Related Experiment Videos
Reliability of the ICD-10 classification of adverse familial and environmental factors
Godelief R W M Willemse1, Tom A Van Yperen, Jan Rispens
1Netherlands Institute of Mental Health and Addiction/Trimbos-instituut, Utrecht, The Netherlands. gwillemse@trimbos.nl
Background:
The tenth revision of the International Classification of Diseases (ICD-10) contains a number of categories and guidelines for coding different types of adverse familial and environmental situations. These categories have been selected on the basis of empirical evidence that the adverse situations might represent important psychiatric risk factors. Prior studies, using written case vignettes or video- or audiotaped semi-structured interviews, showed that the interrater reliability of these classifications is satisfactory.
Method:
We have tested the interrater reliability of the categories in a multicentre study by using an 'in vivo' design. Classifications were performed in day-to-day practice, using information that is normally available. Three hundred children from 0 to 13 years were involved in the study. These cases had been admitted to various institutes for children with psychiatric disorders, developmental delays and/or adverse psychosocial circumstances. Two clinicians and a student classified the psychosocial situation of each child. In total, 51 clinicians and 6 undergraduate students were involved in the study.
Results:
It was found that, with the exception of more or less objective categories, the reliability is not satisfactory. Post hoc analyses showed that the insufficient reliability is only partly due to confounding factors: no clear indications were found for information variance, but observation variance may have played a part. Having additional information from two family questionnaires hardly contributes to a better reliability.
Conclusions:
The reliability of the psychosocial axis of the ICD-10 is not satisfactory if tested in day-to-day practice. In comparing this result with other reliability studies, it seems that the absence of adequate information to code the psychosocial axis may be a fundamental problem in obtaining sufficient interrater agreement. Apparently, the classification of this axis requires information that is often not available in common practice.