Objectives: More and more referrals to memory clinics are expected, but diagnostic routes are already challenged. In order to be able to follow the advice to diagnose dementia more often and earlier in the process, but also to be able to handle the increasing numbers of referrals, a fast but reliable triage test is needed.

According to the Cochrane review, “the MoCA can help identify people who need specialist assessment and treatment for dementia”. It has been validated in multiple institutions and languages. However, many of these studies are designed with a case-control design using healthy, community-based individuals as controls, which can lead to spectrum bias.

Our cohort of referrals to a memory clinic with patients suspected of having cognitive disorders (mild dementia and MCI) after initial assessment in an old age psychiatric clinic, needs to be validated because different settings can give different results.

Design: our reference standard consisted of a consensus-based diagnosis according to international criteria for detecting MCI and MD, and this was compared with patients suspected of MCI/MD - but excluded from cognitive disorders (NoCI)- from the same cohort.

Results: The mean MoCA scores differ significantly between the groups: 24 in NoCI, 21 in MCI and 16.5 in MD. The AUC of MD against non-demented (MCI+NoCI) was 0.83 resulting in 90% sensitivity, 65% specificity, 50% PPV and 94% NPV at a best cut-off of <21 according the Younden index. For CI (MD+MCI) against NoCI the results were respectively 0.77AUC, 95%sens, 47%spec, 88%PPV, 68%NPV at a cut-off <26.

On an individual basis, as a box plot of DSM IV diagnoses showed, the MoCA score has limitations and clinical aspects need to be taken into account: FTD, high education to the upside; MCI including psychiatric etiology to the downside.

Conclusions: by using a cut-off score of <21, 90% of people with positive MoCA have CI, while 94% of people with negative MoCA (≥21) will not have dementia. The MoCA can significantly reduce referrals (50%) by excluding patients for further diagnostic work-up at a memory clinic, even if they are suspected of CI after initial assessment.