Abstract
Purpose:
This study aimed to examine neighborhood effects on the physical and socioemotional health of children from immigrant families, after controlling for parents’ demographic characteristics, socioeconomic status, acculturation, and health care issues.
Design:
Pooled cross-sectional data were merged with community profiles.
Setting:
The United States in 2013, 2014, and 2015.
Participants:
10,399 children from immigrant families in the 2013-2015 National Health Interview Surveys and the U.S. Census Data.
Measures:
Both objective and subjective measures of neighborhood environments were assessed, including neighborhood physical disorder, socioeconomic status, demographic composition, community resources, and social trust.
Analysis:
Descriptive statistics, logistic regression models.
Results:
About half of the sampled children were male (51%); 68% were white; 56% were of Hispanic; and 34% were school-aged. Three neighborhood factors—neighborhood trust, area-level poverty rate, and the presence of primary care physician—were identified as significant predictors for child health outcomes. Foreign-born population, green space, and food desert were not significant. At the individual level, parents’ racial and ethnic minority status, non-marital status, and healthcare issues were found to be risk factors. Families’ financial resources and parental education were identified as protective factors of socioemotional health.
Conclusion:
Intervention approaches to build on neighborhood trust may have broad potential to improve child outcomes. Programs focusing on immigrant families with children in high poverty neighborhoods should be a high priority.
Purpose
Recognizing that children’s health outcomes are influenced by the environmental contexts to which they belong as well as by individual caregivers’ and family characteristics, researchers have examined multiple dimensions of neighborhood environment with a focus on physical characteristics, socioeconomic status, demographic composition, community resources, and collective efficacy. 1 Evidence indicates that disadvantaged neighborhoods adversely impact a wide range of health outcomes including general health status, 2 health behaviors, 3 socioemotional development, 4 mental health, 5 and long-term outcomes of children. 6 The majority of prior studies heavily relied on neighborhood built environment as a predictor of children’s health, while others have focused on perceived measures of neighborhood conditions. 7 Both objective and subjective measures of neighborhood context demonstrate the important role of this environment on child health and development. Despite this evidence, there are some limitations to existing research, including threats to external validity due to using unrepresentative samples and limited measures of neighborhood conditions. Furthermore, although minority children are disproportionally represented in disadvantaged neighborhoods, research examining the effects of neighborhood on child health outcomes in immigrant families remain sparse. 8 Merging nationally sampled individual data and community profiles, the proposed study addresses this gap by testing whether neighborhood characteristics are related to physical and socioemotional health of children in immigrant families after controlling for the effects of parents’ demographic characteristics, socioeconomic status, psychological health, acculturation, and health care accessibility and affordability.
Neighborhood Factors to Determine Child Health
The association between neighborhood conditions and children’s physical and socioemotional health is well-documented. Beneficial neighborhood characteristics are positively associated with child health status, with children in safer neighborhoods engaging in more physical activities. 7 Conversely, adverse neighborhood characteristics were found to be a risk factor for children’s cognitive and socioemotional development. 4 The percentage of individuals who were foreign-born in a neighborhood was a significant predictor of children’s risk behaviors, 9 while the availability of a healthcare provider 10 as well as the presence of public resources and services 7 were identified as protective factors. In addition to neighborhood economic disadvantage and demographic composition, perceived neighborhood social cohesion, social control, and safety are associated with children’s general health status, socioemotional health, and cognitive development. 4,7 Although the aforementioned studies find neighborhood contexts to be predictors of child health, they rely on only a few indicators rather than a comprehensive set of measures, and thus only partially capture neighborhood conditions. To the best of our knowledge, there has not been any attempt to assess multiple dimensions of neighborhood conditions in a single model.
