Reliability study on the Japanese version of the Clinician's Interview-Based Impression of Change

The Japanese version of the Clinician's Interview-Based Impression of Change (CIBIC-plus J) is a semistructured interview format including a 7-point Likert type scale for the Clinician's Global Impression of Change. It comprises subscales of DAD for ADL, Behave-AD for behavioral and psychiatric symptoms and MENFIS for cognitive and emotional impairment. DAD and Behave-AD are used for caregivers and MENFIS for both caregivers and patients. The objective of this study was to examine the interrater reliability of the CIBIC-plus J by the videotape method. Twenty videotaped patients including 13 real and 7 virtual patients were independently assessed by 11 raters. The kappa coefficient from the full set of data of 20 patients and 11 raters was 0.453, and from the limited data of 13 real patients, 0.383. By permitting one-point disagreements on the 7-point scale or by collapsing it into the 3-point scale (improved, no change, or worsened), kappa improved to 0.894 and 0.731, or 0.734 and 0.504, respectively. From these results, it was concluded that the reliability of a 7-point Clinician's Global Impression of Change was halfway, i.e., neither sufficient nor insufficient. However, if one-point disagreements or collapsing into 3-point scale are allowable, the CIBIC-plus J will be sufficiently reliable.