Study design greatly influences reported rates of cognitive impairment and dementia. Important factors that strongly affect results include choice of diagnostic tool, timing of physician input, selection of patients, and selection of outcome variables (physician awareness vs physician documentation; cognitive impairment vs dementia). Most well-designed studies have reported high rates of undetected dementia.
The articles cited by Dr van Hout have considerable design concerns. The study design reported in the articles by Bowers et al1 and Mant et al2 asked physicians to self-select patients from their practice and then comment on the presence of dementia in those individuals. It is likely, in this scenario, that physicians would choose to report on individuals in whom they were confident of their diagnosis. In the latter article, physicians were actually encouraged to include high numbers of individuals thought to be demented. In both studies the Mini-Mental State Examination score was the main criterion for dementia, not the findings from physician examination.