Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Aug 9.
Published in final edited form as: Am J Prev Med. 2024 Feb 27;67(1):15–23. doi: 10.1016/j.amepre.2024.02.015

Impacts of Poverty and Lifestyles on Mortality: A Cohort Study in Predominantly Low-Income Americans

Lili Liu 1, Wanqing Wen 1, Martha J Shrubsole 1, Loren E Lipworth 1, Michael T Mumma 2, Brooke A Ackerly 3, Xiao-Ou Shu 1, William J Blot 1, Wei Zheng 1
PMCID: PMC11312224  NIHMSID: NIHMS2012038  PMID: 38417593

Abstract

Introduction:

Low socioeconomic status has been linked to increased mortality. However, the impacts of poverty, alone or combined with health behaviors, on mortality and life expectancy have not been adequately investigated.

Methods:

Data from the Southern Community Cohort Study was used, including nearly 86,000 participants recruited during 2002–2009 across 12 US southeastern states. Analysis was conducted from February 2022 to January 2023.

Results:

During a median follow-up of 12.1 years, 19,749 deaths were identified. A strong dose-response relationship was found between household incomes and mortality, with a 3.3-fold (95%CI=3.1–3.6) increased all-cause mortality observed for individuals in the lowest income group (<$15,000/year) compared with those in the highest group (≥$50,000/year). Within each income group, mortality monotonically increased with declining healthy lifestyle score. Risk was significantly lower among those in the lowest income but healthiest lifestyle group, compared to those with the highest income but unhealthiest lifestyle (HR=0.82, 95%CI=0.69–0.97). Poor White participants appeared to experience higher all-cause mortality than poor Black participants. Life expectancy was more than 10.0 years shorter for those in the lowest income group compared with those in the highest income group.

Conclusions:

Poverty is strongly associated with increased risk of death, but the risks could be modestly abated by a healthier lifestyle. These findings call for a comprehensive strategy for enhancing a healthy lifestyle and improving income equality to reduce death risks, particularly among those experiencing health disparities due to poverty.

INTRODUCTION

Low socioeconomic status (SES) has been linked to increased mortality,13 among which income inequality plays a key role.47 Individuals with a low income are at an elevated risk of developing cardiovascular diseases, cancer and other noncommunicable diseases.811 They also have worse disease prognoses due in part to limited access to healthcare.12,13 Previous studies investigating associations of income with mortality outcomes were mainly conducted in middle- or upper-middle-income populations. Evidence from low-income and minority populations is limited, especially for those living in developed or high-income countries.5,6,14,15 Given the widening socioeconomic gaps in the U.S. and many other countries over recent decades, improving the understanding of the income-mortality association in low-income populations and uncovering potential impacts of modifiable risk factors on this association could better inform the development of intervention programs to reduce mortality, and ultimately reduce or eliminate the SES disparity in health.3,4

Current evidence indicates that poverty may influence health via multiple pathways, including psychosocial, biological, and behavioral routes.8,10,13,16 Individuals experiencing poverty have fewer economic resources, less access to health care, and more often live in poor and heavily polluted communities.1719 They are also more likely to have unhealthy lifestyle behaviors, such as poor diet, cigarette smoking, heavy alcohol drinking, and being physically inactive.14,20 All of these factors could contribute to excess risk of deaths among individuals with low income and potentially explain racial disparities in longevity.21 Few studies, however, have assessed potential modifying effects of individual health behaviors on the association of income and mortality outcomes, particularly among racial/ethnic minority populations who suffer disproportionately from poverty.

The Southern Community Cohort Study (SCCS) is a large prospective cohort study designed to investigate determinants of disparities in cancer and other chronic diseases in underserved populations in the U.S.22 More than half of the study participants reported an annual household income <$15,000 at enrollment, and approximately two-thirds of cohort members are Black participants, providing a unique opportunity to evaluate the impact of extreme poverty on mortality in a racially diverse low-income population. SCCS data was used to first quantify the association between income and overall/cause-specific mortality and further assess potential modifying effects of behavioral risk factors on this association.

METHODS

Study Population

A detailed description of the SCCS has been published.22 The SCCS was approved by IRBs at Vanderbilt University and Meharry Medical College. All participants provided written informed consent. Briefly, nearly 86,000 participants, aged 40–79 years and not under treatment for cancers within one year before study baseline, were enrolled during 2002–2009 primarily (86%) in partnership with federally qualified community health centers (CHC) across 12 southeastern states that provide health care to low-income populations. The remaining 14% of subjects were recruited from stratified random samples of residents in the same 12 states. Data on household incomes, demographic, medical, lifestyle, and other characteristics were obtained using structured questionnaire at baseline. For the current analysis, participants who died within the first two years after completing the baseline survey to reduce potential bias due to reverse causality (n=1,948), or with missing values for household income (n=2,615) or smoking status (n=2,092) (not mutually exclusive) were excluded, leading to a final study sample of 79,385.

