Inequalities in accelerated cognitive decline: Resolving observational window bias using nested non-linear regression

Area of Science:

• Neuroscience
• Public Health
• Sociology

Background:

• Observational window bias complicates studies on educational differences in Alzheimer's disease and related dementias (ADRD) risk.
• Previous research has yielded conflicting results due to limitations in observing the period preceding accelerated cognitive decline.

Purpose of the Study:

• To introduce a novel model addressing observational window bias in ADRD research.
• To investigate the presence and origins of disparities in accelerated cognitive decline indicative of ADRD.

Main Methods:

• Utilized a large dataset of 167,314 cognitive assessments from 32,441 Health and Retirement Study participants.
• Implemented a parametric non-linear nested longitudinal regression model.
• Reported multivariable-adjusted nodal incidence ratios (aNIR) to quantify risk.

Main Results:

• University degrees were associated with a 74.7% lower incidence of accelerated cognitive decline (aNIR = 0.253, p < 0.001).
• Black participants exhibited a 99.5% higher incidence of accelerated cognitive decline (aNIR = 1.995, p < 0.001), even after adjusting for covariates.
• Sex-stratified analyses revealed reduced educational benefits for women and increased incidence among minoritized women.

Conclusions:

• The novel approach effectively mitigates observational window bias, uncovering substantial social inequalities in ADRD onset.
• Findings demonstrate significant educational and racial/ethnic disparities in the preclinical phase of ADRD.
• The study highlights the intersectionality of sex/gender with race and education in ADRD disparities.