Growth of Pediatric Refugees after Resettlement to the Southeastern United States

Abstract

Objective:

To compare weight status and body mass index z-scores (BMIz) of refugee children upon arrival to the Southeastern US and after resettlement with age- and sex-matched controls of non-refugee children.

Methods:

We identified refugee children resettled between July 2014–June 2018 to Forsyth County, NC. Refugees were assigned age- and sex-matched controls (3 controls: 1 refugee) who received care at the same healthcare site, were Medicaid insured, and had height and weight recorded at time of their matched refugee’s resettlement plus at least 2 additional visits. BMI and BMIz were calculated. Pearson chi-square test assessed differences in weight status at the time of refugee resettlement and at the last measured time point. Multilevel linear mixed-effects regression models assessed change in BMIz by refugee status, adjusting for sex, race/ethnicity, age, and time since resettlement.

Results:

This study examined 139 pediatric refugees and 417 non-refugee controls;46% were female. At the time of resettlement, refugees had a higher rate of underweight (3.3% vs. 1.9%), higher rate of healthy weight (68.9% vs. 48.2%) and lower rate of overweight/obesity (27.8% vs. 50%), compared to controls p<0.001. At the last available time point, compared to controls, refugees had higher rates of underweight (3.3% vs. 3.0%) and healthy weight (69.5% vs. 54.2%) and lower rates of overweight/obesity (27.2% vs. 42.7%), p=0.005. Refugees had a lower predicted BMIz compared with controls (adjusted β: −0.78; 95% CI −0.91, −0.66).

Conclusion:

Resettled pediatric refugees were found to have significantly lower predicted BMIz than their age- and sex-matched non-refugee controls.

INTRODUCTION

Childhood nutrition has important implications in cognitive development, chronic disease risk, and adult weight status.^{1, 2} Refugee children represent a vulnerable expanding patient population in the United States (US), who come from regions with high rates of malnutrition, infection, and poverty.³ In 2018, Homeland Security reported that 22,405 refugees were admitted to the US with 47% being dependent children.⁵ Refugees undergo health screenings prior to departure to the US and again within 30-60 days of arrival to the US, including height and weight measurements.⁶

Many refugee children are affected by undernutrition (wasting, stunting) or overnutrition (overweight, obesity) with disparities in both directions associated with age, country of origin, and location of refugee camp.^7-10 Many refugee children demonstrate catch up growth upon exposure to US diet,¹¹ and numerous studies have shown that refugee children resettling to the Pacific Northwest or Northeast, compared to US children, have a significant rate of increased body mass index (BMI), overweight and obesity following resettlement.^12-15 However, since these studies were conducted, refugee resettlement patterns have evolved in the setting of global events and US priorities.^{16, 17}

US pediatricians report being unprepared to care for immigrant children; therefore, it is imperative that studies evaluate gaps in knowledge uniquely affecting these populations.¹⁸ Currently, the Southeastern US has poor nutrition with high levels of overweight/obesity, however, no studies have investigated the long-term growth outcomes of refugee children after resettlement to the Southeast.^{19, 20} Additionally, given evolving refugee resettlement patterns, more recent data on trends in refugee growth data are necessary to identify differences in various refugee populations. Moreover, few studies compare growth outcomes of refugee children compared to their sex-, age-, and location-matched controls.^{8, 14} Therefore, we aim to compare change in weight status and body mass index z-score (BMIz) of refugee children over time after resettlement with age-, sex-, and location-matched controls in the Southeast.

METHODS

Study Population and Design

Refugee participants were identified utilizing data provided from a local resettlement agency. Refugees were included if they resettled into Forsyth County, North Carolina (NC) between July 2014-June 2018, were <22 years of age at resettlement, received care at 1 of 3 healthcare sites: Wake Forest Baptist Medical Health (academic medical center), Novant Health (community clinic) and United Health Center (Federally Qualified Health Center), and had a height and weight measured at the time of resettlement and at least one year after. Refugees were excluded if they relocated after previous resettlement in another state or if they had documentation in their electronic health record (EHR) of: (1) failure to thrive, (2) any chronic medical contrition that affects growth, or (3) significant developmental delay or intellectual disability.

Control participants were selected by identifying eligible patients with the closest birthdate to the refugee. Controls were included if they were of the same sex, received care at the same primary care clinic, were Medicaid insured, and had height and weight recorded at time of their matched refugee’s resettlement plus at least two additional visits, one of which was at least one year after their refugee’s resettlement. Controls were excluded if they had documentation in the EHR of: (1) refugee status, (2) failure to thrive, (3) any chronic medical contrition that affects growth, or (4) significant developmental delay or intellectual disability. We obtained 3 controls for each refugee participant. This study was approved by the Institutional Review Boards of Wake Forest School of Medicine and Novant Health.

