Population Health, Equity & Outcomes
- June 2026
- Volume 32
- Issue Spec. No. 6
- Pages: eSP22-eSP32
Associations Between Socioeconomic Status, Healthy Lifestyle, and Multimorbidity in the US
Middle socioeconomic status (SES) elevates multimorbidity risk; healthy lifestyles mitigate this. Targeted interventions for middle-SES populations, especially middle-aged non-Hispanic White women, are crucial.
ABSTRACT
Objective: To investigate the relationships between socioeconomic status (SES) and multimorbidity, which can inform public health strategies.
Study Design: Cross-sectional study.
Methods: Data from participants in the National Health and Nutrition Examination Survey 2007-2018 were analyzed. Weighted multivariate logistic regression assessed associations between multimorbidity, SES, and healthy lifestyles. Subgroup analyses stratified by key covariates evaluated stratified risks. Sensitivity analyses verified result robustness.
Results: After adjustment, compared with low-SES individuals, medium-SES individuals showed a significantly increased OR of multimorbidity (1.267; 95% CI, 1.079-1.489), whereas high-SES individuals exhibited a significantly decreased OR (0.723; 95% CI, 0.647-0.807). Healthy lifestyles exerted a protective effect, particularly within the medium-SES subgroup. Middle-aged non-Hispanic White women in the medium-SES group demonstrated elevated risk.
Conclusions: Moderate SES is a critical risk factor for multimorbidity. Lifestyle interventions effectively reduce disease burden, particularly in middle-SES populations, with targeted protection needed for middle-aged non-Hispanic White women. Additionally, heightened vigilance is warranted regarding the expanding disease risks among young and middle-aged populations, alongside advocating for robust research initiatives to optimize evidence-based public health interventions.
Am J Manag Care. 2026;32(Spec. No. 6):eSP22-eSP32.
Global socioeconomic development in recent decades has significantly improved material living standards but exacerbated wealth inequality and health disparities.1,2 Socioeconomic status (SES) remains strongly linked to inequalities in disease burden and health outcomes.3 Studies reveal diverging life expectancy trends: Middle- to high-income groups in the US experience sustained longevity gains, whereas impoverished populations face stagnation or decline,4-6 a pattern mirrored in the UK, India, and other nations.4,6 Chronic noncommunicable diseases (NCDs) now dominate the global disease burden, driven by lifestyle risks (eg, physical inactivity, obesity) and aging, with multimorbidity prevalence rising sharply.7,8 Developing countries such as India confront heightened challenges due to rapid lifestyle transitions.9 Although health care systems exhibit structural overreliance on pharmacotherapy for preventable chronic conditions—often neglecting evidence-based lifestyle interventions—significant populations concurrently face barriers to essential medications due to systemic inequities. Addressing this dual failure requires both scaling accessible nonpharmacological therapies and ensuring equitable pharmacotherapy access. Robust evidence underscores the critical role of healthy lifestyle management.10 For instance, cancer research demonstrates that SES shapes lifestyle through living conditions, nutrition, and occupational exposures, creating cascading effects on disease incidence.11,12
Recent years have witnessed growing scholarly interest in multimorbidity patterns, particularly within populations affected by chronic NCDs. This state is generally characterized by the concurrent manifestation of 2 or more persistent health conditions in an individual patient.13 Clinical investigations demonstrate that the coexistence of multiple chronic disorders substantially compromises health-related well-being, amplifies susceptibility to functional impairment and premature death, and exhibits strong correlations with terminal health outcomes.14-16 Evidence indicates that individuals with multimorbidity demonstrate significantly elevated risks of early mortality, recurrent hospital admissions, and extended inpatient care durations compared with monomorbid patients.7 The complexity of multimorbidity creates multidimensional burdens spanning individual patients, household units, health care delivery mechanisms, and socioeconomic structures, with disproportionate effects observed in underfunded health ecosystems.7,17 Common chronic diseases include cardiovascular diseases, kidney diseases, Following standardized protocols, participants completed structured household interviews before transitioning to clinical assessments combined with biological sample procurement in mobile examination centers. Ethical clearance was secured through the National Center for Health Statistics (NCHS), with informed consent acquired from all study enrollees prior to data collection. Additional details on NHANES consent protocols and the data that support the findings of this study are available on the NCHS website. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology reporting standards for population-based observational studies.