Measures

During the baseline survey of SCCS, each participant was asked “Which of the following describes your total household income last year?”, and there were five choices: <$15,000, $15,000–$24,999, $25,000–$49,999, $50,000–$99,999, or ≥ $100,000. In this current analysis, participants were classified into four groups to avoid sparse distributions: <$15,000, $15,000–$24,999, $25,000–$49,999, or ≥ $50,000. Poverty in this study was defined as having an annual household income below $15,000, which is below the federal poverty line for the vast majority families during the study recruitment period.23 Details for household income measurement are presented in Appendix Methods.

Five behavioral factors measured at baseline were included for their well-established associations with mortality: cigarette smoking, alcohol drinking, physical activity, sedentary behavior, and diet quality. Details on health behavior ascertainment and variable categorization in statistical analyses are described in Appendix Methods. Briefly, participants were classified into four groups based on their smoking status: never smokers, ever smokers, current light smokers, and current heavy smokers. Current smokers who had smoked ≥20 years and ≥20 cigarettes/day were considered as heavy smokers, otherwise as light smoker. Participants reporting >0 drink/day but ≤1 for women or ≤2 for men were considered as moderate drinkers, otherwise as heavy drinkers. For leisure-time physical activity (LTPA), three categories were defined: inactive, fairly active, and active. Participants who did not have any LTPAs were considered as inactive. Those reporting ≥ 150 min/week of moderate activity or ≥ 75 min/week of vigorous activity were considered as active, otherwise as fairly active. Total sitting time per day (hours) was used for assessment of sedentary behaviors. A healthy eating index (HEI) that assessed by compliance with the US Dietary Guidelines for Americans 2010 was used to measure the overall diet quality.24 To measure an individual’s overall lifestyle, the five health behaviors were then combined into a composite lifestyle score for each individual, by taking the sum of the negative of regression coefficients associated with all-cause mortality from the fully adjusted model.25 This coefficient-based score ranged from −0.06 to 1.27, with a higher value representing a healthier lifestyle. Participants were then classified into four quartile groups based on their lifestyle scores: Q1 (unhealthiest), Q2, Q3, and Q4 (healthiest). Authors also created a second lifestyle score (lifestyle score 2) based on the prior knowledge from lifestyle guidelines,26 which ranged from 0 to 5, showing the number of guidelines met by participants across five lifestyle factors. Compared with the lifestyle score, the lifestyle score 2 is more intuitive for public health recommendation but does not fully capture the different magnitude of the associations by lifestyle factors. Authors performed sensitive analyses using the lifestyle score 2. Details for score calculations are presented in Appendix methods.

Information on vital status and causes of deaths was obtained via linkage of the cohort to the National Death Index till December 31, 2019.22 The primary outcomes for this study are all-cause mortality and major cause-specific mortality. Causes of death were grouped according to ICD-10 codes: cardiovascular disease (CVD) (I00–I69), cancers (C00–C97), all other-disease causes excluding CVD and cancers (deaths with codes beginning with the letter D–N), and external causes, such as accidents and injuries (deaths with codes beginning with the letter V, W, X, or Y).

Statistical Analysis

Hazard ratios (HR) and 95% confidence interval (CI) were estimated using Cox proportional hazard regression to assess overall and race-specific associations between household income and all-cause or cause-specific mortality, with ages when individuals entered and exited the cohort as the time scale and stratified by birth year (categorized into 10-year groups). The age at exit was defined as the age at death or December 31, 2019 (the end of follow-up), whichever came first. For associations between household income and cause-specific mortality, the subdistribution hazard models were applied to control for competing risks.27 The base model for income-mortality associations was adjusted for enrollment source (CHC, general population), sex (male, female), racial group (White participants, Black participants, all other participants), and household size (1, 2, 3, 4, ≥5) (Model 1). To evaluate the impacts of sociodemographic and lifestyle factors, additional covariate sets were sequentially included to the base model: education (< high school, high school, > high school), marital status (married, divorced/separated, widowed, single), insurance coverage (yes, no), and area deprivation index (ADI, quartiles; a composite measure indicating socioeconomic disadvantage of individuals’ neighborhoods) in Model 2, and lifestyle score, and body mass index (BMI; <25, 25–30, >30) in Model 3. A sensitivity analysis was conducted by adjusting for five individual lifestyle components instead of lifestyle score in Model 3. Also, all the above covariates were finally included in the fully adjusted model (Model 4). Missing values (0.3%–5.8%) were imputed using multiple imputation chained equation (R package: MICE, M=1), under the assumptions of missing at random.28 Sample sizes for participants missing covariate data are as follows: race n=385, education n=375, marital status n=531, household size n=946, insurance coverage n=449, ADI n=2,233, alcohol intake n=1,325, LTPA n=1,468, diet quality n=4,902, sedentary time n=968, and BMI n=957. Frequency distributions of baseline characteristics were tabulated and compared across income groups. Interactions between household income and lifestyle score, lifestyle score 2, race, age, sex, baseline comorbidities, and ADI were assessed by likelihood ratio tests to compare main effects models with and without the addition of cross-product terms. Trend tests were conducted by treating the categorical income variable as continuous in the model. The proportional hazards assumption was evaluated graphically using the Schoenfeld residuals and was considered met. Sensitivity analyses were conducted among those without any missing values and by including individuals with missing values on smoking, income and those died within the first two years. Stratification analyses were conducted by age, sex, ADI of participants’ residence, and disease status at baseline to investigate their potential modifications on the income-mortality associations. Statistical analysis was conducted from February 2022 to January 2023. All statistical analyses were performed using STATA 15.0 and R 4.1.1, and a p-value <0.05 was considered statistically significant.