Study Measures

Demographic data provided by the local resettlement agency included date of birth, sex, race, date of arrival to the US, and country of origin. Demographic data were extracted for all participants from the EHR, including age, sex, race, ethnicity, language, visit date and health care institution. We categorized race/ethnicity as non-Hispanic White, non-Hispanic Black, Hispanic/Latinx, or other. The age of study participants was recorded as the age of refugee and matched control on the date of refugee initial visit.

Height and weight were extracted from the EHR. We extracted data from the time of refugee resettlement and onward for refugees and one year prior to the matched refugee’s resettlement and onward for controls. Body mass index percentiles and z-score (BMIz) were calculated using age- and sex-specific World Health Organization growth curves.²¹ Children and adolescents with BMI below <5^th percentile were classified as having underweight, 5^th to <85th percentile as healthy weight, and ≥85^th percentile as overweight/obesity.

Statistical Analyses

We used Pearson chi-square test to assess differences in weight status by refugee status at the time of resettlement and at the last measured time point (> 1 year from resettlement). We used multilevel linear mixed-effects modeling to assess differences in BMIz by refugee status, including sex, race/ethnicity, age, age squared, and time since resettlement as fixed effects and repeated measurements of the same individual over time as a random effect. We tested the model with just random intercepts and with both random intercepts and random slopes, but as model results were nearly identical, we utilized the model with only random intercepts for this analysis. We tested a time by refugee interaction term in the model as well, however, this was not significant and did not change the effect size so was not included in the final analyses. We also ran models stratified by weight status (healthy weight and overweight/obesity). Statistical analyses were performed using Stata v14.2 (StataCorp, College Station, TX) and significance was assessed using a 2-sided test at α=0.05.

RESULTS

Study Population

The local resettlement agency provided us with information on 161 refugees. Five refugees had no EHR, three did not have a visit with a primary care provider, four had medical conditions affecting weight gain (neuromuscular disease, growth hormone deficiency, bulimia nervosa, and pregnancy), and ten did not have a subsequent height/weight measured. Our study sample included 139 refugees and 417 age-, sex- and location-matched controls. Participants were an average of 7.4 years old (SD 4.4). Refugees’ country of origin included Syria (42%), Democratic Republic of Congo (10%), Burma (9%), or Iraq (7%). A large percentage of the age- and sex-matched controls identified as Hispanic/Latinx (62%) (Table 1).

Table 1:

Participant Characteristics

Characteristics:	Refugee (%) N=139	Control (%) N=417
Sex
Female	46.4	46.4
Age
0 to 5 years old	41.4	41.4
6 to 10 years old	37.9	37.9
11 to 20 years old	30.7	30.7
Refugee Country of Origin
Syria	47.7	NA
Burma	18.2
Iraq	15.9
Other	18.2
Race/Ethnicity
non-Hispanic White	2.4	7.1
non-Hispanic Black	30.1	28.1
Hispanic/Latinx	1.6	61.8
Other	65.9	2.9
Healthcare Site
Academic Medical Center	75.8	75.8
Community Hospital	8.5	8.5
Federally Qualified Health Center	15.7	15.7
Weight Status at Resettlement
Underweight	2.2	2.1
Healthy Weight	72.7	49.2
Overweight or Obesity	25.2	48.7
Weight Status at Time of Last Visit
Underweight	1.2	1.2
Healthy Weight	65.9	47.5
Overweight or Obesity	32.9	51.3

Weight Status

At the time of resettlement, refugees, compared to controls, had similar rates of underweight, higher rates of healthy weight, and lower rates of overweight or obesity (p<0.001) (Table 1). At the last available time point, refugees still had higher rates of healthy weight compared to controls and lower rates of overweight or obesity (p=0.01). Both refugees and controls had an increase in overweight/obesity at the time of last visit and a slight decrease in underweight and healthy weight (p<0.01 for both).

Growth Trajectoiy

A total of 556 children and 2,849 observations were included in a multilevel linear mixed-effects model that found that refugees had a lower predicted BMIz compared to controls (adjusted β: −0.63; 95% CI −0.94, −0.33) (Table 2). As time since resettlement increased, the confidence interval widened, with a wider confidence interval appreciated in the refugee population than the controls (Figure 1A).

Table 2:

Differences in BMIz by refugee status, results from multilevel linear mixed-effects modeling

	Beta Coefficient (95% Confidence Interval)
Refugee	−0.63 (−0.94, −0.33)***
Male Sex	0.06 (−0.17, 0.29)
Race/Ethnicity	0.10 (−0.05, 0.26)
Health System	0.001 (−0.18, 0.18)
Time Since Resettlement (in years)	0.07 (0.04, 0.18)**
Age (in years)	0.11 (0.04, 0.18)**
Age Squared (in years)	−0.004 (−0.007, −0.0004)*

^*denotes p<0.05

^**p<0.01

^***p<0.001

Figure 1.

Predicted BMIz by refugee status over time for: A) All Participants; B) Refugees with a healthy weight at time of resettlement and their controls; C) Refugees with a weight status of overweight or obesity at time of resettlement and their controls. Time on the x-axis represents the time since resettlement (for refugees) or time since resettlement of the control’s matched refugee (for controls).

