Abstract
Background:
Acculturation and socioeconomic status (SES) are associated with asthma and atopic dermatitis (AD) prevalence in the United States (US), but these relationships and their combined effects are poorly understood in the US Latino population.
Objective:
We sought to estimate the independent and combined contributions of SES and acculturation to asthma and AD prevalence among Latino children in the US.
Methods:
We utilized the National Health Interview Survey (2011–2018), a US population-based survey, to examine the independent and combined associations between proxies for acculturation, SES, and asthma and AD prevalence in US Latino children via summary statistics and a series of survey-adjusted (weighted) logistic regression models.
Results:
English speaking parents and multiple markers of lower SES were independently associated with greater asthma and AD prevalence. The highest disease prevalence was found in children with the combined effects of both greater acculturation and multiple markers of lower SES and the lowest prevalence was found in children with lower acculturation and multiple markers of greater SES.
Conclusion:
Lower SES and proxies for greater acculturation were positively associated with asthma and AD prevalence among US Latino children, and those children who experience both were at the greatest risk.
Keywords: Atopic dermatitis, Asthma, Disparities, Ethnicity, Acculturation, Immigration, Socioeconomic status
Introduction
Risk factors for childhood atopic diseases in Latino children in the United States (US) remain poorly understood.1 Although previous literature has identified poverty and acculturation as two potential risk factors for Latino children, this literature is limited. Studies that have addressed these questions are generally small and/or based on convenience samples from referral clinics and therefore may not be representative of the broader US Latino population.2–8
The US Latino population is young on average, heterogeneous across country of origin, and the largest contributor to US population growth, so is projected to become the largest ethnic group in the US. Asthma and atopic dermatitis (AD) prevalence are also rising faster in Latino children than in non-Latino children.1,9 Furthermore, the characteristics of the US Latino population are shifting; The proportion who are US-born compared with foreign-born is increasing, and the proportion of immigrants who have lived in the US for more than ten years is also increasing.10 These shifting demographic characteristics suggest that acculturation, the process of change that occurs while adapting culturally, socially, and psychologically to a prevailing culture, has increased over time in this population. Any influence that acculturation has on asthma and/or AD risk may also shape temporal trends in these atopic conditions among US Latinos.11,12 All of the above changes within US Latino subgroups may influence atopic disease prevalence and incidence in US Latinos.
Poverty is widely experienced by the Latino population, regardless of immigration status.13 Poverty and acculturation have a complex relationship. They are often inversely correlated and represent distinct concepts.14 Given their relatedness, research findings about acculturation that do not consider poverty, and vice versa, are difficult to interpret.15 Disease burden and disparities in childhood asthma and AD within US Latino children may be underappreciated because there has been less consideration of both socioeconomic status (SES) and acculturation. The heterogeneity of the US Latino population with regard to both SES and acculturation may obscure a greater disease burden among Latino subgroups.3,6,16
To estimate the independent and combined contributions of SES and acculturation to asthma and AD prevalence among Latino children in the US, we analyzed data from a US population-based survey, the National Health Interview Survey, using parental interview language, child birthplace, and child citizenship status as proxies for acculturation.
Methods
National Health Interview Survey
Data from the National Health Interview Survey (NHIS) years 2011–2018 were utilized to perform the analyses.17 This study was exempt from Institutional Review Board review because it used publicly available, de-identified data; informed consent was therefore not required or obtainable.
NHIS asks respondents whether a child is of Hispanic/Latino ethnicity. We use Latino throughout to describe this population, acknowledging that no single term fully reflects all individuals’ preferred ethnic identifiers.18 The analytic population was limited to children aged 0 to 17 years with parent-reported Latino ethnicity, including children reported as Latino regardless of race (including Black). Children were included irrespective of citizenship status.
Asthma was defined using parent report that the child still had asthma, which in NHIS is only asked among children who were ever told by a clinician that they ever had asthma. AD was defined based on parent report that a child had eczema or skin allergy in the past 12 months.
Acculturation Measures
Proxies for acculturation included parent interview language, child’s birthplace, and child’s US citizenship status (eTable 1). NHIS interviews are conducted in English, Spanish, or a combination of English and Spanish; in the public-use files, Spanish-only and English/Spanish interviews are coded together. Interview language may vary across questions depending on respondent proficiency and comfort. For the purposes of this study, interview language was treated as the participants’ preferred language.17
Language use and nativity are commonly used acculturation proxies in population-based Latino health research.5,19,20 English interview language was treated as a marker of greater acculturation, while Spanish or bilingual (English/Spanish) interview language was treated as a marker of lesser acculturation.20 Child birthplace was categorized as US-born versus non-US-born. Child citizenship status was included as an additional acculturation proxy, with acknowledgement that citizenship is also structurally linked to SES through access to resources and services.21,22
Socioeconomic Status Measures
The SES variables were selected a priori based on published literature and work examining racial disparities, in childhood asthma.4,23 These variables represent seven SES domains: income, health care access/coverage, education, wealth, housing, food insecurity, and government assistance (eTable 1).
Statistical Methods
Statistical analyses were performed in Stata/BE 18.0 (Stata-Corp, College Station, Tex) and figures were created using the ggplot2 package (Wickham, 2009) in R Version 4.3.0 (R Foundation, Vienna, Austria).24,25 Survey methods were used to account for the complex survey design, incorporating sampling weights, strata, and primary sampling units. Data from survey years 2011–2018 were pooled for analysis. Age was modeled as a continuous covariate in all regression models. All statistical testing were performed with an alpha level of 0.05.17
For variables with less than 5% missingness, observations with missing values for the outcomes or covariates were excluded. Variables with greater than 5% missingness were recoded to include a missing category, reflecting higher non-response rates for those variables. Details of sample derivation and handling of missing data are shown in eFigure 1.
The NHIS does not collect education information on parents not living in the household. Nearly all children with missing paternal education were coded as “not ascertained” indicating that the father did not live in the household.17 Therefore, paternal education was dichotomized as reported versus not reported. The final analysis population included Latino children aged 0 to <18 years with complete data for all variables, allowing for missing-category exceptions as described above.