RESULTS

Over a median follow-up of 12.1 (2.0–17.8) years, a total of 19,749 deaths were recorded among cohort members, including deaths from CVD (n=9,781), cancer (n=4,980), other diseases (n=3,707), and external causes (n=1,105). Only 9.83% of participants had an annual household income ≥$50,000 at baseline, while more than half fell into the poverty group (<$15,000, 54.74%) (Table 1). Individuals with a lower household income tended to be younger, women, Black, unmarried, and more likely to have lower education levels, lower insurance coverage, and unhealthier lifestyle behaviors, except for sedentary time.

Table 1.

Selected Baseline Characteristics of Participants by Annual Household Income, the Southern Community Cohort Study

Baseline characteristics Whole cohort <$15,000 $15,000–$24,999 $25,000–$49,999 ≥ $50,000 p-value
Participants, N 79,385 43,436 16,807 11,328 7,814
Age, years 52.2 (8.7) 52.0 (8.8) 51.6 (8.7) 52.6 (8.7) 53.9 (8.2) <0.001
Enrollment source <0.001
 Community health center 86.1 95.2 90.0 73.6 44.9
 General population 13.9 4.8 10.0 26.4 55.1
Women 59.7 60.6 61.6 59.6 50.8 <0.001
Racial groups <0.001
 Black 65.9 71.2 69.7 59.4 37.4
 White 30.2 25.3 26.5 35.8 57.4
 Other 3.9 3.5 3.8 4.8 5.2
Household size <0.001
 1 26.2 32.1 22.4 19.8 11.0
 2 33.7 30.8 33.3 36.8 46.2
 3–4 29.7 27.1 32.5 32.9 33.8
 ≥5 10.4 10.0 11.8 10.5 9.0
Education <0.001
 < High school 28.5 40.6 22.3 9.4 2.3
 High school 38.6 39.8 46.0 36.7 18.2
 > High school 32.9 19.6 31.7 53.9 79.5
Marital status <0.001
 Married 35.5 22.6 37.6 52.1 78.8
 Divorced 33.6 38.8 33.5 28.3 12.9
 Widowed 9.7 11.9 9.1 6.8 2.9
 Single 21.2 26.7 19.8 12.8 5.4
Insurance coverage 60.6 52.8 56.5 73.7 93.3 <0.001
Obesitya 44.7 44.2 47.5 46.9 37.9 <0.001
Current smokers 40.8 49.2 39.7 28.2 14.2 <0.001
Non/moderate alcohol drinkers 82.3 79.9 83.5 87.0 86.0 <0.001
Physically activeb 19.6 15.2 20.1 25.3 36.2 <0.001
Sedentary timec, hours 9.3 (5.1) 8.9 (5.1) 9.6 (5.1) 10.0 (5.1) 10.0 (4.5) <0.001
Healthy eating indexd 57.8 (12.0) 56.2 (11.7) 57.7 (11.8) 60.1 (12.2) 63.2 (12.1) <0.001
Area deprivation indexe 77.3 (21.7) 80.7 (20.1) 78.4 (20.4) 74.0 (21.3) 60.8 (24.8) <0.001
Lifestyle scoref 0.6 (0.3) 0.6 (0.3) 0.6 (0.3) 0.7 (0.3) 0.8 (0.3) <0.001

Notes: Results were presented as percentage/mean (standard deviation). p-value were computed by χ2 for categorical variables and ANOVA for continuous variables. Boldface indicates statistical significance (p<0.05).

a

Having a body mass index >30 kg/m2.

b

Reporting ≥150 min/week of moderate activity or ≥75 min/week of vigorous activity.

c

Average sitting time per day, including all sedentary activities over a week.

d

A composite measure of diet quality as assessed by compliance with the US Dietary Guidelines for Americans 2010, ranging from 0 to 100.

e

Constructed by Census 2000 tract characteristics, representing rankings (0–100) of neighborhoods by socioeconomic disadvantage.

f

A coefficient-weighted score summarizing smoking status, alcohol intake, physical activity, diet quality, and sedentary time, with a range from −0.06 to 1.27 and a higher value representing a healthier lifestyle.

