Back

Prevalence of Risk Factors and Established Cardiovascular Disease Among All of Us Participants: Benchmarking Against National Estimates

Yang, M.; Admassu, T.; Krumholz, H.; Huang, C.; Murugiah, K.

2025-07-23 cardiovascular medicine
10.1101/2025.07.22.25331985 medRxiv
Show abstract

IntroductionKnowing how All of Us (AoU) Research Program participants differ in demographics and prevalence of cardiovascular risk factors and established cardiovascular disease (CVD) from the U.S population is important for cardiovascular researchers using these data, and to provide context to their findings. MethodsWe included AoU participants enrolled between May 2017-June 2022. AoU Survey cohort included participants who completed the Personal and Family Health History survey (N=185,232). AoU EHR included participants with linked EHR data (N = 287,012). We identified cardiovascular risk factors and established CVD through survey questions in AoU Survey and National Health and Nutrition Examination Survey (NHANES), and SNOMED codes for AoU EHR. Prevalence was compared to the weighted NHANES 2017-March 2020 cycle (N=9,683 representing 248 million), overall and by demographic sub-groups. Results AoU Survey and AoU EHR were both older than NHANES (56.1{+/-}17 years and 57.6{+/-}17 years versus 47.5{+/-}18 years) with more female participants. Black participants were underrepresented in AoU Survey. Prevalence of cardiovascular comorbidities in AoU Survey and NHANES were similar overall, but lower when stratified by age and among female participants. Black participants had higher rates of many comorbidities like hypertension, diabetes, coronary artery disease and congestive heart failure than the U.S Black population, but lower chronic kidney disease. AoU EHR with its code-based identification had distinctly higher prevalence of most cardiovascular comorbidities. ConclusionAoU includes more historically underrepresented groups, as intended. However, researchers should understand that demographic composition and disease prevalence change based on data availability and how conditions are ascertained.

Matching journals

The top 9 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.