Assessment of Cognition Using Surveys and Neuropsychological Assessment: The Health and Retirement Study and the Aging, Demographics, and Memory Study
Eileen M. Crimmins, Jung Ki Kim, Kenneth M. Langa, and David R. Weir
J Gerontol B Psychol Sci Soc Sci (2011) 66B (suppl 1): i162-i171. doi: 10.1093/geronb/gbr048

One of the best resources for HRS users, beyond the extensive documentation provided by HRS, are the published articles evaluating the HRS sample and data. I rely heavily on these articles for gaining insights into issues surrounding data quality and data use.

Right around the time I started using the cognition data in HRS a paper was published that evaluated the cognitive assessments: Assessment of Cognition Using Surveys and Neuropsychological Assessment: The Health and Retirement Study and the Aging, Demographics, and Memory Study. Eileen M. Crimmins, Jung Ki Kim, Kenneth M. Langa, and David R. Weir. J Gerontol B Psychol Sci Soc Sci (2011) 66B (suppl 1).

If you are using the cognition data in HRS this paper is absolutely a must read, but it’s also useful more generally for understanding the assessment of cognition in large population surveys. Some highlights of the useful information in the paper:

• Describes two approaches for using scores on the cognitive assessment to categorize respondents has having dementia/impairment.
• Compares the categorization from scores to a consensus panel diagnosis of dementia in a subsample of HRS respondents who underwent neuropsychological testing (ADAMS).
• Provides details on how to use other survey information in HRS to assign level of impairment (normal/CIND/dementia) to respondents whose interview was conducted by proxy and who therefore did not take the assessment.

In addition to the useful information about how to use the cognition data in HRS, the main contribution of the paper is the conclusion that level of cognitive impairment can be validly assessed from a combination of performance on the cognitive assessments and, for those who don’t take the assessment, information from interviewer and proxy assessments of respondent functioning.