Exclusion criteria were (1) younger than 20 years, (2) pregnancy, (3) missing or incomplete chronic disease data, (4) incomplete SES or lifestyle data, and (5) mortality data not available for publication. From 34,770 initially eligible individuals in 2007-2018, 26,775 participants met the inclusion criteria for last analysis (Figure).
Chronic NCDs
The primary outcome was multimorbidity (ie, a person having ≥ 2 chronic NCDs at the same time). Secondary outcomes included the number of NCDs, categorized as 0, 1, 2, or 3 or more. Based on prior methodologies25,26 and consistent with NCD definitions across NHANES waves (2007-2018), the following conditions were evaluated via self-reported questionnaires, while avoiding excessive correlation with SES indicators: arthritis, asthma, ischemic heart disease (angina/coronary artery disease), heart failure, acute coronary events (myocardial infarction), cerebrovascular accidents, hypertension, dyslipidemia, chronic obstructive pulmonary disorders (emphysema/chronic bronchitis), thyroid dysfunction, hyperuricemia, renal pathologies (chronic kidney disease and nephrolithiasis), hepatic disorders, depression, and malignancies (all cancer subtypes combined).
SES Assessment
SES was operationalized through 4 established dimensions informed by prior epidemiological framework5,27,28: (1) household income-to-poverty ratio (PIR), calculated as annual family income relative to federal poverty thresholds (categorized as low [PIR < 1], moderate [1-<4], or high [≥ 4])29; (2) educational attainment (less than high school, high school/equivalent, or college/above); (3) employment status (employed or unemployed based on work activity during survey week); and (4) health insurance coverage (private [employer-based/commercial plans], public [Medicare/Medicaid/military plans], or uninsured).
Subsequently, latent class analysis (LCA) was used to structure a nonmeasurable (ie, latent) variable using multiple observed categorical variables, and the SES was estimated with the 4 variables described above.30 We conducted an exploratory LCA with no prespecified number of classes, and we selected the optimal model solely on the basis of statistical fit indices and interpretability. LCA was implemented via the poLCA (v1.6.0.1) R package to derive SES strata, with model parameters including 10,000 maximum iterations and a convergence tolerance of 1 × 10−10. Competing models with 2 to 6 latent classes were evaluated, although solutions beyond 5 classes failed convergence generally. Optimal classes were determined by combining the Akaike and Bayesian information criteria and the statistic of the likelihood ratios (G2). Study participants underwent tripartite socioeconomic stratification (high/intermediate/low tiers) derived from item response probability metrics and clinical validation criteria. Supplementary material documenting the implementation of the methodology is detailed in eAppendix Figures 1 and 2 and eAppendix Tables 1 and 2().
Healthy Lifestyle Assessment
Lifestyle profiles were constructed using self-reported questionnaire data, incorporating 6 evidence-based behavioral domains established in prior epidemiological research5,19,31: sustained nicotine abstinence, low-risk alcohol consumption, physical metabolic expenditure, nutrient-dense dietary patterns, limited sedentary behavior, and optimal sleep duration. Health-conforming indicators were a dichotomous scale (1 point per criterion fulfilled), generating a cumulative metric scaled 0 to 6, where elevated values corresponded to indicate healthier patterns in behavior. The cohort underwent tripartite stratification: excellent lifestyle (4-6 points), moderate lifestyle (2-3 points), and inferior lifestyle (0-1 points). Detailed operational definitions and validation metrics are provided in eAppendix Tables 3 and 4.