Regression Modeling Strategy
Survey-weighted multivariable logistic regression models were fit separately for asthma and AD prevalence to estimate adjusted odds ratios. All models adjusted for age (continuous) and sex, given established variation in asthma and AD prevalence by these factors.
To examine acculturation as a primary exposure, we first fit models including (1) one acculturation variable at a time, followed by (2) all acculturation variables jointly, and (3) all acculturation and all SES variables together.
To examine SES as a primary exposure, we fit models including (4) one SES variable at a time, (5) all SES variables jointly, and (6) all SES and acculturation variables together. Models 3 and 6 are identical in covariate structure; however, they are presented separately because interpretation focuses on acculturation variables in Model 3, and SES variables in Model 6. To assess potential multicollinearity among socioeconomic variables included in the fully adjusted models, we examined pairwise correlations among SES variables. Correlations were modest at most and given that estimates of associations were stable, multicollinearity did not appear to be a threat to the validity of the results.
Combined Effects Prevalence Model
To examine the combined association of acculturation and SES with disease prevalence, we estimated model-based prevalence using a parsimonious combined-effects approach. Based on results from the multivariable models, acculturation was operationalized using parent interview language only. High acculturation was defined as English interview language, and low acculturation as Spanish or bilingual interview language.
Separate survey-weighted logistic regression models were fit for asthma and AD, including acculturation category, age, sex, and all SES variables shown in Tables 2 and 3. Model-estimated prevalence was calculated for combinations of acculturation and SES levels (high acculturation/high SES, high acculturation/low SES, low acculturation/high SES, and low acculturation/low SES). Estimates were generated by setting age to the sample mean (8 years), sex to male, and all other SES variables to their sample means.
Table 2.
Model-based prevalence estimates from the survey-adjusted (weighted) combined effects model results examining the association between childhood asthma prevalence and SES and acculturation (based on parent interview language) among Latino children in the NHIS (2011–2018).
| Spanish or Spanish & English % (95% CI) | English % (95% CI) | |
|---|---|---|
| Food Security | ||
| Food insecure | 7.2 (5.9, 8.5) | 12.6 (11.1, 14.2) |
| Food secure | 4.9 (4.0, 5.7) | 8.7 (7.7, 9.6) |
| Maternal Education | ||
| < High School | 4.5 (3.7, 5.3) | 8.1 (6.8, 9.4) |
| High School - Associate’s Degree | 5.9 (4.8, 7.0) | 10.5 (9.3, 11.6) |
| Bachelor’s Degree or higher | 7.3 (5.5, 9.1) | 12.7 (10.6, 14.9) |
| Father in Home | ||
| No | 6.6 (5.3, 7.9) | 11.6 (10.3, 13.0) |
| Yes | 5.1 (4.2, 6.0) | 9.0 (8.0, 10.1) |
| Housing | ||
| Rent with government assistance | 8.1 (6.0, 10.2) | 14.0 (11.3, 16.8) |
| Rent without government assistance | 5.7 (4.6, 6.7) | 10.0 (8.8, 11.2) |
| Owns or buying | 5.1 (4.1, 6.0) | 9.0 (7.9, 10.1) |
| Received Food Stamps | ||
| Yes | 6.4 (5.1, 7.7) | 11.3 (9.6, 13.0) |
| No | 5.1 (4.2, 6.0) | 9.1 (8.0, 10.1) |
Estimates are from the combined effects model; sex was set at male; age and all other SES variables set at population means.
Table 3.
Model-based prevalence estimates from the survey-adjusted (weighted) combined effects model results examining the association between childhood atopic dermatitis prevalence and SES and acculturation among Latino children in the NHIS (2011–2018).
| Spanish or Spanish/English Prevalence % (95% CI) | English Prevalence % (95% CI) | |
|---|---|---|
| Food Security | ||
| Food insecure | 9.6 (8.1, 11.1) | 14.2 (12.7, 15.7) |
| Food secure | 6.6 (5.6, 7.5) | 9.8 (8.9, 10.7) |
| Maternal Education | ||
| < High School | 5.9 (4.9, 6.9) | 8.9 (7.7, 10.2) |
| High School - Associate’s Degree | 8.1 (6.8, 9.5) | 12.1 (11.0, 13.2) |
| Bachelor’s Degree or higher | 9.5 (7.6, 11.4) | 14.0 (11.8, 16.1) |
| Father in Home | ||
| No | 8.2 (6.9, 9.5) | 12.2 (10.9, 13.4) |
| Yes | 7.1 (6.0, 8.2) | 10.6 (9.6, 11.6) |
| Housing | ||
| Rent with government assistance | 9.1 (6.9, 11.4) | 13.5 (10.7, 16.3) |
| Rent without government assistance | 7.7 (6.5, 8.8) | 11.4 (10.3, 12.6) |
| Owns or buying | 7.0 (5.8, 8.2) | 10.5 (9.4, 11.6) |
| Received Food Stamps | ||
| Yes | 8.7 (7.2, 10.1) | 12.8 (11.3, 14.3) |
| No | 6.8 (5.8, 7.9) | 10.2 (9.2, 11.3) |
| Income-to-Poverty Ratio | ||
| <1 | 6.4 (5.2, 7.6) | 9.6 (8.2, 11.0) |
| 1–1.99 | 7.5 (6.3, 8.7) | 11.1 (9.8, 12.5) |
| 2–3.99 | 8.7 (7.2, 10.4) | 12.9 (11.4, 14.5) |
| >4 | 8.7 (6.6, 10.7) | 12.8 (10.4, 15.2) |
| Received Welfare Assistance | ||
| Yes | 11.1 (7.9, 14.3) | 16.2 (12.3, 20.2) |
| No | 7.3 (6.3, 8.4) | 10.9 (10.0, 11.8) |
| Child has Primary Care Provider | ||
| No | 6.0 (4.6, 7.3) | 9.0 (7.4, 10.5) |
| Yes | 7.6 (6.5, 8.7) | 11.0 (10.0, 12.0) |
Estimates are from the combined effects model; sex was set at male; age and all other SES variables set at population means.
