The Montgomery-Asberg Depression Rating Scale and the Cornell Scale for Depression in Dementia: a validation study with patients exhibiting early-onset dementia.

OBJECTIVE To investigate some validity measures of Montgomery-Asberg Depression Rating Scale (MADRS) and Cornell Scale for Depression in Dementia (CSDD) in patients with early-onset dementia (EOD). DESIGN Cross-sectional design. SETTING Nursing home. PARTICIPANTS A sample of 63 inpatients with EOD. MEASUREMENTS Participants were assessed for the presence of clinical depression using the Provisional Diagnostic Criteria for Depression in Alzheimer Disease. Caregivers were administered the MADRS and CSDD, and two subscales from the Neuropsychiatric Inventory (NPI). RESULTS Depressed patients had higher scores on 6 of the 10 MADRS items and on 3 of the 19 CSDD items than nondepressed patients. Receiver operating characteristic curve analysis revealed a larger area under the curve for MADRS (0.87) than for CSDD (0.76), the difference was significant (p = 0.044). A CSDD optimal cutoff score of "5/6" yielded a sensitivity of 0.833 and a specificity of 0.567. A MADRS cutoff score of "19/20" yielded the highest sum of sensitivity (0.750) and specificity (0.843). A MADRS score of "14/15" with almost the same sum of sensitivity and specificity had a higher sum of sensitivity and a negative predictive value than "19/20." Both scales correlated with the NPI subscales depression/dysphoria (MADRS: rs = 0.70; CSDD: rs = 0.62) and apathy/indifference (MADRS: rs = 0.59; CSDD: rs = 0.50). CONCLUSION The MADRS and CSDD performed well in distinguishing depressed from nondepressed EOD patients and showed good congruent validity. The scales may be used to assess depressive symptoms in EOD. The MADRS intensity grades may be used for development or refinement of depression scales in (early onset) dementia.