In 2009–2010, 17% of United States (US) children ages 2–19 years were obese, with greater odds of obesity among black and Hispanic children. Even as obesity rates appear to have plateaued in select US subgroups, rates have increased in countries such as Mexico that account for a high proportion of immigrants to the US. [1] The percentage of US children ages 0–17 years with at least one parent born outside the US grew from 15% in 1994 to 23% in 2011. [2] Though little is known about the influence of having a non-US born parent on obesity-related behaviors in childhood, [3, 4] designing appropriate interventions for these children will be key to addressing racial/ethnic disparities. To tailor messages and plan effective interventions, public health practitioners must understand how obesity-related behaviors and home environments differ by parental US born status.

Evidence suggests that immigrant adults have lower obesity risk despite lower income and educational attainment. [5] Varying hypotheses seek to explain this “immigrant paradox,” including a buffering effect of strong social networks and family ties, a healthy migrant effect, or reporting bias. Even if associations with birthplace represent true protective effects, there is conflicting evidence as to whether the children of immigrants benefit. Neighborhood level studies show lower rates of childhood obesity in communities with a high proportion of non-US born residents. [6] By contrast, data from the Early Childhood Longitudinal Study indicate that having a non-US born mother is not protective of childhood obesity [7] and may even be associated with elevated obesity risk. [8] Parents’ level of involvement in the ethnic culture of birth or the new culture of settlement can influence parenting behavior directly, through cultural beliefs about effective parenting techniques and what qualities parents should value in their children, and indirectly, through acculturative stress. [9–11] Parental US born status may interact with socioeconomic constraints, level of acculturation and race/ethnicity to influence behaviors from child feeding to mother-child interaction. Adjustment to a new culture takes time, but the rapid adoption of obesogenic dietary and lifestyle habits in immigrant households may combine with perceptions of heavy as healthy to elevate obesity risk in children. [12–14]

This study examined differences by parental US born status in obesity-related behaviors of 2 to 5.9 year old children participating in a randomized controlled trial in the greater Boston area. The behaviors of interest in this study were previously associated with US born status in adults: prior research has documented beneficial behaviors among immigrant families such as longer sleep duration, lower fast food consumption and fewer hours of television (TV) viewing, but also adverse behaviors including less physical activity. [3, 4, 15] We hypothesized that having a non-US born parent would be associated with a similar risk profile in children, e.g. longer sleep duration and greater frequency of meals eaten in the home, but fewer hours of active play.

Study subjects were parent-child dyads participating in the baseline assessment of the Healthy Habits, Happy Homes study, a randomized controlled trial to improve household routines to prevent obesity among children 2–5.9 years of age. [16] The study took place from June, 2011 to August, 2012 and enrolled children who received their pediatric care at 4 federally-qualified community health centers in the greater Boston area. Parent-child dyads were considered eligible if: 1) at least one parent was > 18 years old; 2) the index child was 2 to 5.9 years old; 3) the child had a TV in the room where s/he sleeps; 4) the family was not planning to move from the Boston area within the study period; 5) the child did not have a medical condition requiring chronic care; and 6) at least one parent was comfortable responding to interviews and reading in English or Spanish. There were no BMI eligibility requirements for parents or children.

We telephoned parents to screen for eligibility, obtain verbal consent, collect contact information, and schedule a home visit. Staff collected all baseline data and obtained written consent at a scheduled home visit, which lasted between 40 minutes to 1 hour. If more than one adult headed the household, one adult primary caregiver agreed to be the study participant, respond to survey questions and be present at study home visits. During the home visit, staff measured the child’s height and weight and used a home environment inventory to assess the number of rooms in the home and their use, the number and location of media devices in the home and distance to and safety of outdoor play spaces in the surrounding neighborhood. Staff also administered a baseline survey to the participating adult in either Spanish or English. All study activities were approved by the human subjects committee at Harvard Pilgrim Health Care. Further information on the study site, participant recruitment and eligibility criteria is available in the baseline paper from this trial. [16]

