Comparison of computer assisted scan diagnoses and clinical diagnoses of major mental disorders in Butajira, rural Ethiopia

One of the major difficulties in mental health research is finding suitable research instruments. In the last few decades questionnaires that are supposed to work in different languages and cultures have been developed by WHO to solve this problem. One of such instruments is SCAN, an instrument which uses computer algorithm to make a diagnosis. It is meant to be used for semi-structured interview by qualified psychiatrists or clinical psychologists. This and a few other research instruments have been translated and used in clinical and community settings for psychiatric research in Ethiopia over the last few years. In this study computer assisted SCAN Diagnoses and clinical diagnoses made by trainee psychiatrists in Butajira, a rural setting in Ethiopia, were compared Seven hundred twenty nine persons were diagnosed to have schizophrenia and bipolar disorder using both methods. The agreement between the clinical and SCAN derived diagnosis was shown to be 100% for schizophrenia (Kappa = 1.0). However, the agreement between the two methods of diagnosis was shown to be lower when it comes to subtypes of schizophrenia. The agreement for the diagnoses of bipolar disorder was 95.3%, kappa = 0.9, P < 0.0001 and for depression it was 93.0%, kappa = 0.8, p < 0.0001. Taking into account the limited number of psychiatrists and clinical psychologists in the third world countries like Ethiopia we have shown that using trainees in psychiatry for SCAN interview is a feasible and reliable method to identify major categories of mental disorders in community studies.