A dose-response association between household income and all-cause mortality was observed, with the poverty group having a more than 3-fold elevated mortality compared with those with a household income ≥$50,000 after adjusting for enrollment source, sex, and household size (Table 2). Although the dose-response association remained strong, adjusting for additional other sociodemographic variables (Model 2), lifestyle factors (Model 3), or both sets of variables (Model 4) attenuated the associations. The fully adjusted HRs estimated for the poverty group versus the highest income group were 2.4 (95%CI=2.2–2.6) for all participants, 2.2 (95%CI=1.9–2.4) for Black participants, and 2.5 (95%CI=2.2–2.8) for White participants, respectively. Similar results were identified in sensitivity analyses adjusting for individual lifestyle components (Appendix Table 1, available online). Sensitivity analyses conducted among those without any missing values and including individuals with imputed missing values for smoking status and household income as well as those who died within the first 2 years of the baseline survey, yielded similar results (Appendix Table 2, available online). Dose-response associations of household income were observed for all major cause-specific mortality outcomes (Appendix Table 3, available online). The associations, however, were strongest for other-disease causes (HR = 3.0; 95%CI=2.5–3.6) and weakest for cancer (HR=1.7, 95%CI=1.5–2.0).

Table 2.

Associations Between Household Income and All-Cause Mortality by Race, the Southern Community Cohort Study

Annual household income, $ No. of participants No. of Deaths HR (95%CI)a HR (95%CI)b HR (95%CI)c HR (95%CI)d
All participants
 ≥50,000 7,814 891 1 (Ref.) 1 (Ref.) 1 (Ref.) 1 (Ref.)
 25,000–49,999 11,328 1,947 1.7 (1.6, 1.8) 1.6 (1.5, 1.7) 1.5 (1.4, 1.6) 1.5 (1.4, 1.6)
 15,000–24,999 16,807 3,745 2.4 (2.2, 2.6) 2.1 (2.0, 2.3) 2.0 (1.8, 2.1) 1.9 (1.8, 2.1)
 <15,000 43,436 13,166 3.3 (3.1, 3.6) 2.9 (2.7, 3.1) 2.6 (2.4, 2.8) 2.5 (2.3, 2.7)
  p-trend <0.001 <0.001 <0.001 <0.001
Black participants
 ≥50,000 2,925 340 1 (Ref.) 1 (Ref.) 1 (Ref.) 1 (Ref.)
 25,000–49,999 6,733 1,100 1.5 (1.4, 1.7) 1.5 (1.3, 1.7) 1.4 (1.3, 1.6) 1.4 (1.2, 1.6)
 15,000–24,999 11,702 2,417 2.0 (1.8, 2.2) 1.9 (1.6, 2.1) 1.8 (1.6, 2.0) 1.7 (1.5, 2.0)
 <15,000 30,917 8,915 2.8 (2.5, 3.1) 2.5 (2.2, 2.8) 2.4 (2.1, 2.6) 2.2 (2, 2.5)
  p-trend <0.001 <0.001 <0.001 <0.001
White participants
 ≥50,000 4,478 510 1 (Ref.) 1 (Ref.) 1 (Ref.) 1 (Ref.)
 25,000–49,999 4,046 771 1.7 (1.5, 1.9) 1.5 (1.4, 1.7) 1.4 (1.3, 1.6) 1.4 (1.3, 1.6)
 15,000–24,999 4,463 1,196 2.6 (2.3, 2.9) 2.2 (2.0, 2.5) 2.0 (1.8, 2.2) 2.0 (1.8, 2.2)
 <15,000 10,985 3,835 3.7 (3.3, 4.1) 3.1 (2.7, 3.4) 2.6 (2.4, 2.9) 2.5 (2.3, 2.9)
  p-trend <0.001 <0.001 <0.001 <0.001

Notes: HRs were estimated using Cox models. Results for other racial groups were not presented in the table due to a small number. Race was adjusted in the analysis including all participants. Boldface indicates statistical significance (p<0.05).

a

Model 1: adjusted for enrollment source, sex, race, and household size;

b

Model 2: adjusted for enrollment source, sex, race, household size, education, marital status, insurance coverage, and area deprivation index;

c

Model 3: adjusted for enrollment source, sex, race, household size, lifestyle score, and body mass index;

d

Model 4: adjusted for enrollment source, sex, race, household size, education, marital status, insurance coverage, area deprivation index, lifestyle score, and body mass index.

