Table 1.
Psychometric properties of the AD8
| Psychometric property | Definition | AD8 characteristics |
|---|---|---|
| Concurrent validity | Correlation of the AD8 compared with ‘gold standard’ measures of dementia presence and severity (the CDR) | R = 0.75 (95% CI 0.63–0.88) |
| Construct validity | Correspondence of how well the AD8 measures theorized domains corresponding to CDR and neuropsychological tests. Demonstration of strong correlations (≥40%) support convergent validity, while low correlations (≤30%) indicate no relationship | R > 0.4 for memory, executive function, and activities of daily living; R < 0.3 for semantic memory and language function. |
| Internal consistency | Degree to which the AD8 is free from random error | Cronbach’s α = 0.86 (95% CI: 0.82–0.91) |
| Intra-rater reliability | Degree to which an instrument yields stable scores over time for the same respondent | Weighted κ = 0.67 (95% CI: 0.59–0.75) |
| Inter-modal reliability | Reproducibility of the AD8 across different modes of administration (in person versus telephone) | Weighted κ = 0.65 (95% CI: 0.57–0.73) |
| Inter-rater reliability | Percent agreement between two raters | Intraclass correlation coefficient = 0.82 (95% CI 0.55–0.92) |
| Discriminability | Effectiveness of the AD8 in classifying a CDR score of 0 (non-demented) versus a CDR score > 0.5 (demented) for different demographic characteristics, MMSE scores, CDR stages and clinical diagnoses | Area under the curve = 0.92 (95% CI 0.88–0.95) |
AD8 characteristics are adapted from Galvin et al., 2006.