter this, diagnoses were not formulated by the interviewers, but by one experienced researcher, PhD in clinical psychology, in our study.
Our study comprises only first time admissions, so the generalizability of the findings could be questioned. It can be argued that new patients are more difficult to diagnose than readmitted ones. On the other hand it is reported that a diagnosis of unipolar depression is frequently given following an initial diagnosis of bipolar disorder [7
]. To resolve this question more studies are needed. There were some biases in the study sample. Generally the participants were younger, had more often paid work and were more often voluntarily admitted. It was expected that they were more often of Norwegian ethnicity and had longer lengths of stay in the hospital due to the inclusion criteria and the considerations of the ethics committee. We do not think these biases have affected the results of this study. There is no reason to believe that the patients not included would be more easily or correctly diagnosed in the clinic. The opposite seems more likely.