Lifestyle score is a coefficient-weighted variable summarizing smoking status, alcohol intake, physical activity, diet quality, and sedentary time, with a range from −0.06 to 1.27 and a higher value representing a healthier lifestyle.

CI, confidence interval; HR, hazard ratio; No, number.

All of the five lifestyle factors were significantly associated with all-cause mortality (Appendix Table 4, available online). A dose-response association was found between all-cause mortality and the healthy lifestyle score that summarized the five lifestyle factors (Appendix Table 5, available online). After adjusting for income, mortality risk was 2.3-fold higher (95%CI = 2.2–2,4) among those in the lowest versus highest lifestyle score quartile. Mortality monotonically declined with increasing lifestyle score quartile within each income group (Table 3). For example, among participants having a household income of ≥$50,000, those with the highest lifestyle score were at 73% (HR: 0.27, 95%CI: 0.22–0.33) lower risk of all-cause deaths compared to those with the lowest lifestyle score. Also, an HR of 0.82 (95%CI=0.69–0.97) for all-cause deaths was found for individuals who were in poverty but had the highest healthy lifestyle score compared with those in ≥$50,000 group but had the lowest healthy lifestyle score. On the other hand, both lower household income and lower healthy lifestyle score were associated with higher all-cause mortality (Appendix Figure 2, available online). Individuals who were in poverty and had the lowest lifestyle score were at 6.1-fold (95%CI=5.4–6.9) increase in all-cause mortality compared with those in ≥$50,000 group and had the highest lifestyle score (Appendix Table 6, available online). Similar patterns of the association were found for Black and White participants, as well as for CVD mortality (Appendix Table 6, Appendix Table 7, available online). Sensitivity analyses for lifestyle score based on health guidelines showed similar but somewhat weaker results (Appendix Table 8, available online).

Table 3.

The Joint Association of Household Income and Lifestyle Score With All-Cause Mortality by Race, the Southern Community Cohort Study

Annual household income, $ Lifestyle Score
Q1 (Unhealthiest)
Q2
Q3
Q4 (Healthiest)
No. of Deaths HR (95%CI) No. of Deaths HR (95%CI) No. of Deaths HR (95%CI) No. of Deaths HR (95%CI)
All participants

 ≥50,000 144 1 (Ref.) 189 0.57 (0.45, 0.69) 276 0.38 (0.29, 0.44) 282 0.27 (0.22, 0.33)

 25,000–49,999 436 1.13 (0.92, 1.32) 526 0.81 (0.66, 0.95) 563 0.58 (0.47, 0.67) 422 0.42 (0.35, 0.51)

 15,000–24,999 1,124 1.39 (1.14, 1.62) 1,069 1.02 (0.85, 1.20) 908 0.75 (0.60, 0.86) 644 0.58 (0.49, 0.70)

 <15,000 4,731 1.69 (1.41, 1.97) 3,837 1.32 (1.09, 1.52) 2,806 0.99 (0.83, 1.16) 1,792 0.82 (0.69, 0.97)

Black participants

 ≥50,000 59 1 (Ref.) 87 0.63 (0.45, 0.87) 97 0.34 (0.25, 0.48) 97 0.31 (0.23, 0.42)

 25,000–49,999 202 0.92 (0.69, 1.23) 323 0.79 (0.60, 1.04) 312 0.59 (0.45, 0.78) 263 0.47 (0.36, 0.62)

 15,000–24,999 619 1.17 (0.90, 1.52) 758 0.95 (0.73, 1.23) 578 0.69 (0.53, 0.90) 462 0.60 (0.46, 0.78)

 <15,000 2,873 1.45 (1.12, 1.87) 2,742 1.19 (0.93, 1.53) 1,960 0.95 (0.74, 1.23) 1,340 0.80 (0.62, 1.04)

White participants

 ≥50,000 76 1 (Ref.) 95 0.53 (0.39, 0.72) 167 0.39 (0.30, 0.51) 172 0.28 (0.21, 0.37)

 25,000–49,999 218 1.36 (1.05, 1.76) 180 0.78 (0.60, 1.02) 230 0.53 (0.41, 0.69) 143 0.36 (0.27, 0.48)

 15,000–24,999 452 1.64 (1.28, 2.10) 273 1.09 (0.84, 1.40) 301 0.75 (0.58, 0.97) 170 0.55 (0.42, 0.72)

 <15,000 1,702 2.04 (1.61, 2.58) 968 1.42 (1.12, 1.81) 769 0.99 (0.78, 1.26) 396 0.78 (0.61, 1.01)

Notes: HRs were estimated using Cox models. Models were adjusted for enrollment source, sex, race, household size, area deprivation index, education, marital status, insurance status, and body mass index. Race was adjusted in the analysis including all participants. Results for other racial groups were not presented in the table due to a small number.

