The objective of this study was to assess the psychometric properties of a new instrument, the Delirium Index (DI), to measure changes in the severity of the symptoms of delirium among patients previously diagnosed with delirium. Subjects were medical inpatients aged 65 and over diagnosed with delirium by the Confusion Assessment Method. Interrater reliability of the DI was .78 between research assistants (concordence coefficient) and was .88 between research assistants and geriatric psychiatrists. Criterion validity, assessed by the correlation between DI and Delirium Rating Scale scores (Spearman's correlation coefficient, r), was .84. Construct validity was assessed using correlations of the DI with two measures of current function for convergent validity (r = −.60, −.70) and two measures of function before admission for discriminant validity (r = .26, −.42). We conclude that the DI has acceptable levels of interrater reliability, criterion validity, and construct validity.