Interrater reliability of the clinical dementia rating in a multicenter trial

OBJECTTVE: To test the interrater reliability of the Clinical Dementia Rating (CDR) in a multicenter clinical trial. DESIGN: Observational study. SETTING: Training session for a multicenter trial of milameline, a direct muscarinic agonist, in the treatment of Alzheimer's disease. PARTICIPANTS: Twenty-four raters (physicians and nurses) familiar with drug trials and expert in the care of patients with Alzheimer's disease. METHODS: Independent scoring of the CDR using four videotaped CDR interviews. OUTCOME MEASURE: Interrater reliability, as tested by the Kappa statistic. RESULTS: The overall interrater reliability was 0.62. Within the CDR domains, the global kappas ranged from 0.33 +/- 0.06 to 0.88 +/- 0.06. CONCLUSIONS: The data support moderate to high overall interrater reliability but show important difficulties in the reliable assessment of early dementia.