Back

Diverging Patterns of Cognitive Decline by Sex and Race-Ethnicity in Seriously Ill Older Americans

Mejia-Guevara, I.; Periyakoil, V.

2024-06-28 epidemiology
10.1101/2024.06.27.24309609 medRxiv
Show abstract

ObjectivesDifferences in Cognitive decline are common in older adults in the last years of life, but differences across sex and race-ethnicity are poorly understood. This study investigated if sex and/or race-ethnicity moderated changes in cognitive function in older adults in the last years of life. MethodsData from the Health and Retirement Study (HRS) from 1993 to 2016 were used to analyze imputed cognition summary scores for total word recall and mental status of older adults aged 60-99. Loss of cognitive function was estimated using a multilevel mixed-effects model and accelerated cognitive decline was approximated by incorporating a change-point model using a restricted sample of decedent respondents who died aged 65-99. ResultsNotable disparities were seen in the rates of cognitive decline across sex and race-ethnic groups in the last years of life. Women consistently scored lower than men in word recall but higher in mental status, regardless of race-ethnicity. Non-Hispanic White respondents, men and women, consistently outperformed Hispanic and Black respondents in word recall tasks and mental status. ConclusionsOur study shows that sex and race-ethnicity moderate cognitive decline in older adults during the last years of life. Older adults from underserved communities are at higher risk of cognitive decline. Our study could inform clinical practice and policy focused on mitigating the adverse impact of cognitive decline experienced by marginalized populations of older adults in the last years of life.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.