Psychometric development of an instrument for the diagnosis and assessment of anosognosia in Alzheimer's disease

The purpose this study was to develop a psychometrically sound paper-n-pencil questionnaire for the measuring and diagnosing of anosognosia in Alzheimer's disease (AD). Anosognosia is defined as the lack of awareness one has towards one's own state. It manifests within AD as an unawareness of symptoms the individual is experiencing. The initial 43-item questionnaire was administered to 67 AD patients (age: u = 72.66, SD = 3.40), with 41 females and 26 males. A Cronbach's-alpha of 0.89 was obtained showing the questionnaire had excellent internal reliability. The 43 items in the original questionnaire were reduced to 10 using the internal reliability analysis. The 10-item questionnaire was administered to a new group of 83 AD patients (age: u = 75.83, SD = 3.83), with 58 females and 25 males. Internal reliability of the new questionnaire remained high with an obtained Cronbach's-alpha of 0.87.Correlations between the sample population 10-item questionnaire score and the Mini Mental State Exam (r = -0.24, p < 0.05) and Geriatric Depression Scale (r = -0.30, p < 0.05) showed a low but significant correlation. The 8-Point Clock Drawing (r = - 0.04, p > 0.05), and Zarit's Caregiver Burden Scale (r = 0.04, p > 0.05) showed no correlation. Using 19 of the patients a one-way Intraclass Coefficient (ICC) was used to determine inter-rater agreement (alpha = 0.63). Twenty-one of the patients were used for the purpose of test-retest and resulted in a Pearson-r correlation of r = 0.70 (p < 0.000). Forty-three normal subjects were enrolled in the study (age: u = 73.95, SD = 3.90) with 23 females and 20 males. Using the normals mean + 2SD a cutoff score of 12 was obtained as the point where an AD patient was diagnosed as having anosognosia. Using the cutoff value there were 42 AD patients who had anosognosia which was 51% of the sample population.The questionnaire was found to be reliable though further studies would be needed to confirm the results by expanding the sample size and using more generalized inclusion criteria. Nevertheless, the questionnaire showed little relationship to the other questionnaires administered during the study. This helps to show the questionnaire is measuring a unique phenomenon which is not related to other standard diagnostic questionnaires used with AD patients.