Diagnostic test accuracy of cognitive screeners in older people

Age is the most important risk factor for the development of cognitive decline and dementia. Nevertheless, memory concerns are also prevalent in younger age groups, prompting referral of such patients from primary care to dedicated memory clinics for assessment. Because of the operation of age criteria in some memory clinics, for example those conducted under the auspices of geriatricians and old age psychiatrists (although following a national directive some old age psychiatry clinics now eschew an age cut-off), many of these younger patients are seen in neurologyled memory clinics to be investigated for possible early-onset dementias, the differential diagnosis of which differs somewhat from that in older patients.1 A consequence of this case mix may be to influence the results of pragmatic diagnostic test accuracy studies of CSIs performed in this setting,2 as young patients have a lower prior probability of dementia and mild cognitive impairment (MCI) than older patients. The diagnostic utility of CSIs in older people might therefore be obscured where the case mix is heterogeneous. The aim of the study presented here is to examine the screening utility of five short CSIs for the diagnosis of dementia and MCI versus subjective memory complaint (SMC) in older people (≥65 years), ie a sample enriched for those at greatest risk of cognitive impairment and dementia. Information was collated from several pragmatic diagnostic test accuracy studies undertaken in a secondary care setting (a dedicated neurology-led cognitive disorders clinic based in a regional neuroscience centre) of CSIs taking approximately 5–10 minutes to administer. This type of analysis has previously been reported for the short Montreal Cognitive Assessment (s-MoCA).3 Materials and methods The datasets from several pragmatic prospective diagnostic test accuracy studies undertaken in the authors’ clinic were used,4–9 which examined the following short CSIs: MiniMental State Examination (MMSE);10 Mini-Addenbrooke’s Cognitive Examination (MACE);11 Montreal Cognitive Assessment (MoCA);12 Sixitem Cognitive Impairment Test (6CIT),13 and AD8.14 Study duration was at least one year, for MACE it was two years. For most of these CSIs, lower scores indicate worse cognitive function (MMSE, MACE, MoCA), but others are negatively scored such that lower values are better and higher are worse (6CIT, AD8). Study details (sample size, age range, patient numbers ≥65 years, and prevalence of dementia and MCI in the full cohort and older subgroup) are shown in Table 1. Diagnostic test accuracy of cognitive screeners in older people