Do Higher Neighborhood Incarceration Rates Lower Life Expectancy?

The study investigates whether neighborhood incarceration rates predict life expectancy independent of broader social disadvantage. Using census tract–level data from New York State (2010 incarceration rates, 2010–2015 life expectancy, and the Social Vulnerability Index), the authors found strong negative associations. Residents in the most incarcerated neighborhoods had 5.79 fewer years of life expectancy than those in the least incarcerated, and the most disadvantaged quintile faced 4.70 fewer years. In multivariable models, incarceration rates remained a stronger predictor than social vulnerability, and their interaction showed a synergistic effect, with life expectancy gaps widening in highly incarcerated neighborhoods.

Full Citation and Link to Article

Holaday LW, Simes JT, Wang EA. Neighborhood incarceration rates, social vulnerability, and life expectancy. JAMA Intern Med. Published online July 21, 2025. doi:10.1001/jamainternmed.2025.3174

Extended Summary

Central Research Question

This study asks whether neighborhood incarceration rates are independently associated with life expectancy, even after accounting for neighborhood disadvantage as measured by the Social Vulnerability Index (SVI). It also examines whether incarceration rates amplify the health harms of disadvantage, producing a synergistic effect. The research addresses a gap in public health literature, where incarceration has been studied primarily as an individual or familial exposure, but less often as a neighborhood-level determinant of health outcomes. By embedding incarceration into broader analyses of neighborhood disadvantage, the study seeks to determine if incarceration adds predictive value in explaining health disparities across communities.

Previous Literature

The link between mass incarceration and health has been widely documented. More than 10 million people in the United States are incarcerated annually, with ripple effects touching 113 million family members and 80 million individuals with criminal records. Prior sociological and epidemiological work suggests that mass incarceration exacerbates poverty, disrupts families, reduces labor force participation, and undermines neighborhood cohesion.

Research has shown that incarcerated people disproportionately come from racially segregated and impoverished communities. Incarceration has also been linked with adverse health outcomes such as higher mortality, chronic illness, and psychological stress for both incarcerated individuals and their families. At the neighborhood level, earlier studies found a negative association between incarceration rates and life expectancy, but without controlling for multidimensional indices of disadvantage.

The Social Vulnerability Index, a validated composite measure, incorporates socioeconomic status, household composition, disability, minority status, language barriers, housing, and transportation access. While widely used in disaster response and health disparities research, the SVI does not typically account for incarceration. The study by Holaday, Simes, and Wang builds on both sociological and epidemiological literatures by asking whether incarceration rates add explanatory power beyond standard composite indices of disadvantage, and whether high incarceration intensifies the health costs of neighborhood vulnerability.

Data

The researchers merged three sources of tract-level data for New York State:

1. Life expectancy: U.S. Small-Area Life Expectancy Estimates (2010–2015).

2. Incarceration rates: 2010 census tract–level data, available only for New York State.

3. Neighborhood disadvantage: 2010 Social Vulnerability Index (SVI), developed by the CDC.

The analytic sample included 4,547 census tracts (92.5% of tracts in the state; those without life expectancy data were excluded). Both incarceration rates and SVI were operationalized into quintiles, with quintile 1 representing the lowest exposure (least incarcerated or least disadvantaged) and quintile 5 representing the highest.

Methods

The study estimated three sets of regression models with life expectancy as the outcome:

1. Bivariate models: Tested the association between incarceration quintile and life expectancy, and between SVI quintile and life expectancy, separately.

2. Multivariable model: Included both incarceration rate and SVI quintiles to assess their independent effects.

3. Interaction model: Added an incarceration × SVI interaction term to test for synergy.

All models adjusted for quintiles as categorical predictors, with quintile 1 serving as the reference category. The authors also ran a sensitivity analysis including urbanicity to assess whether urban/rural differences explained the findings. Analyses followed STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.

Findings/Size Effects

The results showed strong, consistent, and statistically significant associations between incarceration, disadvantage, and life expectancy.

• Bivariate models:

  
  

  • Incarceration rates: Residents in the most incarcerated tracts (quintile 5) had 5.79 fewer years of life expectancy compared with those in the least incarcerated tracts (quintile 1) (95% CI −6.07 to −5.51; P < .001).

  • SVI: Residents in the most disadvantaged tracts (quintile 5) had 4.70 fewer years of life expectancy than those in the least disadvantaged tracts (95% CI −5.00 to −4.40; P < .001).

  
  

• Multivariable model: Both incarceration and SVI remained negatively associated with life expectancy when included together, though the effect sizes diminished slightly, especially for SVI.

  
  

  • Incarceration quintile 5 vs. quintile 1: −4.79 years (95% CI −5.13 to −4.45).

  • SVI quintile 5 vs. quintile 1: −1.67 years (95% CI −2.01 to −1.33).

  
  

• Interaction model: The negative impact of disadvantage on life expectancy was significantly stronger in highly incarcerated neighborhoods.

  
  

  • In the least incarcerated tracts, the gap in life expectancy between the most and least disadvantaged tracts was 1.85 years (95% CI −3.01 to −0.69).

  • In the most incarcerated tracts, the same gap widened to 3.72 years (95% CI −5.99 to −1.45).

  • The interaction was statistically significant (P = .004).

  
  

• Sensitivity analysis: Adding urbanicity did not materially change the results.

Overall, incarceration rates had a stronger independent association with life expectancy than the SVI. Moreover, incarceration intensified the adverse health effects of neighborhood disadvantage, suggesting a compounding or synergistic dynamic.

Conclusion

This study demonstrates that neighborhood incarceration rates are a robust and independent predictor of life expectancy, stronger than the widely used Social Vulnerability Index. Incarceration not only reduces life expectancy directly but also magnifies the health harms of disadvantage in already vulnerable neighborhoods.

These findings have several implications:

1. Public health measurement: Incorporating incarceration into composite neighborhood indices could improve the accuracy of health outcome predictions. Indeed, a recently developed structural racism index that included jail incarceration rates outperformed other measures in predicting health outcomes.

2. Mechanisms: Incarceration can affect health through multiple pathways—removing individuals from the labor force, reducing household income, creating legal and social barriers to employment, and eroding community cohesion. Families and neighbors experience stress, stigma, and material hardship, amplifying the effects across entire communities.

3. Policy relevance: Reforms that reduce incarceration—such as alternatives to imprisonment, record expungement, and the removal of barriers to services and employment for people with records—may have population-level health benefits. Targeted efforts in high-incarceration, high-vulnerability neighborhoods could reduce health disparities and extend life expectancy.

4. Limitations: Findings are limited to New York State, the only state with tract-level incarceration data. Cross-sectional design precludes causal inference. Future research should replicate analyses in other states and test whether adding incarceration improves predictions of outcomes beyond life expectancy, such as morbidity and mental health.

In sum, the study reveals that incarceration is not merely a background factor but a central determinant of neighborhood health outcomes. By exacerbating the harms of disadvantage, high incarceration rates emerge as a powerful structural driver of reduced life expectancy. Greater access to neighborhood-level incarceration data will be essential for advancing research and designing interventions that address the intertwined legacies of mass incarceration and health inequity.