Lifestyle score is a coefficient-weighted variable summarizing smoking status, alcohol intake, physical activity, diet quality, and sedentary time, with a range from −0.06 to 1.27 and a higher value representing a healthier lifestyle.

CI, confidence interval; HR, hazard ratio; No, number.

Figure 1A showed cumulative all-cause mortality by race and annual household income. Participants in the <$15,000 group had a consistent higher mortality compared to those with income ≥$50,000 regardless of race. However, among those in the relatively high-income group (≥$50,000), White participants had a lower cumulative mortality risk than Black participants, while among those in poverty, Whites had a higher cumulative mortality than Blacks (p for interaction, <0.001). Approximately 30% of Whites and Black participants had died, with average ages of 82.7 and 81.2 years, respectively, in the income group of ≥$50,000, while the corresponding ages were 67.8 and 70.9 years in the <$15,000 group, resulting in a difference of more than 10 years in life expectancy associated with poverty. Figure 1B showed cumulative all-cause mortality by lifestyle and annual household income. As expected, participants experiencing poverty and living with the least healthy lifestyle had a consistently higher mortality risk compared to individuals with the highest income (≥$50,000) and healthiest lifestyle score. However, participants experiencing poverty but having the healthiest lifestyle score had a lower cumulative mortality than individuals with the highest income and lowest lifestyle score (p for interaction, <0.001).

Figure 1.

Figure 1.

Adjusted cumulative all-cause mortality by annual household income levels with race and lifestyle groups, the Southern Community Cohort Study. (A). All-cause mortality by annual household income levels and race. (B). All-cause mortality by annual household income levels and the lifestyle score.

Stratified analyses were further performed to evaluate whether the positive associations between household income and all-cause mortality were modified by age, sex, ADI of participants’ residence and disease status at baseline (Appendix Figure 1, available online). Stronger associations were observed in participants who were women, younger than 65 years, had no comorbidities at baseline, and lived in less deprived areas. All interaction tests were statistically significant.

DISCUSSION

In this large prospective cohort study conducted in a predominantly low-income and Black population, low household income was strongly associated with elevated risk of all-cause and major cause-specific mortality. Healthy lifestyles were shown to significantly reduce mortality and partially negate the adverse effects of low income on mortality outcome. Compared to those having an annual household income ≥$50,000, individuals in poverty (<$15,000) had a more than 10-year decrease in life expectancy. This estimate is alarming and calls for evaluation of policy and public health strategies to reduce mortality and health disparities in low-income Americans.

To the best of current knowledge, this is the first study conducted in an extremely low-income and predominantly Black population to quantify the effect of poverty on total and cause-specific mortality risk. The findings for a substantially elevated mortality associated with low household income, even after adjusting for other risk factors, are supported, in general, by previous studies conducted in middle and upper-middle income populations.13,15,17 This current study, with a large sample size, extended the dose-response association between household income and mortality to extreme low-income groups.

Poor White participants experienced higher all-cause mortality than poor Black participants. Compared with their counterparts in the highest income group, the impoverished Whites and Blacks had a decreased longevity of 14.7 years and 10 years at a 30% cumulative mortality, respectively. A more than 10 years of difference in life expectancy was also reported in a large study comparing the 1% wealthiest Americans with the 1% poorest.6 However, that study did not report any racial disparities. Reasons for the racial difference in mortality in low-income populations are not entirely clear and further studies are needed.

Healthy behaviors can help to mitigate some of the risk associated with lower income. However, healthy behaviors alone cannot fully mitigate the adverse effects of poverty on mortality. This study calls for multifaceted strategies to reduce mortality disparities, not only by implementing relevant government policies to reduce income inequality, but also through lifestyle modifications. People living below the poverty level face several challenges, including limited availability of healthy food, inadequate access to a safe walkable environment, insufficient access to green space, and an excessive presence of cigarette and alcohol advertisements.1719 Tailored preventive strategies should be implemented specifically for this population and early establishment of a healthy lifestyle is encouraged to maximize the potential benefits and mitigate the burden of mortality caused by income inequality.

Limitations

The major strength of this study was the inclusion of large numbers of participants who lived at or below poverty thresholds, which provides a unique opportunity to investigate the impact of poverty in adulthood on total and cause-specific mortality in this under-representative population. Comprehensive data on healthy behavioral factors were collected at the outset, which made it possible to evaluate the combined associations of multiple lifestyle factors and poverty on mortality outcomes. The sample size of this study was large, and the statistical power was high. There were some limitations of this study. First, information on household income was self-reported and collected in five levels (categories), which limited the ability to estimate per capita income for association analyses. However, the vast majority of families with a household income of $15,000 in the study are indeed in poverty and suffer from substantial financial stress. Second, all cancers were combined as a single mortality outcome in the study, and the results might be driven by some major cancer sites that are differentially related to socioeconomic status. Future studies could further investigate the associations of income with incidence and mortality of site-specific cancers. Last, reverse causation and unmeasured confounding may still be present due to the observational nature of this study. In addition, future studies could also be conducted to investigate potential influences of material well-being and psychosocial stressors in the association between poverty and mortality outcomes.