Results
Population Characteristics
The unweighted analysis population included 21,281 Latino children 0 to <18 years of age from 2011 to 2018, representing a weighted population of approximately 15.4 million children. The majority of the children were Mexican or Mexican American (65.9%), and a large proportion experienced socioeconomic disadvantage, with 30.0% of households reported income below the federal poverty level. Sample characteristics are presented in eTable 2 and missingness is presented in eTable 3.
Atopic Disease Prevalence
All disease estimates are generalized to the US population of Latino children using survey weights and accounting for the complex NHIS sampling design. The estimated prevalence of asthma among Latino children in the US was 8.5%, 95% CI [8.0, 9.0] and estimated prevalence of AD was 10.7%, 95% CI [10.2, 11.3] (Table 1). There was heterogeneity in asthma and AD prevalence across Latino subgroups, with higher prevalence observed among Puerto Rican and Dominican children compared with Mexican and Mexican American children.
Table 1. Survey-weighted.
AD and asthma prevalence among all Latino children in NHIS 2011–2018 by SES or acculturation indicator.
| Variable of Interest | Unweighted Sample Size | Survey-weighted AD prevalence (95% CI) | Survey-weighted asthma prevalence (95% CI) |
|---|---|---|---|
| Overall | 21,281 | 10.7% (10.2, 11.3) | 8.5% (8.0, 9.0) |
| Sex | |||
| Female | 10,381 | 10.7% (9.9, 11.6) | 7.5% (6.9, 8.2) |
| Male | 10,900 | 10.7% (10.0, 11.5) | 9.4% (8.7, 10.2) |
| Latino Subgroup | |||
| Multiple Hispanic | 1,007 | 13.4% (10.9, 16.3) | 11.1% (8.7, 14.1) |
| Puerto Rican | 1,891 | 14.9% (12.9, 17.1) | 17.1% (15.0, 19.5) |
| Mexican | 5,052 | 8.1% (7.1, 9.1) | 6.7% (5.8, 7.7) |
| Mexican-American | 8,715 | 10.8% (10.0, 11.8) | 7.7% (7.0, 8.4) |
| Cuban/Cuban American | 592 | 7.4% (5.1, 10.7) | 9.1% (5.8, 13.9) |
| Dominican (Republic) | 640 | 15.1% (11.8, 19.1) | 9.1% (6.5, 12.5) |
| Central or South America | 2,865 | 10.6% (9.1, 12.5) | 6.8% (5.7, 8.1) |
| Other Latin American | 98 | 17.8% (9.6, 23.1) | 6.3% (2.9, 13.5) |
| Other Spanish | 421 | 12.5% (9.3, 16.7) | 11.4% (7.8, 16.4) |
| Interview Language (parent) | |||
| Spanish or [Spanish & English] | 5,772 | 6.8% (6.0, 7.8) | 5.5% (4.8, 6.3) |
| English | 15,509 | 12.1% (11.5, 12.8) | 9.6% (8.9, 10.2) |
| Born in USA (child) | |||
| No | 1,475 | 5.2% (4.0, 6.8) | 5.5% (4.3, 7.2) |
| Yes | 19,806 | 11.1% (10.5, 11.7) | 8.7% (8.2, 9.2) |
| Citizen of US (child) | |||
| No | 1,014 | 3.9% (2.7, 5.7) | 4.3% (3.0, 6.3) |
| Yes | 20,267 | 11.0% (10.5, 11.6) | 8.7% (8.1, 9.2) |
| Income/Poverty Ratio | |||
| <1 | 5,989 | 10.0% (9.0, 11.0) | 9.7% (8.8, 10.8) |
| 1–2 | 6,325 | 10.4% (9.3, 11.5) | 8.2% (7.3, 9.2) |
| 2–4 | 4,911 | 12.0% (10.9, 13.3) | 7.7% (6.8, 8.7) |
| >4 | 2,606 | 12.7% (11.2, 14.3) | 7.4% (6.2, 8.7) |
| Undefinable or unknown | 1,450 | 8.2% (6.3, 10.6) | 9.0% (7.2, 11.1) |
| Food Insecurity | |||
| Yes (food insecure) | 6,940 | 12.7% (11.6, 13.8) | 11.1% (10.1, 12.2) |
| No (food secure) | 14,341 | 9.8% (9.2, 10.5) | 7.3% (6.7, 7.9) |
| Interest | |||
| No | 19,643 | 10.5% (10.0, 11.1) | 8.5% (8.0, 9.1) |
| Yes | 1,638 | 13.4% (11.5, 15.6) | 8.2% (6.6, 10.1) |
| Dividends | |||
| No | 20,455 | 10.6% (10.0, 11.2) | 8.5% (8.0, 9.1) |
| Yes | 826 | 14.4% (11.4, 18.0) | 7.6% (5.7, 10.0) |
| Maternal Education Level | |||
| <High School | 6,333 | 7.6% (6.8, 8.5) | 7.0% (6.3, 7.9) |
| High School thru associate’s, technical, or some college | 10,508 | 12.1% (11.3, 13.0) | 9.3% (8.5, 10.1) |
| Bachelor’s or higher | 2,918 | 13.6% (12.1, 15.3) | 9.1% (7.8, 10.5) |
| Refused, not ascertained, don’t know | 1,522 | 8.8% (6.4, 11.9) | 8.1% (6.5, 10.2) |
| Paternal Education Level | |||
| Not reported | 7,023 | 12.1% (11.1, 13.2) | 11.6% (10.6, 12.7) |
| Reported | 14,258 | 10.1% (9.5, 10.8) | 7.1% (6.5, 7.7) |
| Housing | |||
| Rents with government assistance | 1,282 | 14.3% (12.0, 17.0) | 15.9% (13.4, 18.7) |
| Rents without government assistance | 10,282 | 10.6% (9.7, 11.5) | 8.3% (7.7, 9.1) |
| Owns or buying | 9,354 | 10.5% (9.7, 11.3) | 7.5% (6.8, 8.3) |
| Other arrangement | 363 | 9.3% (6.2, 13.7) | 12.1% (8.7, 16.6) |
| Cash Welfare Benefits | |||
| Yes | 1,241 | 12.0% (9.7, 14.7) | 12.1% (10.0, 14.7) |
| No | 20,040 | 10.7% (10.1, 11.3) | 8.3% (7.7, 8.8) |
| Other Welfare Benefits | |||
| Yes | 650 | 18.9% (15.1, 23.5) | 14.2% (10.9, 18.3) |
| No | 20,621 | 10.5% (9.9, 11.1) | 8.3% (7.8, 8.9) |
| Food Stamps | |||
| Yes | 6,836 | 11.9% (10.9, 13.0) | 10.7% (9.6, 11.8) |
| No | 14,445 | 10.1% (9.4, 10.8) | 7.3% (6.8, 7.9) |
| Commercial Health Insurance | |||
| No | 14,098 | 10.2% (9.5, 10.9) | 8.9% (8.3, 9.5) |
| Yes | 7,183 | 11.9% (11.0, 12.8) | 7.7 (6.9, 8.6) |
| Has Primary Care Provider | |||
| No | 2,190 | 7.9% (6.6, 9.4) | 7.7% (6.0, 9.8) |
| Yes | 19,091 | 11.0% (10.5, 11.7) | 8.6% (8.0, 9.1) |
Asthma
In adjusted analyses, proxies for greater acculturation (English parent interview language, child US-born, and child US citizen) were positively associated with asthma prevalence when examined individually. When all acculturation variables were included jointly, only parent interview language remained statistically significant (eTable 4). Further adjustment for SES did not meaningfully change the magnitude or direction of the association between asthma and interview language (eTable 4, Figure 1).
