Comparing retrospective informant assessments to prospectively collected cognitive measures in the Health and Retirement Study.

INTRODUCTION: Clinicians and researchers frequently ask informants about changes in a person's cognition, but whether informant assessments correspond to objectively measured change is unclear.

METHODS: A subset ( = 2710) of US Health and Retirement Study participants and their informants completed the Harmonized Cognitive Assessment Protocol (HCAP). Using generalized estimating equations, we compared informant-reported change in memory and daily functioning to prospectively collected delayed word recall and instrumental activities of daily living (iADL) in the 10 years preceding HCAP.

RESULTS: Informant reports of worsened memory were associated with declining word recall, and informant-reported iADL loss was associated with declining iADLs. Informant-reported memory impairment was more strongly associated with declining word recall when informants saw the respondent weekly or more compared to one to three times monthly or less ( < 0.0001 for interaction).

DISCUSSION: Informant assessments of memory and iADLs are generally consistent with prospective measurements, but this relationship depends significantly on frequency of informant contact.

HIGHLIGHTS: Informant ratings of a person's memory and daily functioning are generally reliable.Quality of reporting depends on frequency of informant contact.Knowing an informant's characteristics is important for interpreting responses.