Individual and Family Factors to Determine Child Health
Research investigating predictors of child health has included individual and family characteristics as covariates. Children’s health disparities were found to reflect gender and racial differences, indicating that boys and white children tend to demonstrate more behavior problems than girls and other racial groups, respectively. 11 Parents’ marital status was significantly associated with children’s socioemotional development, with children from two-parent families less likely to have behavior problems. 12 Parents’ health status was identified as 1 of the most robust factors to determine their children’s health outcomes. 13 Evidence indicates that higher socioeconomic status, 14 higher levels of psychological health with less stress, depression, and anxiety, 15 more accessible and affordable health care services, 16 and more access to green space 17 predict better health outcomes of children. Children in rural areas were found to have poorer health outcomes compared to those in urban areas. 18 Both parents’ and their children’s immigration status were also related to child health. 19 Given that children from immigrant families are among those at greatest risk for poor health outcomes, researchers have accounted for acculturation through factors such as legal immigration status, 20 years in the U.S., 21 literacy and language proficiency, 10 and cultural barriers. 22 Accessibility and affordability of health care resources were also found to be significantly related to health outcomes of children in immigrant families. 22,23
Conceptual Model
Informed by the ecological perspective that multiple dimensions of children’s environments influence their health and development, 24 we developed a conceptual model to investigate individual-, family-, and neighborhood-level factors associated with children’s physical and socioemotional health. The present study aims to address the research gaps discussed earlier and extend the current knowledge by: (a) incorporating both neighborhood built environment and perception measures, (b) focusing on disproportionately vulnerable children from immigrant families, and (c) using nationally representative samples merged with neighborhood profiles. This study examines whether neighborhood conditions are related to physical and socioemotional health of children from immigrant families after controlling for the effects of parents’ demographic characteristics, socioeconomic status, and health care accessibility and affordability.
Methods
Study Design and Sample
We used public- and restricted-use data from the 2013-2015 National Health Interview Survey (NHIS), a nationally representative, cross-sectional household survey conducted by the National Center for Health Statistics (NCHS). Using a geographic identifier (ie, census tract), we merged the NHIS data with external datasets including the 2013-2015 American Community Survey (ACS), the Primary Care Service Area (PCSA) version 3.1 data, the 2013-2015 Provider of Services (POS) files, the Food Access Research Atlas (FARA), and the 2011 National Land Cover Database (NLCD). Geographic variables including state, county, and tract were used to merge Census variables which provided neighborhood contextual information. State, county, and tract are restricted-use variables, and these data were accessed through the Research Data Center. Once annual datasets were merged separately by year, the merged datasets were appended vertically to create a single three-year pooled dataset. The final sample includes 10,399 children whose parents were not born in the U.S.
Table 1 presents the characteristics of the sampled children. About half of children were male (50.7%); 18.5% were Asian, 11.5% were Black; and 56.2% were of Hispanic origin. U.S. citizens composed 85.8% of the subsample, and the remaining children were either foreign-born U.S. citizens (4.6%) or foreign-born non-U.S. citizens (9.6%). Approximately, 21.0% were 3 years of age or younger; 17.2% were between 4 and 6 years of age; 34.7% were school-aged between 7 and 12 years of age; and 27.4% were between 13 and 17 years of age. Regarding parent respondents, most of them were female (97.9%). Over half were Hispanic (57.4%); 20.2% were Asian, and 11.1% were Black. Most parents were non-U.S. citizens (56.3%), and most were in their 30’s (47.0%) or 40’s (33.6%). Over a third had a high school education or less (37.7%); 53.7% were employed; and 77.1% were married.
Frequencies for Sampled Children and Parents from a Pooled Analysis of 2013-2015 Data (n = 49,775,187; Weighted).
Compared to the U.S. Census Bureau data, the current samples are considered representative of the population of immigrant children. The American Community Survey (ACS) in 2014 indicated that 51.2% of immigrant children were male; 20.2% were Asian; 10.0% were Black; and 54.9% were of Hispanic origin. Non-U.S. citizens comprised 9.9% of the immigrant children. Regarding the child’s age, 21.0% were 3 years of age or younger, 17.2% were between 4 and 6 years of age, 34.7% were between 7 and 12 years of age, and 27.0% were between 13 and 17 years or age. 25
Measures
Neighborhood characteristics
Parents’ perception of neighborhood trust was assessed with a four-item scale (α = .89). Parents were asked to indicate the extent to which they agreed with the following statements: (a) people in this neighborhood helped each other out, (b) there were people they could count on in this neighborhood, (c) people in this neighborhood could be trusted, and (d) this was a close-knit neighborhood. Seven objective measures of neighborhood characteristics were included. Poverty rate was measured as the percentage of families living under the federal poverty level. Foreign-born rate was measured as the percentage of foreign-born individuals among census tract population. The presence of healthcare centers was constructed from a summarized number of hospitals, health centers, home health agencies, and rural health clinics and transformed into a dichotomous variable for sample distribution. The number of primary care physicians was categorized into a dichotomous variable to indicate the presence of a primary care physician in each census tract. Food desert was measured as the percentage of individuals who are beyond 1 mile away from the nearest supermarket for urban areas and 10 miles from the nearest supermarket for rural areas. Green space was measured by green space accessibility index including developed open space, grass, shrub, and more.
