METHODS

Self-reported memory complaints are common in elderly individuals. Depending on the population studied, and the manner in which the question was asked, between 20% and 56 % of elderly persons report problems with their memory. Recent population-based studies have shown that subjective memory loss (SML) predicts dementia or cognitive decline in seniors with normal cognition.1-5 How-ever, the usefulness of subjective memory complaints in identifying persons with cognitive impairment is not clear: some studies report a weak or no association between SML and cognition,6-9 while others report an association between SML and cognitive status.10,11 It is also unclear if the association between SML and cognition is due to potential confounding by depression12: those with depres-sive symptoms may be more likely to report SML and may have more impairment of cognition. SML is important for clinicians for 2 reasons: First, patients may present themselves with memory complaints. Clinicians then need to know if they should proceed with further cognitive assessment. Second, SML could be use-ful as a very simple screening test for cognitive impairment, if it accurately reflects true cognition. To clarify the association between cognitive status and SML, we conducted a secondary analysis of an exist-ing data set (the Manitoba Study of Health and Aging). Specifically, our objectives were the following: 1. to determine if subjective memory complaints are associated with Mini-Mental State Exam (MMSE) scores; 2. to determine if this association is independent of other factors such as age, sex, education, and depres-sive symptoms; and 3. to determine the sensitivity and specificity of subjec-tive memory complaints for the presence of cognitive impairment (dementia or cognitive impairment, no dementia)