RESUMEN
BACKGROUND: Early diagnosis of schizophrenia can improve its outcome. Hence, screening policies should be held and suitable tools must be available for general practitioners. AIM: To translate the section G (about schizophrenia) of the Composite International Diagnostic Interview to the Tunisian dialect and to check its validity and reliability. METHODS: The Arabic version of section G of the Composite International Diagnostic Interview has been translated to the Tunisian dialect by psychiatrists speaking fluently both languages. Metric features of the instrument (sensitivity, specificity and predictive values) were assessed by checking its results against those of the gold standard i.e. the expert's opinion owing to the criteria of the Diagnostical and Statistical Manual of mental disorders 3rd revised edition. Reliability has been measured by the index of observer agreement. RESULTS: The instrument showed a low sensitivity of 45% [32% - 58%] and a high specificity of 96% [93% - 99%]. It looked as a diagnostical test that can avoid wrong diagnoses of such a serious and stigmatizing illness. These values are similar or even better than those of literature. The observer agreement index was 0.83 showing a very good reliability. The interviews mean duration was 20 minutes. The instrument showed no variability towards the sex, the age or the educational level of interviewees. CONCLUSION: Even though section G of the Composite International Diagnostic Interview failed to detect most cases with schizophrenia and showed a poor sensitivity, this instrument can be useful for screening strategies carried out by lay interviewers in the general population.