Structured interviews for assessing children.

Structured interviews provide a valuable means of obtaining and quantifying information about the mental status of children. This review indicates that children can reliably self-report and that the information they provide can concur with the opinion of adults knowledgeable about them. However, considerably more research is warranted before it may be assumed that these interviews are adequately reliable and valid. In general, it appears that the task of documenting the psychometric soundness of these interviews has not been taken seriously, as if content validity were sufficient. For example, except for the CAS and the DISC, there has been little effort to study contrast group validity (i.e. whether the interview even differentiates "known groups"). More specifically, review of the reliability and validity data relevant to DSM-III-R diagnoses provides support for the CAS, DICA, ISC and K-SADS, with the validity data for the DICA being weaker than for the others. One limitation of these data for the DICA, ISC and K-SADS is that the diagnoses were clinician-generated, rather than algorithm-generated. Unfortunately the processes for generating clinical diagnoses were not specified, except for criterion reference to DSM-III-R. The findings for the DISC and the DISC-R are notably weak. There is no evidence for DISC reliability, except for adolescents, and the validity studies have demonstrated only weak relationships. There has been limited study of the psychometric properties of symptom scales. In fact, for two interviews (i.e. DICA and ISC), there are no data available. Reliability for the DISC scales is adequate only for adolescents. Psychometric data have been generated for the CAS and the K-SADS, with considerably more studies conducted with the CAS. The relative paucity of interest in scale scores is striking given that they provide a continuous variable which can indicate extent of symptoms. Other measures of mental status, besides presence/absence of diagnosis, will become increasingly important as research in child psychopathology progresses toward more sophisticated studies (i.e. treatment effects, risk factors). These interviews are labor intensive and costly to the researcher as well as time-consuming and tedious for the children and parents. Given this commitment, researchers should invest in developing other ways of exploiting the richness of the data generated. An example is the CAS "content" scales, which generate scores reflecting on the child's functioning in various areas (e.g. school, friends, family). As the evolution of these interviews continues, it will be important to remain attentive to the developmental limitations of children.(ABSTRACT TRUNCATED AT 400 WORDS)