Objective To test diagnostic accuracy of the mini-Addenbrooke's Cognitive Examination (m-ACE) compared to the MMSE for the diagnosis of dementia and MCI in consecutive referrals to a dedicated cognitive disorders clinic.

Results: Of 135 consecutive new outpatients seen over 6 months (June–November 2014) administered the mini-ACE (F:M=64:71, 47% female; age range 18–88 years, median 60), 24 were diagnosed with dementia (DSM–IV–TR criteria) and 39 had MCI (Petersen criteria). Using the cutoffs defined in the index paper (≤25/30 and ≤21/30), m-ACE was sensitive (1.00, 0.92) but not specific (0.28, 0.61) for dementia diagnosis; it also proved useful for MCI diagnosis (sensitivities 1.00, 0.77; specificities 0.43, 0.82). Area under the ROC curve was 0.86. Effect size (Cohen's d) for m-ACE for dementia vs. no dementia was 1.53 (large) and for MCI vs no cognitive impairment was 1.59 (large); for MMSE the corresponding figures were 1.56 and 1.26. Weighted comparison suggested a small net loss for m-ACE vs MMSE for dementia diagnosis (–0.13) but a large net benefit for MCI diagnosis (0.38).

Conclusions: In this pragmatic study, m-ACE proved quick, easy to use, and acceptable to patients, with metrics comparable to MMSE for dementia diagnosis and better for MCI diagnosis.