CONCLUSIONS

Over the past four decades, income inequality in the USA has increased, along with the enlarging health disparity across SES groups.4 Life expectancy has risen among middle- and high-income Americans while stagnated or even declined among the poor.6 Without interventions on this specific population to decouple low income and health, further widening and hardening SES gaps might be seen in health. This study, based on the SCCS data, quantified mortality differentials by income and provided valuable evidence for elimination of the gaps in life expectancy in the USA. However, addressing racial/income disparities in mortality requires comprehensive research and policy considerations, including improving access to healthcare, addressing social determinants of health, promoting health equity policies. Additionally, efforts to improve income equality are crucial in reducing mortality and health disparities among low-income Americans.

Supplementary Material

Appendix of tables and figures

ACKNOWLEDGMENTS

The authors thank the study participants and research team members for their contributions to the study, Mr. Fangcheng Yuan for his help in data analysis and manuscript preparation, and Rachel Mullen for technical supports in preparing the manuscript. Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number U01CA202979. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare no conflicts of interest. No financial disclosures have been reported by the authors of this paper.

Footnotes

CREDIT AUTHOR STATEMENT

Lili Liu: Conceptualization, Methodology, Formal analysis, Writing – original draft, Writing – review & editing, Visualization. Wanqing Wen: Methodology, Formal analysis, Writing – review & editing. Martha J. Shrubsole: Funding acquisition, Writing – review & editing. Loren E. Lipworth: Writing – review & editing. Michael T. Mumma: Writing – review & editing. Brooke A. Ackerly: Writing – review & editing. Xiao-Ou Shu: Writing – review & editing. William J. Blot: Funding acquisition, Writing – review & editing. Wei Zheng: Conceptualization, Methodology, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Resources, Funding acquisition.

SUPPLEMENTAL MATERIAL

Supplemental materials associated with this article can be found in the online version at https://doi.org/10.1016/j.amepre.2024.02.015.

