Within-Group Racial and Ethnic Differences in County-Level Socio-Behavioral Risk Across Cancer Mortality Tertiles in the United States

This cross-sectional study of 3,141 U.S. counties reveals that within-group socio-behavioral risk profiles vary significantly across local cancer mortality gradients, demonstrating that analyzing domain-specific determinants rather than relying on group averages is essential for developing precise, equity-oriented strategies to address breast and prostate cancer disparities.

The Gist

Imagine the United States as a giant, sprawling garden with 3,000+ different plots (counties). In this garden, two specific types of weeds—breast cancer and prostate cancer—are causing the most trouble.

For a long time, gardeners (researchers and policymakers) have looked at the garden and said, "Oh, the Black section of the garden has more weeds than the White section, and the Hispanic section has fewer." They've been watering the whole section based on that average.

But this new study says: "Wait a minute! That's like looking at a whole neighborhood and saying 'everyone here is rich' or 'everyone here is poor.' It's not true. There are rich people in poor neighborhoods and poor people in rich neighborhoods."

Here is the breakdown of what this study found, using simple analogies:

1. The "Average" Trap

Most studies look at the average risk for a whole group (like all Black women or all Hispanic men).

• The Problem: If you take a group of 100 people, and 90 are doing great but 10 are in deep trouble, the "average" looks okay. You might miss the 10 people who are drowning.
• The Study's Fix: Instead of looking at the group average, the researchers zoomed in on specific counties (plots of land) and sorted them into three buckets: Low Risk, Medium Risk, and High Risk for cancer deaths. Then, they looked inside each racial group to see what was happening in those specific buckets.

2. The "Risk Score" (The Weather Report)

The researchers created a "Socio-Behavioral Risk Score." Think of this like a weather report for a community, but instead of rain and wind, it measures things like:

• Economic Livelihoods: Do people have jobs and money?
• Touchpoints with Care: Can they actually see a doctor?
• Lifestyle Behaviors: Is there access to healthy food and exercise?
• Community Solidarity: Is the neighborhood supportive or isolated?

The Big Finding: In almost every group (White, Black, and Hispanic), the counties with the worst cancer death rates also had the worst "weather reports" (highest risk scores). The worse the social conditions, the higher the cancer deaths.

3. The "Hidden Danger Zones" (The Hispanic Paradox)

This is the most interesting part.

• The Old View: Hispanic populations generally have lower cancer death rates on average. It's like saying, "The Hispanic section of the garden is mostly healthy."
• The New View: The study found that while the average is good, there are specific Hispanic counties with very high cancer death rates.
• The Analogy: Imagine a car that usually gets great gas mileage (low average deaths). But in a few specific cities, that same car is breaking down constantly because of bad roads and no gas stations.
• The Result: In those high-death Hispanic counties, the "weather" was terrible specifically regarding jobs, money, and access to doctors. If we only looked at the "average," we would have ignored these specific towns that desperately need help.

4. The "Black Community" Reality

For Non-Hispanic Black populations, the study found that the "weather" was bad everywhere, even in the counties with lower cancer death rates.

• The Analogy: It's like living in a house where the foundation is shaky everywhere. Even in the "better" parts of the Black community, the risk factors (like lack of healthcare access or economic struggle) are still much higher than in other groups. This means you can't just focus on the "worst" areas; you need to support the whole community because the baseline risk is high everywhere.

5. The "Recipe" for Success

The study also found that the type of problem changes depending on the cancer and the group.

• Breast Cancer: The biggest problems were often about lifestyle (food, exercise) and community support.
• Prostate Cancer: The biggest problems were often about money and access to doctors.
• The Lesson: You can't use the same "cure" for every problem. If you have a town where the main issue is "no doctors," giving them a pamphlet about "eating more broccoli" won't fix the cancer deaths. You need to fix the doctor access first.

The Bottom Line (The "So What?")

This paper is telling us to stop using a one-size-fits-all map.

• Old Way: "We need to help Black women with breast cancer." (Too vague).
• New Way: "We need to help Black women in County X because their main problem is no doctors, but we need to help Hispanic women in County Y because their main problem is no money for healthy food."

By looking at the specific "ingredients" of the problem in each specific town, we can stop wasting resources and start fixing the exact things that are causing people to die. It's about moving from a blurry group photo to a sharp, high-definition picture of exactly who needs help and what kind of help they need.

Technical Summary

1. Problem Statement

Persistent racial and ethnic disparities in breast and prostate cancer mortality in the United States are well-documented, with Non-Hispanic Black populations experiencing significantly higher mortality rates than Non-Hispanic White and Hispanic populations. However, existing research primarily focuses on between-group differences and relies on population averages or single composite indices of social disadvantage.

This approach obscures within-group heterogeneity, failing to identify high-need communities within specific racial/ethnic groups. Furthermore, it does not characterize how specific socio-behavioral determinants of health (SBDOH) vary across local gradients of cancer mortality. The study addresses the gap in understanding how SBDOH profiles differ within racial/ethnic groups across low, moderate, and high-mortality counties to inform more precise, equity-oriented interventions.

2. Methodology

Study Design and Population

• Design: Cross-sectional ecological study of U.S. counties.
• Sample: 3,141 U.S. counties.
• Stratification: Counties were stratified by race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic) and cancer type (Breast, Prostate).
• Data Sources:
  • Mortality: Age-adjusted breast and prostate cancer mortality rates (2018–2022) from the National Cancer Institute (SEER).
  • SBDOH Indicators: Derived from the American Community Survey (ACS) and other national sources.

