Attentional control typically declines with normal aging, but how modifiable social and behavioral factors shape that decline in large, diverse samples is not fully described. The authors used data from the All of Us Research Program to examine whether four potentially modifiable risk factors — smoking intensity, alcohol binge drinking, social satisfaction, and blood pressure — are associated with differences in sustained attention across adult ages.
The primary objective was to quantify associations between each risk factor and sustained attention as measured by the GradCPT and to express effect sizes in clinically interpretable, age-equivalent years of cognitive aging.
The study sample comprised 13,866 adults drawn from the All of Us Registered Tier Curated Data Repository. Cognitive performance was assessed using the Gradual-Onset Continuous Performance Task (GradCPT), a continuous-performance paradigm that yields multiple metrics related to sustained attention and response control.
Four modifiable exposures were evaluated:
Analyses adjusted for age, sex and race/ethnicity. The manuscript reports that the All of Us Research Program provides de-identified data accessed under the Registered Tier; participants provided electronic informed consent for data use.
To summarize GradCPT performance into a single outcome, the authors derived a composite cognitive efficiency score using principal component analysis on four GradCPT metrics: d′ (perceptual sensitivity), response criterion, median reaction time and reaction time variability. This composite served as the dependent variable in linear regression models.
Multiple linear regression models examined associations between cognitive efficiency and each risk factor, adjusted for age, sex and race/ethnicity. Effect sizes were converted to age-equivalent years of cognitive aging by dividing the regression coefficient for a given risk factor by the regression coefficient for age, providing an interpretable metric (age-equivalent years = coefficient_risk / coefficient_age).
Bootstrapped analyses assessed combined exposures to test whether joint effects followed an additive model or suggested interaction (greater-than-additive) effects.
GradCPT performance declined progressively with chronological age (p < 0.001), confirming expected age-related reductions in sustained attention within this cohort.
Among the four examined risk factors, self-reported social satisfaction had the largest association with cognitive efficiency. Participants reporting low social satisfaction exhibited an estimated additional 5.8 age-equivalent years of cognitive aging relative to those reporting high social satisfaction.
Smoking showed a dose–response relationship with the composite measure. Heavy smokers (> 20 cigarettes/day) demonstrated an approximate 5-year age-equivalent deficit in sustained attention compared with never-smokers.
Stage 2 hypertension showed a modest, negative association with cognitive efficiency, but this association did not reach statistical significance in the reported analyses.
Binge drinking exhibited limited and inconsistent significant associations with the composite cognitive measure. The authors note that observed patterns for alcohol may be confounded by survivor bias or abstainer bias present in cross-sectional observational samples.
Bootstrapped analyses examining combinations of exposures indicated that joint risk-factor profiles, particularly those that included low social satisfaction, produced larger deficits in cognitive efficiency than would be expected under a strictly additive model. This suggests possible interaction effects among social and behavioral risk factors on attention-related cognitive outcomes.
The authors interpret these findings as evidence supporting an interactive model of lifestyle-related cognitive vulnerability rather than a purely additive model.
In this large, diverse cross-sectional sample from the All of Us Research Program, sustained attention measured by the GradCPT declined with age and varied systematically with certain modifiable social and behavioral factors. Low social satisfaction and heavy smoking were associated with age-equivalent decrements in sustained attention of approximately 5–6 years. Stage 2 hypertension and binge drinking showed smaller or inconsistent associations.
The authors propose that the GradCPT can serve as a subclinical measure of attention-based cognitive risk in non-clinical populations and that interventions targeting social satisfaction and smoking cessation might be relevant to cognitive aging, although causality cannot be inferred from cross-sectional data.
The manuscript does not present detailed limitations beyond the points noted about potential survivor and abstainer biases affecting alcohol associations; other specific limitations and sensitivity analyses are reported in the full text and supplementary files.
The analyses used de-identified data from the All of Us Registered Tier Curated Data Repository (specified versions in the manuscript). Access to the data requires institutional registration and training. Summary statistics are reported in the manuscript and Additional file 1; analysis code is available from the corresponding author on reasonable request. The study was reviewed under existing All of Us IRB oversight and the authors' institutional review determined the secondary analysis exempt from full review. Funding sources listed include RTI International, the Foundation for the National Institutes of Health, and RCMI Morgan State Center for Urban Health Disparities Research and Innovation.
Note: this manuscript is a preprint and has not been peer reviewed; the authors explicitly state it should not be used to guide clinical practice.