Figure 1.
Odds ratios for asthma diagnosis predicted by acculturation after accounting for SES (Model 3 results). Diamonds represent odds ratios (OR), lines with bars represent the 95% confidence intervals (CI). Lines that cross 1.0 (the red dotted line) indicate statistically insignificant OR.
Associations between SES and asthma prevalence were more complex than for acculturation. In models that included individual SES indicators, multiple markers of socioeconomic disadvantage, including lower income, rental housing, receipt of government assistance, food insecurity, father not in the household, and public health insurance, were significantly associated with greater asthma prevalence (eTable 5). Other markers of poverty (no receipt of investment interest or dividends) had associations in the same direction, although some estimates did not reach statistical significance. In contrast, some indicators typically associated with lower socioeconomic disadvantage (greater maternal education level and the child having a primary care provider [PCP]) were associated with greater asthma prevalence. When all SES variables were included jointly, and when all acculturation measures were added to these models, individual estimates were attenuated or strengthened; however, the overall pattern of associations between SES and asthma prevalence remained consistent (eTable 5, Figure 2).
Figure 2.
Odds ratios for asthma diagnosis by SES after accounting for acculturation (Model 6 results). Diamonds represent odds ratios (OR), lines with bars represent the 95% confidence intervals (CI). Lines that cross 1.0 (the red dotted line) indicate statistically insignificant OR.
Abbreviations: Ed = education; HS = high school; Assoc = associates degree; Govt = government; PCP = primary care provider.
Results from the combined effects prevalence model showed that across all examined combinations, children in the high acculturation/low SES groups had the highest asthma prevalence and children in the low acculturation/high SES groups had the lowest asthma prevalence (Table 2). For example, among 8-year-old boys with Spanish-speaking parents and no food insecurity, the estimated asthma prevalence was 5.0% (95% CI 4.0, 6.0). In contrast, among 8-year-old boys with English-speaking parents and food insecurity the estimated asthma prevalence was 13.0% (95% CI 11.0, 14.0), holding other SES characteristics constant.
Atopic Dermatitis
Similar to asthma, adjusted results showed that greater acculturation was positively associated with AD prevalence for all three indicators when considered individually and jointly; however, only parent interview language remained statistically significant when accounting for all acculturation variables in the same model (eTable 6, Figure 3).
Figure 3.
Odds of AD diagnosis predicted by acculturation after accounting for SES (Model 3 results). Diamonds represent odds ratios (OR), lines with bars represent the 95 confidence intervals (CI). Lines that cross 1.0 (the red dotted line) indicate statistically insignificant OR.
Several markers of lower SES, such as father not in the household, food insecurity, and use of government assistance (including food stamps, rental assistance, or non-cash welfare assistance), were positively associated with AD prevalence (eTable 7). Receipt of government assistance, particularly non-cash welfare assistance, was also positively associated with AD prevalence. Conversely, indicators typically reflecting higher SES, including higher maternal educational attainment, child having private insurance, child having a PCP, home ownership, and markers of greater wealth (receipt of investment interest or dividends), were also positively associated with AD prevalence. Accounting for all SES variables in a single model, and subsequently accounting for all acculturation variables, attenuated or strengthened some associations; however, the overall pattern of results remained similar (eTable 7;Figure 4).
Figure 4.
Odds of AD diagnosis predicted by SES after accounting for acculturation Model 6 results). Diamonds represent odds ratios (OR), lines with bars represent the 95% confidence intervals (CI). Lines that cross 1.0 (the red dotted line) indicate statistically insignificant OR.
Results from the combined effects prevalence model showed that SES variables like receipt of government assistance and experiencing food insecurity that are better at distinguishing very poor from not very poor individuals had associations with disease prevalence similar to asthma; children in the high acculturation/low SES groups generally had the highest AD prevalence and children in the low acculturation/high SES groups generally had the lowest AD prevalence (Table 3). For example, among 8-year-old boys with Spanish-speaking parents and no food insecurity, the estimated AD prevalence was 7% (95% CI 6, 8), compared with 14% (95% CI 13, 16) among 8-year-old boys with English-speaking parents and food insecurity.