Family characteristics
We included the sampled households’ residential area, income level, food insecurity, and healthcare accessibility and affordability issues as covariates. Food insecurity was assessed with an 8-item scale (α = .88). Parents were asked whether: (a) they worried whether their food would run out, (b) the food did not last and they did not have money to buy more, (c) they could not afford to eat balanced meals, (d) they ever cut the size of their meals or skip meals, (e) they ever ate less than they should, (f) they were ever hungry but did not eat, (g) they lost weight because there was not enough money to buy food, or (h) they did not eat for a whole day. Healthcare accessibility issues were measured with a 5-item scale (α = .59). Parents were asked whether their child had delayed medical care for the past 12 months because: (a) they could not get through on the phone, (b) they could not get an appointment, (c) they had to wait too long, (d) the clinic was not open, or (e) they did not have transportation. Healthcare affordability issues were measured with a 6-item scale (α = .66). Parents were asked if their child had not received health care for the past 12 months because they could not afford: (a) to see a specialist, (b) follow-up care, (c) prescription medicines, (d) mental health care or counseling, (e) dental care, or (f) eyeglasses.
Individual characteristics
The sampled children’s sex, race, ethnicity, nativity and citizenship, and age were measured. Such characteristics of the parents as sex, race, ethnicity, nativity and citizenship, age, educational attainment, marital status, employment status, and general health status were included.
Child health outcomes
Both physical and socioemotional health were measured based on parents’ perception. Parents were asked to indicate their children’s general health status with response options ranged from 1 (“poor”) to 5 (“excellent”). This variable is heavily left-skewed because approximately 85% of parents rated their children’s health status as good or better; therefore, we defined “poor physical health” as being rated as either poor or fair in this study, using this dichotomized variable (1 = “poor or fair,” 0 = “good, very good, or excellent”) as the physical health outcome. Socioemotional health was measured with a 6-item scale (α = .56). Parents were asked whether their child: (a) was well behaved, (b) worried or seemed worried, (c) was unhappy, depressed, or tearful, (d) got along better with adults than peers, (e) had good attention, and (f) had difficulties with emotions or concentration, all for the past 6 months. Defining “poor socioemotional health” as the 15th percentile of the scale composite score, we used a dichotomous variable in our model. Similar to this method, previous studies have created dichotomous variables for child outcomes using reclassification or percentiles. 26,27
Results
Table 2 presents the odds ratios (OR) of the predictors in the physical and socioemotional health models. As shown in the left columns in Table 2, girls were approximately 15% less likely to demonstrate poor physical health than boys (OR = .848). Older parental age decreased odds of poor physical health in children (OR = .985), indicating that 30-year-old parents are 15% less likely to have children with poor health status than 20-year-old parents. Compared to white-immigrant parents, Hispanic (OR = 1.448) and Asian and other immigrants (OR = 1.614) were about 1.4 to 1.6 times more likely to have children with poor physical health. Higher levels of parents’ educational attainment were significantly associated with lower odds ratios of their child’s poor health status (OR = .921). Parents’ general health status was also found as a significant factor to reduce the odds of children’s poor health status (OR = .256). Healthcare accessibility or affordability issues were associated with increased odds of poor health status (OR = 1.208). Higher levels of perceived neighborhood trust were significantly related to reduced odds of poor health status (OR = .911). Children from immigrant families who resided in higher poverty areas were more likely to demonstrate poor health status than those in lower poverty areas (OR = 1.049). Children living in neighborhoods where there was a clinically active primary care physician were approximately 14% less likely to have poor health status than those in neighborhoods without a primary care physician (OR = .858).
