Learning from Two Decades of Blood Pressure Data: Demography-Specific Patterns Across 75 Million Patient Encounters

TL;DR

The paper addresses the problem that BP distributions and their clinical interpretation vary across demographic groups, challenging one-size-fits-all thresholds. It analyzes a massive cuff-based BP dataset from Emory Healthcare (2000–2022), cleans the data, and aggregates medians by age, sex, and race to quantify population-level statistics and the interdependence between $SBP$ and $DBP$. Key findings include higher BP in males, age-related increases in $SBP$, a peak in $DBP$ during the 40s, and higher BP and dispersion in African American/Black individuals, along with a significant population-level correlation between $SBP$ and $DBP$ across groups. The work underscores the importance of demography-aware BP analysis to enable personalized hypertension risk assessment and to inform future precision medicine approaches for BP monitoring and management.

Abstract

Hypertension is a global health concern with an increasing prevalence, underscoring the need for effective monitoring and analysis of blood pressure (BP) dynamics. We analyzed a substantial BP dataset comprising 75,636,128 records from 2,054,462 unique patients collected between 2000 and 2022 at Emory Healthcare in Georgia, USA, representing a demographically diverse population. We examined and compared population-wide statistics of bivariate changes in systolic BP (SBP) and diastolic BP (DBP) across sex, age, and race/ethnicity. The analysis revealed that males have higher BP levels than females and exhibit a distinct BP profile with age. Notably, average SBP consistently rises with age, whereas average DBP peaks in the forties age group. Among the ethnic groups studied, Blacks have marginally higher BPs and a greater standard deviation. We also discovered a significant correlation between SBP and DBP at the population level, a phenomenon not previously researched. These results emphasize the importance of demography-specific BP analysis for clinical diagnosis and provide valuable insights for developing personalized, demography-specific healthcare interventions.

Figures (4)

• Figure 1: Male vs female blood pressure distributions. The contours correspond to 95% percentile range of the data. Red/blue dots show the mean SBP and DBP of males/females.
• Figure 2: Blood pressure distribution across race/ethnicity. The contours correspond to 95% percentile ranges. Dots show the mean SBP and DBP in each racial/ethnic group.
• Figure 3: Mean$\pm$STD pulse pressure (top) and blood pressure (bottom) by age, for males and females.
• Figure 4: Blood pressure distribution of different age groups. The contours correspond to 95% percentile ranges. Dots show the mean SBP and DBP in each age group.