Family of Origin Bilingualism and ADRD: An Epidemiologic Study of 377,000 Older Adults in the US - Project Summary/Abstract: Evidence for links between bilingualism and Alzheimer’s Disease and Related
Disorders (ADRD) is mixed and arises from a range of study designs. Few prospective cohort studies exist in
the US, where associations between bilingualism and ADRD may be confounded by socioeconomic status
(SES) and immigration status. We propose to fill this gap by examining the association between bilingual family
of origin and ADRD over 50 years later in Project Talent, a sample of 377,000 individuals representative of the
current cohort of older Americans. Project Talent participants completed a comprehensive battery of measures
in 1960 when they were in high school (age 13-17), including data on languages spoken in the home, cognitive
tests, socioeconomic factors, and parental immigration status. We will extend assessment of family
immigration status back to 1950 by linking participant data to the newly released 1950 individual census
records. The cohort already has established linkages with the Center for Medicaid and Medicare Services
(CMS) and National Death Index (NDI) which we propose to update. This will ascertain ADRD outcomes when
the cohort is 76-80, a period of rapid increasingly incidence. We will address three major questions through a
life course epidemiology lens. First, has been unclear if bilingualism is associated with ADRD incidence in
general, or with age of diagnosis. Our follow-up period allows us to address both questions, which reduce to an
issue of shifting relative risks with age. The design also permits us to address another issue that has been
largely ignored, competing risk censoring. Second, SES and immigration status may be important confounders
of any links. Detailed measurements of both family SES and immigration status will permit us to determine
whether one or both of these factors confound any observed associations between family of origin bilingualism
and ADRD. As well, adolescent cognitive ability may function as a mediator, explaining a portion of the
association. Third, our sample is large enough to explore questions of moderation by language type, and
speculations about mediation or moderation by attained education (the latter within a subsample followed up at
age 29). In summary, the proposed study would be the largest to date on these questions of which we are
aware and features a sample directly relevant to the current generation of US older adults.