Selected characteristics of parent-child dyads in the study sample are shown in Table 1. Over 90% of adult respondents were the participating child’s biological mother; 6 fathers, 1 aunt and 1 grandparent also participated. Nearly 60% of the sample had annual household incomes at or below $20,000. The mean (SD) age of children was 4.0 (1.1) years, and mean (SD) BMI z-score was 0.94 (1.16). Neither age nor BMI z-score differed between groups. Children in the study were predominantly Hispanic (52%) or Black (34%); race/ethnicity differed between groups: children of non-US born parents were 69% Hispanic and 19% Black, while children of US born parents were 33% Hispanic and 51% Black. Nearly half of the non-US parents were from Central America; parents were also from other countries in Latin America, and a small number of parents were from South Asia and Africa. Children of non-US born parents were less likely to be enrolled in childcare (61% v. 83%). Overall, non-US born households had lower educational attainment and a greater number of adult residents.

Table 2 shows the prevalence of obesity-related behaviors by US born status. Children of non-US (v. US) born parents drank more sugar-sweetened beverages (11.3 v. 5.9 servings/week), engaged in less active play (1.2 v. 1.4 hours/day), had later weekday bedtimes (9:23 v. 8:40p.m.) and wake-up times (7:42 v. 7:14 a.m.) and were less likely to have a consistent, calming nighttime routine (41% v. 65%). Both groups of children drank 100% juice an average of more than once per day, and there were no differences by US born status in the frequency of family meals, fast food consumption, child or parent sleep duration, co-sleeping, how consistently TVs were on in the home (e.g. during meals or when no one was watching), nor whether the parent put limits on TV-viewing. We found no differences in access to outdoor play space or in the proximity and safety of nearby parks by US born status.

Table 3 shows bivariate and adjusted associations of non-US born status and obesity-related behaviors. After adjustment for parental educational attainment, children of non-US (v. US) born parents consumed more soda (0.67 additional servings/week; 95% CI: 0.02, 1.32) and total sugar-sweetened beverages (4.42 additional servings/week; 95% CI: 0.35, 8.49), and engaged in less active play (0.16 fewer hours/day; 95% CI: −0.29, −0.03). Children of non-US born parents watched 44 fewer minutes of TV daily than children of US born parents (0.74 fewer hours/day; 95% CI: −1.35, −0.14) after adjusting for parental educational attainment. Children of non-US born parents also had later weekday bedtimes (0.75 hours later; 95% CI: 0.32, 1.17) and wake-up times (0.51 hours later; 95% CI: 0.13, 0.89) after adjusting for parental educational attainment.

As shown in Table 3, further adjustment for child race/ethnicity removed the protective association of non-US born status with child TV-viewing, but strengthened associations with lower odds of a calming nighttime routine and later weekday bed and wake times. Adjustment for child race/ethnicity did not substantially affect the association of non-US born status lower child active play and higher child soda consumption. Additional adjustment for child BMI z-score did not substantially change any results (data not shown).

Among the adults, in education-adjusted models non-US (v. US) born parents drank fewer sugar-sweetened beverages (5.30 fewer servings/week; 95% CI: −9.10, −1.50); were exposed to fewer total hours of screen time (1.41 fewer hours/day; 95% CI: −2.17, −0.66); and were less likely to eat out one or more times per week (OR = 0.29; 95% CI: 0.08, 0.998). Parent fast food consumption, sleep duration and physical activity did not differ by US born status. Further adjustment for race/ethnicity removed the protective association of non-US born status with parent sugar-sweetened beverage intake and frequency of eating out, but associations with lower screen time remained.

The prevalence of co-sleeping (sharing a bed or bedroom with a parent) did not differ between children of US born parents and non-US born parents (14% v. 30% prevalence of co-sleepers); however, children that did co-sleep were less likely to sleep the recommended ≤ 11 hours per day (OR = 0.46; 95% CI: 0.22, 0.97; data not shown).

Within the non-US born group, number of years spent in the US did not predict parent or child screen time, sleep, or physical activity, but longer parental residence in the US was associated with somewhat higher child fast food consumption (0.04 meals/week; 95% CI: 0.02, 0.07) and parent sugar-sweetened beverage consumption (0.06 servings/day; 95% CI: 0.01, 0.10), even after adjusting for educational attainment.

