Objectives: The aim of this population-based study was to investigate the usefulness of the Geriatric Depression Scale 15-item version (GDS-15) to assess depressive symptoms among very old people with differing levels of cognitive function.
Methods: The 834 participants were aged 85 and over. Feasibility of GDS-15 was evaluated as the proportion of people who completed the scale. Concurrent criterion validity was evaluated by calculating correlations between GDS-15 and Philadelphia Geriatric Center Morale Scale (PGCMS). PGCMS measures psychological wellbeing which is closely related with depressive symptoms. Correlations were calculated within groups according to cognitive function assessed with Mini-Mental State Examination (MMSE); 0-4, 5-9, 10-14, 15-19, 20-24, 25-27, and 28-30, using Pearson's two-sided correlation and compared using Fisher r-to-z transformation. Internal consistency of the GDS-15 was evaluated by calculating Cronbach's α in each group.
Results: In total, 651 (78%) of the 834 participants completed the GDS-15. For the two MMSE-groups with scores of <10, the proportion who completed GDS-15 were 1% and 42%, respectively, compared to 65-95% in the MMSE-groups with scores of ≥10. Cronbach's α in each MMSE-group ranged from 0.636 (MMSE 28-30) to 0.821 (MMSE 5-9). The level of correlation between GDS-15 and PGCMS did not significantly differ between MMSE-groups with scores of 5-27 compared to the MMSE-group with scores of 28-30.
Conclusions: The GDS-15 seems to have an overall usefulness to assess depressive symptoms among very old people with an MMSE score of 10 or more. More studies are needed to strengthen the validity of GDS-15 among older people with MMSE scores of 10-14. For older people with MMSE scores lower than 10, there is a need to develop and validate other measurements.