| Abstract | Within the United States, many large-scale, nationally representative studies exist with
the goal of tracking and monitoring aspects of health. These studies are often used to establish
the prevalence of dementia and subjective cognitive decline (SCD) in the population. The goal of
the current study is to examine how different population-based studies probe respondents about
conditions related to cognitive impairment, and to assess similarities and differences in point
estimates. We reviewed eight studies and identified comparable items related to dementia and
SCD. We calculated design-appropriate point prevalence estimates and compared weighted
estimates across studies, finding a wide range and statistically significantly different estimates
for dementia (estimates ranging from 2.7% - 9.9%) and for SCD (5.6% - 46.6%). Close analysis
of item construction revealed meaningful differences in the use of terminologies and timeframes
that could account for these differences. Moreover, subtle but consequential sampling differences
were also discovered within study documentation that also could be responsible. Given the
importance of prevalence estimates for research, practice, and policy, our findings highlight the
need for harmonization across methodology in these large studies, even at their most basic level,
to establish the true burden of these conditions.
|