In this study of low-income, urban, racial/ethnic minority families, we hypothesized that having a non-US born parent would be associated with a profile of both adverse and protective obesity-related behavior in children; however, in models adjusted for both education and race/ethnicity we observed only adverse associations. In initial models adjusted for education only, children of non-US born parents watched fewer hours of TV but consumed more sugar-sweetened beverages and engaged in less active play. We also found several obesity-protective behaviors among the adults when adjusting for education only: non-US (v. US) born parents had less screen time exposure and ate fewer meals outside the home, and, in contrast to their children, consumed fewer sugar-sweetened beverages. However, the influence of non-US born status on certain obesogenic behaviors appeared to vary by race/ethnicity: adjustment for child race/ethnicity removed protective association of parental non-US born status with child TV-viewing, parent sugar-sweetened beverage intake and frequency of eating out, but strengthened associations with lower odds of a calming nighttime routine and later weekday bed and wake times, implying that both nativity and race/ethnicity should be consider when designing intervention messages. Tenure in the US had a small impact on parent and child behaviors: among non-US born parents, the number of years living in the US was associated with modestly higher child fast food consumption and parental sugar-sweetened beverage intake.

Consistent with previous studies of low-income and racial/ethnic minority preschool children, consumption of 100% juice and sugar-sweetened beverages was common for all participants. [40, 41] We found soda consumption to be more prevalent among children of non-US (v. US) born parents. Prior studies have found soda to be more available in Hispanic households, and hypothesized that other forms of sugar-sweetened beverages, such as fruit drinks, might be more common amongst African Americans. [42] Sugar-sweetened beverage consumption is an important intervention focus for low-income families in general, and for children of immigrants in particular. [43] Intervention messages should include attention to culturally-specific sugar-sweetened beverages, including aguas frescas, fruit drinks and flavored milk. A potential reason for the higher prevalence of soda intake among children of immigrants could be differing perceptions of children’s risk of obesity; studies of Latina mothers suggest that perceptions of chubby children as healthy or that children’s excess weight will resolve with growth are more common amongst immigrant v. US born mothers. [12, 13]

We also found that children of non-US born parents engaged in less active play. In contrast to this finding, a study using data from the National Survey of Children's Health documented greater physical activity among immigrant children; this same study found also lower TV-viewing among immigrant children, as was observed in our sample when adjusting for education only. [44] It is possible some of the differences in physical activity between children of US and non-US born parents could be accounted for by unmeasured factors such as neighborhood characteristics. However, all of our study participants lived in the same set of neighborhoods in the Boston area, and we found no differences in access to outdoor play space or in the proximity and safety of nearby parks by US born status.

To our knowledge, our study is the first to examine sleep routines among children of non-US born parents. Child sleep duration did not differ by US born status. However, we observed later bed and wake times and lower odds of a calming nighttime routine among children of non-US born parents. Bedtime may represent an important intervention focus: later bedtimes are associated with shorter sleep duration in preschool children. [30] As the children of non-US born parents enter school, late bedtimes could result in insufficient sleep since the late weekday wakeup times will no longer be possible. Several important contextual factors could account for late bedtimes: first, one third of non-US born parents reported sharing a bed or bedroom with their child. Second, later bedtimes, co-sleeping and lack of a calming routine may be due to space constraints because of the greater numbers of adults in the homes of non-US born parents. The prevalence of co-sleeping did not differ significantly by US born status; however, co-sleeping children were less likely to sleep the recommended ≤11 hours. Considering these aspects of the home environment (co-sleeping and number of adults) will ensure that interventions targeting sleep and bedtime routines develop strategies compatible with the realities of the immigrant families’ living situations.

Consistent with other studies that have found foreign birthplace to be protective with respect to dietary patterns, [45] we found less frequent eating out and lower sugar-sweetened beverages among non-US born parents in education-adjusted models; however, further adjustment for race/ethnicity removed protective associations. We did not find differences in parent fast food consumption or frequency of family meals, though foreign birthplace has been protective in other studies. [46] Non-US born parents were also exposed to an hour less of screen time daily than non-US born parents; this could be due in part to employment status: unemployed parents were exposed to more hours of screen time than parents who worked or studied part or full time, and a greater proportion of unemployed parents were born in the US in this sample.

A systematic review examining children’s diets showed no definitive associations with parental acculturation; [47] however, in this study, we found the number of years living in the US was associated with modest increases in parent sugar-sweetened beverage intake among non-US born parents as well as in the fast food consumption of their children.