The associations were less straightforward among SES variables that may be better at distinguishing wealthy from not wealthy, and the difference in prevalence was minimal for some variables. For example, among 8-year-old boys with Spanish-speaking parents and an income-to-poverty ratio greater than 4, estimated AD prevalence was 9% (95% CI 7, 11), compared with 10% (95% CI, 8–11) among those with English-speaking parents and an income-to-poverty ratio less than 1, holding all other SES characteristics constant.
Sensitivity Analyses
In a sensitivity analysis restricted to children with evidence of healthcare access (children with commercial health insurance and a usual source of care, N = 6,696), associations between SES and acculturation-related factors and AD were largely consistent with those observed in the primary analysis (eTable 8).
When the sample was restricted to Mexican and Mexican American children, the largest Latino subgroup, associations between SES and acculturation-related factors and asthma and AD were similar to those in the primary analysis (eTable 9).
Given that odds ratios can overstate associations when outcomes are not rare, we also conducted a sensitivity analysis refitting the fully adjusted primary models for asthma and AD using survey-weighted Poisson regression with robust standard errors to estimate prevalence ratios. Findings from this sensitivity analysis were directionally consistent with the primary odds ratio–based findings (eTable 10).
Discussion
In a nationally representative population of Latino children in the US, markers of greater acculturation and multiple markers of lower SES were independently and positively associated with current asthma and AD prevalence. Furthermore, Latino children with the combination of greater acculturation and lower SES had the highest prevalences of asthma and AD, which generally were twice as high as those with less acculturation and higher SES. Although SES and acculturation have been recognized individually as risk factors for asthma and AD prevalence, estimates of asthma and AD prevalence among Latino children vary widely.4,6,27,28 While this variability appears to be explained, in part, by heterogeneity across Latino ethnicity subgroups, it may also be explained by heterogeneity in SES and acculturation in Latino children in the US.3,7,10,12 We found that US Latino children experiencing the combination of greater acculturation and several markers of greater poverty had an approximately 2-fold higher prevalence of current asthma or AD than those with lower acculturation and without poverty.10 Notably, the asthma and AD prevalence in this group of Latino children with greater acculturation and socioeconomic disadvantage was similar to the prevalence reported among Black children in the US, who have among the highest asthma prevalence.4,23,29 Although most Latino children (aside from Puerto Rican and Dominican) have generally been considered to be at substantially lower risk for asthma and AD, these findings highlight a high burden among US Latino children experiencing socioeconomic disadvantage, particularly when combined with higher acculturation as proxied by parent interview language.27,28,30 These findings are particularly salient given the US Latino population’s rapid growth and shift toward a lower proportion of immigrants, with those who have immigrated having spent more time in the US.10
Several of the associations between SES and asthma and/or AD prevalence are well-established in the literature, while others have not been or are rarely discussed. Several domains of lower SES were significantly associated with greater asthma prevalence: lower income, rental housing, receipt of government assistance, food insecurity, father not in the household, and the child having public health insurance. The results for SES and AD were mostly in line with those for asthma, but not completely. Several markers of greater poverty (use of government assistance [food stamps, rental assistance, or non-cash welfare assistance], experiencing food insecurity, father not in the household) were positively associated with AD prevalence. Our hypothesis, supported by the literature, is that the absence of a father in the home (represented by paternal education not reported) reflects broader household vulnerability and is a marker for other risk factors such as material hardship and maternal stress.31,32 Interestingly, higher income, higher maternal education level, child having private insurance, home ownership, and markers of greater wealth (receipt of investment interest or dividends) were also positively associated with AD prevalence. Other researchers have reported similar mixed associations between SES and AD prevalence, and one explanation is that markers such as higher maternal education level and wealth may identify children who are more likely to receive a diagnosis.33 Accordingly, these measures of wealth may not reflect greater underlying disease burden but may reflect differential diagnosis opportunity through increased access to quality health care.34,35 To address concerns that differential diagnostic opportunity may explain positive associations between higher SES and AD, we conducted a sensitivity analysis restricted to children with evidence of healthcare access (commercial insurance and a usual source of care). Associations in this restricted sample were largely consistent with the primary analysis, suggesting that differential diagnostic opportunity alone does not fully explain the observed associations. These findings expand upon published literature implicating lower SES with greater odds of asthma or AD prevalence by utilizing a multidimensional measure of SES. An important contribution of this study is the use of a multidomain SES construct that captures multiple dimensions of socioeconomic context (income, education, wealth, housing, food insecurity, government assistance, and healthcare access) rather than relying on a single SES indicator. This approach provides a more comprehensive view of how different socioeconomic domains may relate to asthma and AD prevalence, and it helps contextualize prior work that has often examined one SES measure at a time.