In analyses stratified by baseline weight status, refugee children who had a healthy weight at the time of resettlement did not have a significantly different predicted BMIz (adjusted β: −0.11; 95% CI −0.32, 0.09) compared to controls who had a healthy weight at the time of resettlement (Figure 1B). Similarly, among children who had overweight or obesity at the time of resettlement there were no statistically significant differences in predicted BMIz over time between refugees and controls (adjusted β: −0.29; 95% CI −0.64, 0.05) (Figure 1C).

DISCUSSION

To our knowledge, this study is the first to evaluate pediatric refugee patients’ growth longitudinally in the Southeastern US compared to age-, sex- and location-matched controls. This study found that refugees had similar rates of underweight, higher rates of healthy weight, and lower rates of overweight/obesity compared to controls upon resettlement and post resettlement. Refugees and controls both experienced an increased in overweight or obesity at the time of last visit compared to the resettlement visit. Refugees had lower predicted BMIz compared with controls over time. Both refugees and controls had a widening of BMIz confidence intervals over time and refugee populations had wider BMIz confidence intervals compared to controls.

Our results were similar to studies that found that rates of overweight and obesity increased after resettlement for pediatric refugee patients. A study in Rhode Island evaluating change in BMI percentiles in refugee children found an increase in prevalence of overweight and obesity at initial intake to 3 years post resettlement.¹² Similarly, a study evaluating the longitudinal weight trajectory of pediatric refugees in New York found an increase in obesity rates in the 9 years after resettlement.¹³ Our study shows that the increasing rates of overweight and obesity in refugees found in resettlement to the Northeast holds true in NC. Our stratified analysis did not show significant differences between refugees and controls when examining children with overweight/obesity separately from children with healthy weight. For refugee children with overweight/obesity, however, the confidence intervals widen considerably at four years after resettlement, likely due to increased variation in growth and the smaller sample size of the refugee participant population. Taken together, these findings support the importance of creating obesity prevention programs that are culturally effective for refugee populations in order to limit the increasing rate of obesity.

Our results differed slightly from previous studies which have found that over time refugees had a steeper increase in the rate of overweight or obesity compared to that of controls. One study evaluating growth trajectories found that refugees had a higher risk of obesity over 10-24 months post resettlement than nonrefugees.¹⁴ One reason for this difference could be the Southeast’s higher rates of overweight/obesity, and therefore controls in the Southeast likely have higher rates of overweight/obesity than controls in other studies.^{19, 20} This suggests that despite the obesity rates of controls being higher at baseline, there is still an importance on focusing on healthy lifestyle behaviors in refugee patients to prevent obesity. Another possible reason for the difference is that our study evaluated BMI status within 1-2 years of resettlement; therefore our, study could be displaying the “healthy immigrant effect” in which recently arrived refugee and immigrant populations experiencing less acculturation consistently associate with better health outcomes.²² Moreover, there could have been existing differences between our control and refugee populations that we were unable to measure that could have influenced weight status, such as social support and acculturation.

Our results also differed from previous studies finding that refugees have an underweight status on arrival. In our population, however, only 2% of refugee children had underweight. This could be due to differences in weight status by country of origin. In our study, most refugees were from Syria, while in similar studies most refugees were from Bhutan, Iraq, Somalia, or Burma.^{8, 14, 15} Our findings of low underweight status in refugee children are consistent with previous literature in the Syrian population. Syrian refugees aged 6 to 59 months had a low prevalence of undernutrition (<5%) and a high prevalence of overweight or obesity (10.6%) prior to resettlement. The nutritional status of Syrian refugees living in Iraq, Jordan, and Lebanon was similar to that of host communities. The authors hypothesized that the low levels of acute malnutrition could in part be due to Syria’s strong agriculture and animal husbandry and to large relief interventions of international organizations during the Syrian crisis.^{23, 24}

One limitation of our study is the small sample size of eligible refugees, which may have affected study reliability. Additionally, after resettlement refugees were only required to have one subsequent visit for data collection, which limited our ability to track growth status over time. Furthermore, we were unable to evaluate refugee patients who subsequently moved out of Forsyth County or did not receive further medical care. Lastly, because our study was conducted in one region of the US, the Southeast, our study’s findings may not be generalizable to other refugee populations resettling outside our region.

Conclusion

Our results found that refugees had higher rates of healthy weight and lower rates of obesity compared to controls upon resettlement and at the last recorded visit. Given these differences, future studies could benefit from determining why these variations exist by exploring refugee patient populations’ eating habits, food procurement and preparation practices, and how and from whom they receive nutritional information. However, both refugees and controls had an increase in overweight and obesity over time. This suggests the importance of targeting culturally effective obesity prevention programs towards all children.

What’s New:

Refugees had similar rates of underweight, higher rates of healthy weight, and lower rates of overweight and obesity compared to controls upon resettlement and post resettlement. Refugees and controls both experienced an increased in overweight or obesity over time.