REFERENCES

  • 1.Stringhini S, Sabia S, Shipley M, et al. Association of socioeconomic position with health behaviors and mortality. JAMA. 2010;303(12):1159–1166. 10.1001/jama.2010.297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Stringhini S, Carmeli C, Jokela M, et al. Socioeconomic status and the 25 × 25 risk factors as determinants of premature mortality: a multicohort study and meta-analysis of 1.7 million men and women. Lancet. 2017;389(10075):1229–1237. 10.1016/s0140-6736(16)32380-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Pillay-van Wyk V, Bradshaw D. Mortality and socioeconomic status: the vicious cycle between poverty and ill health. Lancet Glob Health. 2017;5(9):e851–e852. 10.1016/s2214-109x(17)30304-2. [DOI] [PubMed] [Google Scholar]
  • 4.Bor J, Cohen GH, Galea S. Population health in an era of rising income inequality: USA, 1980-2015. Lancet. 2017;389(10077):1475–1490. 10.1016/s0140-6736(17)30571-8. [DOI] [PubMed] [Google Scholar]
  • 5.Kinge JM, Modalsli JH, Øverland S, et al. Association of household income with life expectancy and cause-specific mortality in Norway, 2005-2015. JAMA. 2019;321(19):1916–1925. 10.1001/jama.2019.4329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Chetty R, Stepner M, Abraham S, et al. The association between income and life expectancy in the United States, 2001-2014. JAMA. 2016;315(16):1750–1766. 10.1001/jama.2016.4226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Chokshi DA. Income, poverty, and health inequality. JAMA. 2018;319(13):1312–1313. 10.1001/jama.2018.2521. [DOI] [PubMed] [Google Scholar]
  • 8.Ridley M, Rao G, Schilbach F, Patel V. Poverty, depression, and anxiety: causal evidence and mechanisms. Science. 2020;370(6522):eaay0214. 10.1126/science.aay0214. [DOI] [PubMed] [Google Scholar]
  • 9.Niessen LW, Mohan D, Akuoku JK, et al. Tackling socioeconomic inequalities and non-communicable diseases in low-income and middle-income countries under the Sustainable Development agenda. Lancet. 2018;391(10134):2036–2046. 10.1016/s0140-6736(18)30482-3. [DOI] [PubMed] [Google Scholar]
  • 10.Lund C, De Silva M, Plagerson S, et al. Poverty and mental disorders: breaking the cycle in low-income and middle-income countries. Lancet. 2011;378(9801):1502–1514. 10.1016/s0140-6736(11)60754-x. [DOI] [PubMed] [Google Scholar]
  • 11.Williams J, Allen L, Wickramasinghe K, Mikkelsen B, Roberts N, Townsend N. A systematic review of associations between non-communicable diseases and socioeconomic status within low- and lower-middle-income countries. J Glob Health. 2018;8(2):020409. 10.7189/jogh.08.020409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Schwandt H, Currie J, von Wachter T, Kowarski J, Chapman D, Woolf SH. Changes in the relationship between income and life expectancy before and during the COVID-19 pandemic, California, 2015-2021. JAMA. 2022;328(4):360–366. 10.1001/jama.2022.10952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Zonderman AB, Mode NA, Ejiogu N, Evans MK. Race and poverty status as a risk for overall mortality in community-dwelling middle-aged adults. JAMA Intern Med. 2016;176(9):1394–1395. 10.1001/jamainternmed.2016.3649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Zhang YB, Chen C, Pan XF, et al. Associations of healthy lifestyle and socioeconomic status with mortality and incident cardiovascular disease: two prospective cohort studies. BMJ. 2021;373:n604. 10.1136/bmj.n604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Makaroun LK, Brown RT, Diaz-Ramirez LG, et al. Wealth-associated disparities in death and disability in the United States and England. JAMA Intern Med. 2017;177(12):1745–1753. 10.1001/jamainternmed.2017.3903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Yoshikawa H, Aber JL, Beardslee WR. The effects of poverty on the mental, emotional, and behavioral health of children and youth: implications for prevention. Am Psychol. 2012;67(4):272–284. 10.1037/a0028015. [DOI] [PubMed] [Google Scholar]
  • 17.Glei DA, Lee C, Weinstein M. Assessment of mortality disparities by wealth relative to other measures of socioeconomic status among US adults. JAMA Netw Open. 2022;5(4):e226547. 10.1001/jamanetworkopen.2022.6547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Warren Andersen S, Blot WJ, Shu XO, et al. Adherence to cancer prevention guidelines and cancer risk in low-income and African American populations. Cancer Epidemiol Biomarkers Prev. 2016;25(5):846–853. 10.1158/1055-9965.Epi-15-1186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Patel SA, Krasnow M, Long K, Shirey T, Dickert N, Morris AA. Excess 30-day heart failure readmissions and mortality in black patients increases with neighborhood deprivation. Circ Heart Fail. 2020;13(12):e007947. 10.1161/circheartfailure.120.007947. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Petrovic D, de Mestral C, Bochud M, et al. The contribution of health behaviors to socioeconomic inequalities in health: a systematic review. Prev Med Aug. 2018;113:15–31. 10.1016/j.ypmed.2018.05.003. [DOI] [PubMed] [Google Scholar]
  • 21.Himmelstein KEW, Lawrence JA, Jahn JL, et al. Association between racial wealth inequities and racial disparities in longevity among US adults and role of reparations payments, 1992 to 2018. JAMA Netw Open. 2022;5(11):e2240519. 10.1001/jamanetworkopen.2022.40519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Signorello LB, Hargreaves MK, Steinwandel MD, et al. Southern community cohort study: establishing a cohort to investigate health disparities. J Natl Med Assoc. 2005;97(7):972–979. [PMC free article] [PubMed] [Google Scholar]
  • 23.The US Census Bureau. Poverty thresholds by size of family and number of children. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-poverty-thresholds.html.
  • 24.Yu D, Sonderman J, Buchowski MS, et al. Healthy eating and risks of total and cause-specific death among low-income populations of African-Americans and other adults in the southeastern United States: a prospective cohort study. PLoS Med. 2015;12(5):e1001830. 10.1371/journal.pmed.1001830. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Warren Andersen S, Zheng W, Sonderman J, et al. Combined impact of health behaviors on mortality in low-income Americans. Am J Prev Med. 2016;51(3):344–355. 10.1016/j.amepre.2016.03.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kushi LH, Doyle C, McCullough M, et al. American Cancer Society Guidelines on nutrition and physical activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin. 2012;62(1):30–67. 10.3322/caac.20140. [DOI] [PubMed] [Google Scholar]
  • 27.Austin PC, Fine JP. Practical recommendations for reporting Fine-Gray model analyses for competing risk data. Stat Med. 2017;36(27):4391–4400. 10.1002/sim.7501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Azur MJ, Stuart EA, Frangakis C, Leaf PJ. Multiple imputation by chained equations: what is it and how does it work? Int J Methods Psychiatr Res. 2011;20(1):40–49. 10.1002/mpr.329. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix of tables and figures

RESOURCES