Even after accounting for SES, greater acculturation (parents completed interview in English) was associated with a nearly 2-fold increase in asthma prevalence and a nearly 1.5-fold increase in AD prevalence. Some studies have reported greater asthma prevalence and/or disease burden among children whose primary household language was English (vs. Spanish), but the studies were not nationally representative, reported smaller magnitude differences, and/or were among children who were at high risk of asthma instead of a general population.36–39 Other literature ties birthplace outside the US and less time residing in the US with lower AD prevalence among ethnically and racially diverse populations; however, to our knowledge there is limited nationally representative evidence examining parent interview language as an acculturation proxy in relation to AD prevalence, specifically among Latino children.3,6
We do not interpret parent interview language as a causal exposure; rather, it is a marker of acculturation and related social and contextual factors that may correlate with asthma and AD prevalence. Despite high rates of poverty and less access to quality health care, which are associated with increased atopic disease risk among the general population, more recent immigrants tend to have lower asthma and AD prevalence.6,7,40 This pattern is in line with the Hispanic (or Immigrant) Paradox, with evidence that this advantage attenuates with longer time in the US.41 Proposed explanations include differences in environmental exposures, diet, and other lifestyle factors associated with the acculturation process.42,43 Recent immigrants with lower acculturation may also be healthier than those they left behind in their country of origin (healthy migrant effect) and thus lower SES may be less of a risk factor for diseases among this population.44 Acculturation is also associated with marginalization experiences such as discrimination, xenophobia, segregation, and immigrant stress, which may contribute to psychosocial stress.45 Psychosocial stress has been tied to worse asthma and AD outcomes and may be one pathway linking acculturation-related experiences with greater disease burden.46,47
We utilized surrogate variables in lieu of a validated acculturation instrument, but use of these variables is supported in the literature as proxies for acculturation in population-based studies. NHIS public-use files code Spanish-only interviews and bilingual (Spanish and English) interviews together, precluding separate evaluation of these language contexts. Citizenship was treated as an additional proxy for acculturation but with the acknowledgement that it is also tied to SES because citizenship confers access to resources. Importantly, when modeled together not all acculturation variables retained associations with AD and asthma prevalence; associations with language were retained but associations with nativity and citizenship waned. The limited variability in nativity and citizenship in our sample (approximately 7.0% were foreign-born and 5.0% were non-citizens), may have contributed to less precise estimates of association for these measures. These observations are consistent with the NHIS study design, which represents the civilian, non-institutionalized US population and may underrepresent non-citizen or more recently immigrated children. The observed associations may not be generalized to Latino subgroups, other than Mexican and Mexican Americans, for whom results were consistent with the larger Latino population.
Several additional limitations should be considered. The NHIS is cross-sectional, with exposures and outcomes measured at the same time; therefore, temporality cannot be established. Accordingly, the associations observed in this study should not be interpreted as definitive evidence that acculturation or socioeconomic conditions cause asthma or AD, nor as evidence of specific causal pathways. Notably, the NHIS does not include an objective clinical assessment of asthma and AD. Asthma in NHIS is based on parent report of whether the child ever had asthma and still has asthma, while AD is captured from parent report of eczema or any kind of skin allergy in the past 12 months.17 Our estimates of prevalence are more likely to reflect current prevalence rather than lifetime prevalence of asthma and AD. Additionally, reliance on parent-reported clinician diagnosis without severity information limits our ability to comment on gradients in disease severity. We could not account for the number of years a child had lived in the US because the NHIS did not collect this as a continuous measure. Thus, future studies could examine duration of US residence. Many SES domains were included, but it is not possible to capture all aspects of SES. The study also did not incorporate additional exposures such as diet, secondhand smoke, allergens, outdoor air pollution, neighborhood safety, housing quality, or discrimination-related stressors, nor directly evaluate mechanisms linking acculturation and SES to asthma and AD prevalence. Examining multiple socioeconomic indicators across two outcomes also increases the potential for type I error due to multiple testing; individual associations should therefore be interpreted cautiously. Finally, we included NHIS data through 2018 because the NHIS questionnaire and weighting were redesigned in 2019, precluding including more recent years.
The shifting demographics of the US Latino population reflect increasing acculturation in a population that also experiences high rates of poverty, and both lower SES and greater acculturation were associated with asthma and AD prevalence in this analysis.3,10,13 Given these shifting sociodemographic characteristics, asthma and AD prevalence in this population may change over time. The finding that the combination of lower SES and greater acculturation was strongly associated with disease prevalence highlights the importance of accounting for both factors when examining asthma and AD prevalence in the US Latino population. These findings also have implications for clinical screening and public health resource allocation. Latino children experiencing the intersection of greater acculturation and socioeconomic disadvantage may warrant targeted screening for asthma and atopic dermatitis and prioritization for prevention and management resources.
Supplementary Material
Funding declaration:
Funding is provided by the National Institutes of Health (grants T32HL140290 [Croce & Matsui], R01DK118354 [Parast], KL2 TR002646 [Bhavnani], K24AI114769 [Matsui], R01ES035131 [Matsui], R01ES034803 [Matsui], R01ES023447 [Matsui], R01ES026170 [Matsui], and R34HL159126 [Matsui]), the Childhood Eczema Challenge Grant jointly sponsored by the Pediatric Dermatology Research Alliance (PeDRA) Research and the National Eczema Association (NEA) [Croce], and core funds of the Dell Medical School at the University of Texas at Austin (Bhavnani). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Abbreviations:
- AD
atopic dermatitis
- SES
socioeconomic status
- US
United States
- PCP
primary care provider
- PSU
primary sampling unit
- CI
confidence interval
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest:
The authors have no relevant conflicts of interest to report.
Contributor Information
Emily Croce, The University of Texas at Austin Dell Medical School, Austin, TX 78712; The University of Texas at Austin Steve Hicks School of Social Work, Austin, TX 78712; Dell Children’s Medical Group, Austin, TX 78723.
Darlene Bhavnani, The University of Texas at Austin Dell Medical School, Austin, TX 78712.
Layla Parast, The University of Texas at Austin Department of Statistics and Data Sciences, Austin, TX 78705.
Khamron Micheals, The University of Texas at Austin Dell Medical School, Austin, TX 78712; The University of Texas at Austin Steve Hicks School of Social Work, Austin, TX 78712.
María Teresa García Romero, Dermatology Department, Instituto Nacional de Pediatría, Mexico City, Mexico 04530.
Gabriela Livas, The University of Texas at Austin Department of Human Development and Family Sciences, Austin, TX 78712.
Elizabeth C. Matsui, The University of Texas at Austin Dell Medical School, Austin, TX 78712; Dell Children’s Medical Group, Austin, TX 78723.