Exposures: Socio-Behavioral Determinants of Health (SBDOH)
The study utilized a composite index comprising 59 indicators across 7 domains:

1. Community Solidarity: (e.g., incarceration rates, childcare access, residential segregation).
2. Education, Health Literacy, and Digital Connectivity: (e.g., high school completion, internet access, health literacy).
3. Quality of Care: (e.g., hospital/pharmacy access, preventable hospital admissions).
4. Housing and Environmental Risk: (e.g., crowding, air pollution, housing cost burden).
5. Economic Livelihoods: (e.g., income, unemployment, medical debt).
6. Lifestyle Behaviors: (e.g., tobacco/alcohol use, physical activity, diet).
7. Touchpoints with Care: (e.g., lack of primary provider, missed screenings, preventive care gaps).

Analytical Approach

• Scoring: Race/ethnicity-specific census tract-level indicators were aggregated to the county level using population-weighted averaging. Relative risk scores were generated via percentile ranking (0 = lowest risk, 100 = highest risk).
• Grouping: Counties were classified into mortality tertiles (Low, Moderate, High) based on race/ethnicity-specific age-adjusted mortality rates.
• Statistical Analysis:
  • Population-weighted mean SBDOH scores were calculated for each tertile.
  • Due to non-normal distributions, Kruskal-Wallis rank-sum tests were used to assess differences across tertiles.
  • Dunn's multiple comparison tests with Bonferroni adjustment were used for pairwise post-hoc analysis.

3. Key Contributions

1. Within-Group Focus: Shifts the analytical lens from between-group disparities to within-group heterogeneity, demonstrating that high-need communities exist within every racial/ethnic group.
2. Domain-Level Profiling: Moves beyond a single deprivation score to provide cancer-specific and race-specific domain profiles. This identifies which specific barriers (e.g., economic livelihoods vs. touchpoints with care) drive mortality in specific contexts.
3. Dual Cancer Analysis: Simultaneously analyzes breast and prostate cancer, revealing that socio-behavioral gradients are cancer-specific (e.g., monotonic patterns for breast cancer vs. more complex patterns for prostate cancer).
4. Actionable Framework: Proposes a framework linking local mortality strata with domain-level SBDOH profiles to guide targeted intervention design.

4. Key Results

Overall Trends

• Higher county-level socio-behavioral risk was consistently associated with higher breast and prostate cancer mortality across all racial and ethnic groups.
• Non-Hispanic Black populations exhibited the highest mortality rates and persistently elevated socio-behavioral risk across all mortality tertiles.
• Hispanic populations had the lowest average mortality but showed the steepest within-group gradients in socio-behavioral risk.

Breast Cancer Findings

• Risk Gradient: Socio-behavioral risk increased monotonically across mortality tertiles for all groups.
• Magnitude of Increase: The largest absolute increases in risk occurred between low- and moderate-mortality counties for Hispanic and Non-Hispanic Black women.
• Domain Specificity:
  • Non-Hispanic Black: Largest gaps in touchpoints with care, lifestyle behaviors, and community solidarity.
  • Hispanic: Largest gaps in lifestyle behaviors, touchpoints with care, and economic livelihoods.
  • Non-Hispanic White: Largest gaps in touchpoints with care, economic livelihoods, and lifestyle behaviors.

Prostate Cancer Findings

• Risk Gradient: Generally increased across tertiles but was less monotonic than breast cancer, particularly for Hispanic and Non-Hispanic White men.
• Domain Specificity:
  • Non-Hispanic Black: Largest gaps in economic livelihoods, lifestyle behaviors, and community solidarity.
  • Hispanic: Largest gaps in touchpoints with care, economic livelihoods, and lifestyle behaviors.
  • Non-Hispanic White: Largest gaps in touchpoints with care, housing/environmental risks, and community solidarity.

The "Hispanic Paradox" Nuance
While Hispanic populations had lower aggregate mortality, high-mortality Hispanic counties exhibited pronounced risks in lifestyle behaviors, economic livelihoods, and touchpoints with care. This suggests that lower population averages mask severe localized vulnerabilities.

5. Significance and Implications

Clinical and Public Health Impact

• Precision Interventions: The study argues that relying on group averages leads to "one-size-fits-all" strategies that miss high-need sub-populations. By mapping specific domain barriers to mortality strata, stakeholders can design interventions that target the dominant barriers in a specific geography (e.g., improving economic livelihoods in high-mortality Hispanic counties vs. improving community solidarity in high-mortality Black counties).
• Equity-Oriented Prioritization: The framework allows funders and policymakers to move beyond broad socioeconomic indices to identify exactly where and why mortality is elevated within specific communities.

Limitations

• Ecological Fallacy: Associations are at the county level and cannot be inferred at the individual level.
• Data Suppression: Mortality data suppression for small counties reduced sample sizes for Hispanic and Non-Hispanic Black strata.
• Cross-Sectional Nature: Causality cannot be established; the study identifies associations, not causal mechanisms.
• Temporal Misalignment: SBDOH indicators and mortality data were drawn from slightly different time windows, though community conditions generally change slowly.

Conclusion
The paper concludes that within-group heterogeneity in socio-behavioral risk is substantial. Linking population-specific, domain-level SBDOH profiles to cancer mortality provides a superior basis for equity-oriented cancer control strategies compared to traditional reliance on group averages or composite indices. This approach enables the alignment of intervention strategies with the specific structural and behavioral barriers faced by high-mortality communities.