Disparities exist within the danger of dementia amongst black and white adults

The ratio of dementia risk among black-and-white non-Hispanic people in the US doesn’t seem to have changed between 2000 and 2016, researchers found in a study published in JAMA Neurology.

Several studies have reported that non-Hispanic black people are at higher risk of dementia than non-Hispanic white people. Using data from the US Health and Retirement Study (HRS), a nationally representative study of adults aged 50 and over, the researchers wanted to investigate whether there are relative racial differences in dementia in the US, both in terms of relative prevalence and relative Incidence of dementia changed from 2000 to 2016.

The researchers analyzed data from 9 HRS waves from 2000 to 2016. Each wave (ranging from 6322 to 7579 eligible participants per wave) became a cross-sectional study to quantify racial differences in the prevalence of dementia at intervals of 2 years from the researchers estimated the development of racial differences during this time using data from all 9 waves.

Subcohorts (ranging in size from 5322 to 5961 participants) with a follow-up period of 4 years were nested within the larger longitudinal HRS in order to quantify the racial differences over calendar periods with the base years 2000 to 2012. All data from the subcohorts were then combined.

Participants who had an algorithmic dementia diagnosis in the base year were excluded from subcohorts, and subcohort participants who had no diagnosis at the end of 2 follow-up waves were censored.

The researchers performed all analyzes using the LASSO (Least Absolute Shrinkage and Selection Operator) expert model and modified Hurd algorithms that predict dementia status based on socio-demographic characteristics, physical health, and cognitive health collected during the HRS interview. All three algorithms were designed for use in HRS-based studies of racial and ethnic differences in dementia because they have similar out-of-sample sensitivity and specificity for non-Hispanic white and non-Hispanic black subgroups.

The study’s authors used HRS sample weights to adjust the different inclusion probability and considered geographic stratification and clustering when estimating standard errors. Data from non-Hispanic black and non-Hispanic white HRS participants who were 70 years of age and older in each wave and who could be attributed an algorithmic diagnosis of dementia were analyzed. Weighted regression models were used to estimate crude dementia prevalence rates and to estimate and quantify time trends in the incidence of dementia. The race-specific distribution of age and gender in 2000 was standardized.

The expert model and the LASSO algorithms classified an average of 20% of the participants as having dementia, while the modified Hurd algorithm classified 18% of the participants as having dementia. All three algorithms found that non-Hispanic black participants had an approximately 1.5 to 1.9 times higher prevalence of dementia than non-Hispanic white participants in all waves, in both rough and standardized estimates.

Overall, the prevalence of dementia decreased over time for both raw and standardized estimates, and point estimates indicated a slight decrease in the raw prevalence rate compared to groups in later years, but the relative prevalence of dementia in both groups did not change over time essential. After standardizing age and gender, there was no evidence of a decrease in racial differences over time.

In all subcohorts, non-Hispanic black participants had about 1.4 to 1.8 times the incidence of dementia compared to non-Hispanic white participants. Analyzes using the expert model or modified Hurd algorithms to determine the dementia status indicated no change over time, while analyzes using the LASSO model indicated that the incidence of dementia only decreases over time in age- and gender-standardized analyzes.

The limitations of the study included the use of an algorithm to detect dementia and an emphasis on the prevalence and incidence of clinical dementia syndrome rather than biomarker-based diagnoses of Alzheimer’s disease.

“Although our results indicate a stable or declining risk of dementia overall, we did not find any evidence that the relative racial differences in the risk of dementia decreased between 2000 and 2016,” the authors say.


Power MC, Bennett EE, Turner RW, et al. Trends in the relative incidence and prevalence of dementia among black and white non-Hispanic individuals in the US, 2000-2016. JAMA Neurol. Published online November 30, 2020. doi: 10.1001 / jamaneurol.2020.4471

This article originally appeared on Psychiatry Advisor

Related Articles