Screening for dementia in general hospital inpatients: a systematic review and meta-analysis of available instruments

Objective: Dementia is common and often undiagnosed. Improving rates of diagnosis has become a key part of current dementia guidelines. Older people admitted to hospital are a potential target population for screening for dementia. The objective was to report whether instruments advocated in screening for dementia had been validated in hospital inpatients and to make recommendations on evidence-based screening for dementia in this population. Design: a systematic review was performed by an initial electronic database search using three key search criteria. Studies were then selected in a systematic fashion using specific predetermined criteria. Pooled meta-analysis was performed. Inclusion criteria were studies where the study group were inpatients in general hospitals, including a clearly defined group of older people (60 or older), they used a recognised screening instrument compared with a reference standard, and included at least 10 cases of dementia. Demographic data as well as sensitivity and specificity were recorded from the selected studies. Results: in total nine studies describing validation of six discreet instruments satisfied all our criteria and we were able to perform meta-analysis with one instrument, the Abbreviated Mental Test Score (AMTS). With a cut-off of < 7, pooled analysis of the AMTS showed a sensitivity of 81%, a specificity of 84% and an area under the curve (AUC) of 0.88. Conclusion: a small number of instruments have been validated for screening for dementia in general hospital. Understanding strengths and weaknesses of currently available instruments allows informed decisions about screening in this setting.