A study of patients living in institutes may bias the findingsBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.6997.126 (Published 08 July 1995) Cite this as: BMJ 1995;311:126
EDITOR,--Alewijn Ott and colleagues report their estimates of the prevalence of dementia in a suburb of Rotterdam and state that the fact that 27% of the potential population was unscreened was likely to lead to an underestimate of the rates.1 This fails, however, to account for the bias of living in an institution.
Altogether 8% of the total population, a high rate for the cut off age of 55, was living in an institution, but all those in institutions were included in the authors' screened sample--a rate of 12%. Since the authors state that many of the subjects with dementia were living in an institution the true prevalence is likely to be lower than their estimate. Unfortunately, they detail neither the number in institutions considered to be demented nor the quality of the diagnostic work up that this sizeable subgroup had. It is also unclear whether this part of Rotterdam is typical in terms of the number of residential homes per capita. Perhaps better educated people with dementia are more able to avoid admission to an institution in the Netherlands and fall into the unevaluated 23%, while those with lower educational attainment end up in institutions.