Background/aims: No single tool is available in geriatric clinical settings to quantitatively measure global cognitive ability at different stages ranging from normal functioning to severe impairment. We developed an adaptive test to measure cognitive ability in geriatric populations (Geriatric Rapid Adaptive Cognitive Estimate, GRACE); however, the test failed to discriminate between high-functioning individuals. This study aimed to estimate the extent to which adding more difficult items to the GRACE would improve measurement precision in the upper range of cognitive ability.
Methods: The original data used to develop the GRACE was merged with newly collected data of patients who presented at local geriatric clinics. The Rasch analysis was used to estimate the difficulty level of the newly added items and evaluate whether the psychometric properties of the GRACE were improved.
Results: One newly added item (Sequencing 5) had a higher difficulty level than all of the previous items in the GRACE. The rest of the new items were located in the high difficulty range.
Conclusion: The psychometric properties of our adaptive screening tool were improved, and we were able to distinguish between individuals who had higher levels of cognitive functioning.