New cognitive screening test seems more sensitive to detection of AD, easier to administer than MMSE, data suggest

A new cognitive screening test developed by researchers at Addenbrooke's Hospital in the United Kingdom is easier to administer than the Mini-Mental State Examination (MMSE) and has greater sensitivity for the detection of Alzheimer's disease (AD), according to findings from a cross-sectional study. The researchers designed the tool, which is called Test Your Memory (TYM), to be suitable for widespread use, even by nonspecialists. Specifically, they wanted to create a screening test that...

A new cognitive screening test developed by researchers at Addenbrooke's Hospital in the United Kingdom is easier to administer than the Mini-Mental State Examination (MMSE) and has greater sensitivity for the detection of Alzheimer's disease (AD), according to findings from a cross-sectional study.

The researchers designed the tool, which is called Test Your Memory (TYM), to be suitable for widespread use, even by nonspecialists. Specifically, they wanted to create a screening test that requires minimal operator time to administer, assesses a reasonable range of cognitive functions, and is sensitive to the detection of mild AD. Although the MMSE has been the standard short cognitive test for decades, it does not satisfy these 3 critical requirements, they noted.

The TYM is filled out by patients, with healthy control subjects completing it in approximately 5 minutes, according to data from the current study. It includes 10 tasks (orientation, ability to copy a sentence, semantic knowledge, calculation, verbal fluency, similarities, naming, 2 tasks evaluating visuospatial abilities, and recall of a copied sentence). Scores for these tests, along with a score reflecting the patient's ability to complete the test, are combined for a maximum possible score of 50.

To evaluate the TYM's usefulness in the detection of AD, the investigators recruited 679 subjects to complete the TYM as well as Addenbrooke's cognitive examination, which contains the MMSE. The participants included 94 subjects with AD, 14 subjects with amnestic mild cognitive impairment (MCI), 31 subjects with non-AD degenerative dementia, and 540 healthy controls.

In an analysis of the 94 patients with AD and 282 age-matched controls, the mean TYM score was 33.2 among the patients with AD, compared with a score of 46.6 among the controls (P<.001). After correction for multiple comparisons, differences between the cases and controls were significant for all subtests, except for sentence copying. The authors added that TYM scores, Addenbrooke scores, and MMSE scores were strongly and significantly correlated in the patients with AD (P<.001 for all 3 correlations).

A score of <=42 had a sensitivity of 93% for detecting AD, compared with a sensitivity of 52% with the MMSE. The specificity at this cutoff was 86%. The negative predictive value (NPV) of a TYM score of <=42 was 100% in a population with an AD prevalence of up to 5%; the positive predictive value (PPV) was 26%. The NPV and PPV of a TYM score of <=42 were 99% and 42%, respectively, in a population with an AD prevalence of 10%.

A score of >=43 correctly excludes a diagnosis of AD in 97% of cases even when the prevalence of AD within a population is as high as 30%, the researchers noted. Given the high NPV with this cutoff, they suggested that the TYM might be a valuable screening test for memory problems.

The low PPV, particularly in unselected groups, indicates that the TYM cannot be used alone to diagnose AD, although the authors noted that the PPV would be higher in select groups with a higher disease prevalence, such as older adults with memory complaints.

Among the 31 patients with non-AD dementia, the average TYM score (38.9) was below the cutoff for dementia (<=42). By contrast, the average MMSE score was 25.3, which was greater than the standard cutoff for dementia with this test (<=23), suggesting that the TYM was superior to the MMSE for the diagnosis of non-AD dementia. (Brown J, et al. BMJ 2009;338:b2030.)