Predictors to Determine Poor Physical and Social Emotional Health of Children from Immigrant Families (n = 52,593,647).
Note. *P < .05, **P < .01, ***P < .001; weighted model statistics: observations = 10,292, strata = 553, PSUs = 3,147, population size = 52,593,647.
Logistic regression models for children’s socioemotional health (Table 2; right columns) also indicated children’s gender (OR = .702), parents’ general health status (OR = .807), healthcare issues (OR = 1.732), and neighborhood trust (OR = .807) as significant predictors. Compared to white parents, non-white groups were associated with increased odds of poor socioemotional health (ORs ranged from .554 to .676). Children from unmarried parents were about 50% more likely to demonstrate poor socioemotional health than those from married parents (ORs ranged from 1.523 to 1.564). Family income (OR = .977) and food insecurity (OR = 1.598) were also found as a significant factor. Neighborhood trust was the only neighborhood-level variable significantly associated with children’s socioemotional health in this model.
Figures 1 and 2 summarize risk and protective factors by the size of odds ratios. The physical health model identifies parents’ racial and ethnic minority status, healthcare issues, and the area-level poverty as risk factors and parents’ education, perceived neighborhood trust, the presence of primary physician, child’s sex (being female), and parents’ self-rated health status as protective factors (see Figure 1). In the socioemotional health model, we found that healthcare issues, food insecurity, and non-marital status significantly increased the risk of having poor socioemotional health while parents’ health status, neighborhood trust, being a female child, and racial and ethnic minority status decreased the risk (see Figure 2). Parents’ neighborhood trust, racial and ethnic minority status, health status, and their children’s sex (being female) were found as common predictors to determine both physical and socioemotional health of children in immigrant families.

Risk and protective factors for poor physical health by odds ratio.

Risk and protective factors for poor socioemotional health by odds ratio.
Discussion
Our findings highlight the significance of both perceived and objective neighborhood conditions for the health and well-being of children in immigrant families. Neighborhood trust assessed by immigrant parents’ perception was identified as a protective factor that facilitates both physical and socioemotional health of children. Immigrant parents who reside in a neighborhood environment with more social support and trust tend to report better physical and socioemotional development of their children. 4 The presence of primary care physician was another protective factor for children’s physical health, which supports prior research indicating that distance from a primary care physician can contribute to a reluctance to seek care, and consequently, to poorer health outcomes. 28,30,31 The current findings suggest that immigrant families living in high poverty areas are likely to experience difficulties in obtaining necessary resources and supports for their children 32 ; consequently, the children are at greater risk for poor health outcomes than those in more affluent neighborhoods.
Immigrant families’ financial resources were also identified as robust predictors of children’s socioemotional health. Lower levels of family income predicted poorer socioemotional health, which indicates that children without proper economic means are at high risk for the underuse of healthcare services, and consequently, behavioral and psychological problems. 33 Children in food insecure households had significantly higher odds of poor socioemotional health, supporting evidence that household food insecurity is associated with chronic stress and poor mental and emotional health outcomes in children and parents. 34
Other risk factors at individual and family levels such as non-marital status and racial and ethnic minority status should be highlighted. Children had poorer socioemotional health outcomes when parents were never-married, widowed, divorced, or separated, which is consistent with prior findings. 12 We find that immigrant children of parents of color experience poorer physical health outcomes and better socioemotional health outcomes compared to immigrant children of white parents. This contrast indicates that the social consequences of how parents (and thereby children) are racialized influence child health. White racial privilege functions within immigrant groups, 35 even after controlling for parental education, income, English proficiency, and access to health care. In the racial context of the United States, the social effects of perceived whiteness may open the doors to resources for optimal physical health, while promoting narratives that undermine mental health. 36 In addition, immigrant children of non-European origin tend to have stronger social networks and family cohesion which may boost their psychological health. 37
Turning to other protective factors, it should be noted that child’s sex, parental age and education, and parental health status predicted child health outcomes. Female children had higher levels of both physical and socioemotional health than male children. 11 Children with older parents showed better physical health but not socioemotional health. Aligned with previous research, higher parent education predicted better physical health. 14,16 Higher parental self-rated health was significantly associated with better parent ratings of child physical and socioemotional health for all children, which also aligns with previous evidence. 38 Concerning healthcare access and affordability, reduced access to care results in worse health outcomes for children. 39 Our finding supports that the presence and availability of healthcare services is a protective factor for children’s health and development 7,10