Covariates
Covariates included age-stratified categories (20-44, 45-64, ≥ 65), biological sex, racial/ethnic categories (Mexican American, non-Hispanic Black, non-Hispanic White, others), and marital status.
Statistical Analysis
Analyses followed NHANES analytic guidelines for weighted data. Classification variables are presented as frequencies (%), normally distributed continuous variables as mean (SD), and nonnormal variables as median (IQR). Between-group comparisons employed χ2 tests, analysis of variance, or Kruskal-Wallis tests, as appropriate.
Weighted binary logistic regression models examined associations between SES and NCD multimorbidity across 3 models: model 1 (unadjusted), model 2 (demographic-adjusted covariates [sex, race, education]), and model 3 (age-adjusted). Outcomes were expressed as OR (95% CI), with stratification by demographic strata and health tiers, and sensitivity analyses assessed robustness. Statistical analyses were performed with R 4.3.2 (R Foundation for Statistical Computing) and IBM SPSS Statistics 25.0 (IBM), with an inference threshold of α = .05.
RESULTS
Population Characteristics
This study included 26,775 participants categorized into low (22%; n = 8197), middle (25%; n = 7503) and high (53%; n = 11,075) SES groups. The study participants had a mean age of 48 years, with 18.84% aged 65 years and older, and 48.52% were male. The majority were non-Hispanic White (68.19%), and 64.09% were married. In terms of lifestyle, 63.82% maintained an intermediate healthy lifestyle, whereas only 22.50% met optimal standards. Chronic disease burden was substantial: 37.49% had at least 3 chronic conditions, and only 10.93% were disease-free. Socioeconomic indicators revealed that 62.28% had higher education, 63.66% were employed, and private insurance accounted for the majority (including mixed insurance, 64.99%) (Table 1).
The middle-SES group had higher proportions of women, older adults (≥ 65 years), non-Hispanic White individuals, married individuals, and those with public insurance only and mixed insurance, with poverty levels concentrated in the PIR range of 1 to less than 4. This group also exhibited the highest chronic disease prevalence. The high-SES group had the highest education attainment (65.55%), wealthiest economic status, highest rate of private insurance, and highest employment rates. The low-SES group showed balanced poverty distribution (similar proportions in PIR < 1 and PIR 1-<4) and the highest uninsured rate. The middle-SES group exhibited a higher incidence of unemployment or nonparticipation, whereas high-SES and low-SES groups demonstrated higher employment rates.
Association of SES and Lifestyle With Multimorbidity
As shown in Table 2, in adjusted logistic regression analyses, socioeconomic disparities in multimorbidity ORs persisted after covariate adjustment. Compared with the low-SES group, moderate-SES individuals exhibited 26.7% increased odds of multimorbidity (OR, 1.267; 95% CI, 1.079-1.489), whereas high SES correlated with a nearly 30% odds reduction (OR, 0.723; 95% CI, 0.647-0.807). Key SES components revealed divergent associations: Compared with being uninsured, mixed insurance (OR, 1.135; 95% CI, 0.732-1.757) and public insurance only (OR, 1.406; 95% CI, 1.153-1.714) showed a positive correlation with the multimorbidity outcome, and only private insurance (OR, 0.838; 95% CI, 0.733-0.958) showed a negative correlation. Lifestyles demonstrated graded protection: Compared with the poor lifestyle tier, moderate and optimal tiers exhibited a positive association, with optimal behaviors (eg, sustained nicotine abstinence: OR, 0.573; 95% CI, 0.493-0.665; healthy diet: OR, 0.620; 95% CI, 0.521-0.739) showing stronger benefits. Notably, a healthy lifestyle significantly reduced the odds of multimorbidity among moderate-SES populations, with optimal lifestyle behaviors showing the strongest protective effect (OR, 0.465; 95% CI, 0.246-0.879) (eAppendix Tables 5-6).