References
- 1.Choragudi S, Yosipovitch G. Trends in the Prevalence of Eczema Among US Children by Age, Sex, Race, and Ethnicity From 1997 to 2018. JAMA Dermatol. 2023;159(4):454–456. doi: 10.1001/jamadermatol.2022.6647 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cardet JC, Chang KL, Rooks BJ, Carroll JK, Celedon JC, Coyne-Beasley T et al. Socioeconomic status associates with worse asthma morbidity among Black and Latinx adults. J Allergy Clin Immunol. 2022;150(4):841–849.e4. doi: 10.1016/j.jaci.2022.04.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gold DR, Acevedo-Garcia D. Immigration to the United States and acculturation as risk factors for asthma and allergy. J Allergy Clin Immunol. 2005;116(1):38–41. doi: 10.1016/j.jaci.2005.04.033 [DOI] [PubMed] [Google Scholar]
- 4.Croce EA, Levy ML, Adamson AS, Matsui EC. Reframing racial and ethnic disparities in atopic dermatitis in Black and Latinx populations. J Allergy Clin Immunol. 2021;148(5):1104–1111. doi: 10.1016/j.jaci.2021.09.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Eldeirawi KM and Persky VW. Associations of Acculturation and Country of Birth with Asthma and Wheezing in Mexican American Youths. J Asthma. 2006; 43(4):279–286. doi: 10.1080/0277090060022869 [DOI] [PubMed] [Google Scholar]
- 6.Silverberg JI, Simpson EL, Durkin HG, Joks R. Prevalence of Allergic Disease in Foreign-Born American Children. JAMA Pediatr. 2013;167(6):554–560. doi: 10.1001/jamapediatrics.2013.1319 [DOI] [PubMed] [Google Scholar]
- 7.Iqbal S, Oraka E, Chew GL, Flanders WD. Association Between Birthplace and Current Asthma: The Role of Environment and Acculturation. Am J Public Health. 2014;104(S1):S175–S182. doi: 10.2105/AJPH.2013.301509 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Martinez A, delaRosa R, Mujahid M, Thakur N. Structural Racism and its Pathways to Asthma and Atopic Dermatitis. J Allergy Clin Immunol. 2021;148(5):1112–1120. doi: 10.1016/j.jaci.2021.09.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.US Census Bureau. New Estimates Highlight Differences in Growth Between the U.S. Hispanic and Non-Hispanic Populations. Census.gov. Accessed September 24, 2024. https://www.census.gov/newsroom/press-releases/2024/population-estimates-characteristics.html
- 10.Moslimani M and Noe-Bustamante L. Facts on Latinos in the U.S. Pew Research Center. August 16, 2023. Accessed April 28, 2025. https://www.pewresearch.org/race-and-ethnicity/fact-sheet/latinos-in-the-us-fact-sheet/
- 11.Abraído-Lanza AF, Armbrister AN, Flórez KR, Aguirre AN. Toward a theory-driven model of acculturation in public health research. Am J Public Health. 2006;96(8):1342–1346. doi: 10.2105/AJPH.2005.064980 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Garcia DS. Acculturation: A Core Concept of Hispanic Health Research. Hisp Health Care Int Off J Natl Assoc Hisp Nurses. 2015;13(3):115–118. doi: 10.1891/1540-4153.13.3.115 [DOI] [PubMed] [Google Scholar]
- 13.Thiede BC, Brooks MM, Jensen L. Unequal From the Start? Poverty Across Immigrant Generations of Hispanic Children. Demography. 2021;58(6):2139–2167. doi: 10.1215/00703370-9519043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ablow Measelle E, McClure HH, Snodgrass JJ, Martinez CR Jr, Jiménez R, Isiordia L. Climbing the ladder of decline: Income and acculturation associated with chronic inflammation among Mexican immigrants. Am J Hum Biol. 2019;31(4):e23271. doi: 10.1002/ajhb.23271 [DOI] [PubMed] [Google Scholar]
- 15.Mendoza FS. Health Disparities and Children in Immigrant Families: A Research Agenda. Pediatrics. 2009;124(Supplement_3):S187–S195. doi: 10.1542/peds.2009-1100F [DOI] [PubMed] [Google Scholar]
- 16.Montoya-Williams D, Williamson VG, Cardel M, Fuentes-Afflick E, Maldonado-Molina M, Thompson L. The Hispanic/Latinx Perinatal Paradox in the United States: A Scoping Review and Recommendations to Guide Future Research. J Immigr Minor Health. 2021;23(5):1078–1091. doi: 10.1007/s10903-020-01117-z [DOI] [PubMed] [Google Scholar]
- 17.NHIS - Methods. June 14, 2021. Accessed June 17, 2023. https://www.cdc.gov/nchs/nhis/methods.htm
- 18.Passel MHL Jens Manuel Krogstad and Jeffrey S. Who is Hispanic? Pew Research Center. September 12, 2024. Accessed April 28, 2025. https://www.pewresearch.org/short-reads/2024/09/12/who-is-hispanic/
- 19.Gee GC, Morey BN, Walsemann KM, Ro A, Takeuchi DT. Citizenship as Privilege and Social Identity: Implications for Psychological Distress. Am Behav Sci. 2016;60(5–6):680–704. doi: 10.1177/0002764216632834 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Arcia E, Skinner M, Bailey D, Correa V. Models of acculturation and health behaviors among Latino immigrants to the US. Soc Sci Med 1982. 2001;53(1):41–53. doi: 10.1016/s0277-9536(00)00310-5 [DOI] [PubMed] [Google Scholar]
- 21.Philipsborn RP, Sorscher EA, Sexson W, Evans HH. Born on U.S. Soil: Access to Healthcare for Neonates of Non-Citizens. Matern Child Health J. 2021;25(1):9–14. doi: 10.1007/s10995-020-03020-3 [DOI] [PubMed] [Google Scholar]
- 22.Chinchilla M, Yue D, Ponce NA. Housing Insecurity Among Latinxs. J Immigr Minor Health. 2022;24(3):656–665. doi: 10.1007/s10903-021-01258-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Croce EA, Parast L, Bhavnani D, Matsui EC. Lower socioeconomic status may help explain racial disparities in asthma and atopic dermatitis prevalence: A mediation analysis. J Allergy Clin Immunol. 2024;153(4):1140–1147.e3. doi: 10.1016/j.jaci.2023.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Stata Statistical Software. Published online 2025.
- 25.RStudio Team. RStudio. Published online 2024.