Despite these findings, several limitations should be noted. First, we used objective measures of neighborhood based on census tract; however, census-defined units may not necessarily match residents’ perspectives of neighborhood boundaries. 40 Immigrants typically do not perceive their own neighborhoods as spatially restricted to specific residential or administrative boundaries; instead, they are more likely to construct their daily routine activities to obtain services and goods further from their residential areas. 41 Second, other subjective measures of neighborhood—except neighborhood trust—were not included due to data unavailability. Previous research has demonstrated the significance of subjective perceptions of various neighborhood social processes and conditions, including neighborhood physical characteristics, safety, crime, social cohesion, and collective efficacy. 29,4 Third, the use of the subjective predictors may have introduced shared variance with parent-rated health outcomes, as both measures were parent-report. Fourth, the legal residency status of immigrants was not accounted for. Undocumented immigrants have limited access to health care and other public services; and barriers to health care for undocumented immigrants go beyond policy and range from financial limitations, to discrimination and fear of deportation, 42 which may ultimately impact on their children’s health. Finally, it is important to note that many factors may influence immigrants’ self-reported health status. Immigrants’ self-reported health can be influenced by official language proficiency, discrimination experience, age, healthcare access barriers, neighborhood social environment, 43 acculturation and dietary changes, 44 and other health-related and socioeconomic variables. 45 Several of these factors were not controlled for in this analysis.
In conclusion, our multi-layered analysis of ecological factors that predict children’s physical and socioemotional health points to neighborhood trust as a key point intervention point with broad potential to lead to improved health outcomes for immigrant children. Given that immigrants in racially and ethnically diverse areas tend to have low levels of neighborhood trust, 46 culturally sensitive strategies to promote interpersonal contact, social interaction, and local community building would be beneficial. Importantly, programs focusing on immigrant families with children who suffer from economic hardship in high poverty neighborhoods should be a high priority. Given our finding that low-income and racial ethnic minority immigrant children experience poorer health outcomes, current immigration policies that restrict access to public assistance and welfare programs for children in low-income households (eg, food and medical assistance programs) should be reassessed to ensure that disadvantaged children have access to health care and social support. 47
So What?
What is already known on this topic?
Neighborhood is a prominent environmental context that influences children’s physical and socioemotional health. Although minority children are disproportionally represented in disadvantaged neighborhoods, research examining the effects of neighborhood on child health outcomes in immigrant families remain sparse.
What does this article add?
This article adds to current knowledge by finding the significance of both perceived and objective neighborhood conditions for the health and well-being of children in immigrant families. In particular, this article highlights the importance of neighborhood trust, area-level poverty rate, and the presence of primary care physician as significant predictors for child health outcomes.
What are the implications for health promotion practice?
Future health promotion efforts should leverage culturally sensitive strategies to promote interpersonal contact, social interaction, and local community building. Programs focusing on immigrant families with children in high poverty neighborhoods should be a high priority.
Footnotes
Authors’ Note
J. Choi conceived the study, analyzed the data, and wrote the first draft of the article. M. Kelley contributed to the conception of the study, interpreted the findings, and wrote the first draft. D. Wang and H. Kerby conducted literature reviews and interpreted the data. All authors reviewed, revised, and approved the article. Ethics approval was not required for this study because the data used were publicly available and posted without an expectation of privacy. Preliminary findings from this study were presented at the American Public Health Association Annual Meeting; Philadelphia, PA; November 2-6, 2019. The findings and conclusions in this paper are those of the author(s) and do not necessarily represent the views of the Research Data Center, the National Center for Health Statistics, or the Centers for Disease Control and Prevention.
Acknowledgments
The authors wish to thank Dr. Jinyoung Lee for her contributions to the RDC research proposal. The findings and conclusions in this paper are those of the author(s) and do not necessarily represent the views of the Research Data Center, the National Center for Health Statistics, or the Centers for Disease Control and Prevention.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Central Plains Research Data Center [grant number 25-0544-0002-014].