Subgroup analyses identified significant heterogeneity (Table 3). Women exhibited a marginally elevated OR of multimorbidity compared with men (1.297 vs 1.214, respectively). Compared with their low-SES counterparts, middle-SES women and middle-SES men both showed increased risk, though the elevation was marginal. Significant racial disparities were observed: Both non-Hispanic White and non-Hispanic Black individuals within the moderate-SES group had higher ORs of multimorbidity than their low-SES counterparts; notably, the OR for non-Hispanic White individuals (1.163) was significantly greater than that for non-
Hispanic Black individuals. Age significantly confounded the association. After age adjustment (model 3), the effect attenuated, with a significant interaction observed: Middle SES showed a positive age association with multimorbidity, whereas high SES showed an inverse association. Middle-aged and older individuals’ OR decreased significantly by 16.74%. Lifestyle further amplified SES gradients: Moderate-SES individuals with poor lifestyles faced more than double the odds (OR, 2.186; 95% CI, 1.397-3.423). These stratified associations underscore the importance of tailored interventions addressing demographic-specific OR profiles.
Sensitivity Analysis
To verify the robustness of the results, multiple imputation was performed to handle missing values in the original data set. Although the robustness improved after imputation, this may not fully reflect the true underlying situation. After model adjustment, compared with the low-SES group, the odds increased for the medium-SES group, whereas they significantly decreased for the high-SES group (OR, 0.735; 95% CI, 0.661-0.817). Furthermore, multivariable logistic regression with overlap weighting (OW) propensity score adjustment was performed. The fully adjusted model showed increased ORs for medium SES (OR, 1.267; 95% CI, 1.079-1.489) and decreased ORs for high SES (OR, 0.723; 95% CI, 0.647-0.807) relative to low SES. The OW method achieved covariate balance through inverse probability weighting, with stabilized standardized mean differences after adjustment. Scenarios restricting the analysis to specific age ranges were also performed. The effect sizes and directions were generally consistent across methods. Considering the correlation between age and insurance types, the insurance categories were refined by separately classifying Medicare-only and Medicaid-only groups in the analysis. After this refinement, the risk associated with Medicare surged significantly (OR ≈ 20), whereas Medicaid showed only a weak association. The combined insurance indicator shifted from being protective to nonsignificant, with estimates exhibiting substantial fluctuations and a considerable decline in robustness. (See detailed results in eAppendix Tables 7-13.)
DISCUSSION
This nationwide cross-sectional study leveraging NHANES 2007-2018 data reveals significant socioeconomic gradients in NCD multimorbidity risks, with notable modifications by demographic factors and lifestyle behaviors. The findings demonstrate that moderate-SES individuals face disproportionately elevated multimorbidity risks compared with their low-SES counterparts, a pattern potentially mediated by occupational instability, insurance disparities, and suboptimal health behaviors. Conversely, high SES confers protective effects, likely through enhanced health care access and socioeconomic resilience. The constructed 6-component healthy lifestyle score demonstrates robust protective associations, attenuating SES-related disparities by up to 3-fold in moderate-SES subgroups.
The observed SES gradient aligns with previous reports on individual chronic diseases32 yet extends critical insights into multigenerational risk patterns. Notably, 24% of younger adults (aged 20-44 years) exhibit multimorbidity, challenging conventional age-centric risk paradigms. Although occupational hazards and delayed health care–seeking behaviors may explain elevated vulnerability among men,33,34 our stratification reveals unprecedented effect sizes: Moderate-SES men demonstrate a 7% higher OR than women, which is basically consistent with traditional female-predominant morbidity patterns.34 This pattern likely reflects evolving gender roles in occupational stress and health literacy disparities.
Age-stratified analyses identified midlife (45-64 years) as a critical prevention window, where moderate SES confers 64% excess OR elevation compared with the low-SES population. This aligns with biological aging thresholds—mitochondrial dysfunction and cellular senescence accelerate after 45 years, amplifying environmental stressors.35 Younger high-SES adults’ paradoxical risks may reflect occupational burnout in competitive labor markets, warranting targeted workplace Associations Between healthy Socioeconomic Status