- 26.FAQ: Missing standard error because of stratum with single sampling unit. | Stata. Accessed June 17, 2023. https://www.stata.com/support/faqs/statistics/standard-error-because-of-stratum/
- 27.Brunner PM, Guttman-Yassky E. Racial differences in atopic dermatitis. Ann Allergy Asthma Immunol Off Publ Am Coll Allergy Asthma Immunol. 2019;122(5):449–455. doi: 10.1016/j.anai.2018.11.015 [DOI] [PubMed] [Google Scholar]
- 28.Choragudi S, Yosipovitch G. Trends in the Prevalence of Eczema Among US Children by Age, Sex, Race, and Ethnicity From 1997 to 2018. JAMA Dermatol. 2023;159(4):454–456. doi: 10.1001/jamadermatol.2022.6647 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Martinez A, de la Rosa R, Mujahid M, Thakur N. Structural racism and its pathways to asthma and atopic dermatitis. J Allergy Clin Immunol. 2021;148(5):1112–1120. doi: 10.1016/j.jaci.2021.09.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Salari K, Burchard EG. Latino populations: a unique opportunity for epidemiological research of asthma. Paediatr Perinat Epidemiol. 2007;21 Suppl 3:15–22. doi: 10.1111/j.1365-3016.2007.00880.x [DOI] [PubMed] [Google Scholar]
- 31.Livingston G. A Tale of Two Fathers. Pew Research Center’s Social & Demographic Trends Project. June 15, 2011. Accessed June 20, 2023. https://www.pewresearch.org/social-trends/2011/06/15/a-tale-of-two-fathers/ [Google Scholar]
- 32.Livingston G, Parker K. Chapter 1. Living Arrangements and Father Involvement. Pew Research Center’s Social & Demographic Trends Project. June 15, 2011. Accessed September 28, 2023. https://www.pewresearch.org/social-trends/2011/06/15/chapter-1-living-arrangements-and-father-involvement/ [Google Scholar]
- 33.Chung J and Simpson EL. The socioeconomics of atopic dermatitis. Ann Allergy Asthma Immunol. 2019;122(4):360–366. doi: 10.1016/j.anai.2018.12.017 [DOI] [PubMed] [Google Scholar]
- 34.Siegfried EC, Paller AS, Mina-Osorio P, Vekeman F, Kaur M, Mallya UG, et al. Effects of variations in access to care for children with atopic dermatitis. BMC Dermatol. 2020;20(1):24. doi: 10.1186/s12895-020-00114-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Servattalab SE, Lee M, Hlobik M, Song H, Huang JT. Examining racial and ethnic disparities in diagnosis and access to care in infantile atopic dermatitis in the USA: a retrospective cohort study. Br J Dermatol. 2024;192(1):144–151. doi: 10.1093/bjd/ljae322 [DOI] [PubMed] [Google Scholar]
- 36.Emmanuel M, Margolis R, Badh R, Kachroo N, Teach SJ, Parikh K. Caregiver Language Preference and Health Care Utilization Among Children With Asthma. Pediatrics. 2023;152(6):e2023061869. doi: 10.1542/peds.2023-061869 [DOI] [PubMed] [Google Scholar]
- 37.Mosnaim GS, Sadowski LS, Durazo-Arvizu RA, Sharp LK, Curtis LM, Shalowitz MU, et al. Parental language and asthma among urban Hispanic children. J Allergy Clin Immunol. 2007;120(5):1160–1165. doi: 10.1016/j.jaci.2007.08.040 [DOI] [PubMed] [Google Scholar]
- 38.Banta JE, Ramadan M, Alhusseini N, Aloraini K, Modeste N. Socio-demographics and asthma prevalence, management, and outcomes among children 1–11 years of age in California. Glob Health Res Policy. 2021;6(1):17. doi: 10.1186/s41256-021-00199-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Thakur N, Borrell LN, Ye M, Oh SS, Eng C, Meade K, et al. Acculturation is associated with asthma burden and pulmonary function in Latino youth: The GALA II study. J Allergy Clin Immunol. 2019;143(5):1914–1922. doi: 10.1016/j.jaci.2018.12.1015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Siañez M, Highfield L, Balcazar H, Collins T, Grineski S. An Examination of the Association of Multiple Acculturation Measures with Asthma Status Among Elementary School Students in El Paso, Texas. J Immigr Minor Health. 2018;20(4):884–893. doi: 10.1007/s10903-017-0627-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis. 2001;11(3):496–518. [PubMed] [Google Scholar]
- 42.Tham EH, Loo EXL, Zhu Y, Shek LPC. Effects of Migration on Allergic Diseases. Int Arch Allergy Immunol. 2019;178(2):128–140. doi: 10.1159/000494129 [DOI] [PubMed] [Google Scholar]
- 43.Wong QYA, Chew FT. The Association between Migration and Prevalence of Allergic Diseases: A Systematic Review and Meta-Analysis. Int Arch Allergy Immunol. 2024;185(11):1099–1122. doi: 10.1159/000539382 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ruhnke SA, Reynolds MM, Wilson FA, Stimpson JP. A healthy migrant effect? Estimating health outcomes of the undocumented immigrant population in the United States using machine learning. Soc Sci Med 1982. 2022;307:115177. doi: 10.1016/j.socscimed.2022.115177 [DOI] [PubMed] [Google Scholar]
- 45.García Coll C, Lamberty G, Jenkins R, McAdoo HP, Crnic K, Wasik BH, et al. An integrative model for the study of developmental competencies in minority children. Child Dev. 1996;67(5):1891–1914. [PubMed] [Google Scholar]
- 46.Arndt J, Smith N, Tausk F. Stress and Atopic Dermatitis. [DOI] [PubMed]
- 47.Chen E, Hanson MD, Paterson LQ, Griffin MJ, Walker HA, Miller GE. Socioeconomic status and inflammatory processes in childhood asthma: The role of psychological stress. J Allergy Clin Immunol. 2006;117(5):1014–1020. doi: 10.1016/j.jaci.2006.01.036 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.