Psychometric properties of the schedules for clinical assessment in neuropsychiatry (SCAN-2.1)

Abstract

Background: The Schedules for Clinical Assessment in Neuropsychiatry (SCAN), the successor of the ninth version of the Present State Examination (PSE-9), is one of the latest instruments developed by the World Health Organisation for the assessment of psychiatric disorders. So far, the psychometric properties have only been established for certain sections of the instrument. The present study is the first to test the psychometric properties of SCAN-2.1 for most of the disorders covered by the SCAN, and was carried out prior to a survey conducted in the Nijmegen Health Area (the Netherlands).

Methods: Interviewers were psychology graduates with little clinical experience. Two designs were used. In one design, pairs of independent live interviews with the same respondent were compared (test-retest situation). In the other, ten videotaped interviews by experts were rated by each of the interviewers (standardized situation), and the outcomes were compared with those of the other interviewers as well as with a reference score.

Results: In the test-retest situation the kappa coefficient for diagnostic caseness was qualified as substantial (0.62) and for diagnostic categories and diagnostic groups as moderate to good (0.24 to 0.64). In the standardized situation using videotaped interviews by experts, sensitivity as well as specificity proved to be substantial to almost perfect. The agreement per interviewer with regard to the reference diagnoses ranged from 87% (diagnostic group) to 94% (diagnostic caseness). Agreement on the syndrome level (without duration and interference criteria of DSM-IV) was excellent.

Conclusions: Although the instrument is traditionally used by experienced clinicians, this study shows that less experienced (but well trained) interviewers can apply SCAN reliably. Special attention should be paid to the items without explicit interview questions, as they tend to be more sensitive to neglect than the items with interview questions.