Dietary Factors and BMI Among Preschool-Aged Head Start Children and Home Caregivers in Remote Alaska Native Communities

Abstract

Background

American Indian and Alaska Native children experience disparities in obesity. Home caregiver diet and weight status may be important for obesity prevention efforts. The goals of this study were (1) to examine the concordance of diet and body mass index (BMI) between Yup’ik caregivers and preschoolers in remote Alaska and (2) to examine the association of dietary factors with BMI for caregivers.

Methods

Study data came from “Got Neqpiaq?”, a culturally centered multilevel intervention focused on Yup’ik preschool-aged children (n = 155) and caregivers (n = 144) in 12 communities in Southwest Alaska. Dietary factors of interest were measured using biomarkers: traditional food intake (nitrogen stable isotope ratio), processed food intake (carbon stable isotope ratio), and vegetable and fruit intake (skin carotenoid concentration). Associations among variables were evaluated using confounder-adjusted linear regression with BMI modeled as a continuous outcome.

Results

Dietary biomarkers were highly concordant between caregivers and children, but caregiver and child BMI were not. Among caregivers, traditional food intake was positively associated with BMI (beta = 1.57; 95% confidence interval [CI] = 0.45, 2.68; p = .006), vegetable and fruit intake was negatively associated with BMI (beta = − 0.02; 95% CI = − 0.04, − 0.004; p = .02), and processed food intake was not associated with BMI (beta = 0.74; 95% CI = − 1.07, 2.55; p = .42).

Conclusions

There was high concordance between caregiver and child diet which suggests that dietary-related obesity prevention efforts in Yup’ik communities could focus on family-level interventions. Future work must move beyond a singular focus on obesity and consider holistic health, which aligns with an Alaska Native worldview.

Background

There are considerable racial and ethnic disparities in childhood obesity in the U.S. Nationally representative data for the U.S. indicate that the overall prevalence of obesity is 12.7% among children aged 2–5 years [1]. In comparison, recent studies with Alaska Native communities in Southwest Alaska estimate that about 35% of Yup’ik children aged 3–5 years have obesity [2]. Long-term consequences of obesity include a higher risk of heart disease, type 2 diabetes, and some cancers [3, 4]. Despite disparities, obesity prevention and control efforts fail to include diverse racial and ethnic populations or to account for the cultural and contextual risk factors that drive disparities in diverse racial and ethnic groups [5–7].

Some interventions aimed at reducing health disparities and preventing obesity in Yup’ik populations have focused on modifiable risk factors for obesity like diet [2]. Other important modifiable risk factors in Yup’ik communities may include the diet, weight status, and feeding practices of the home caregiver [8]. Home caregivers shape child feeding practices through control of meal routines, what foods are accessible to children, and the social context of eating occasions [9–11]. These behaviors are impacted by the family context. According to the Family Ecological Model, caregiver feeding practice results from the context in which the family resides, including education level, cultural background, social support, and other organizational, community, policy, and media-related factors [12]. In Yup’ik communities, feeding practices are likely also driven by geographic, economic, and cultural factors such as adherence to traditional lifestyles/diets.

The association of caregiver dietary behavior with child diet in nationally representative U.S. populations has been reported extensively in the literature, but less is known about these relationships in Yup’ik populations. One study found that Yup’ik caregiver’s intake of traditional foods was positively associated with child intake of traditional foods and negatively associated with child body mass index (BMI) [13]. It is unknown whether caregiver intake of processed foods or vegetables and fruits influences child diet in Yup’ik communities. Improving the understanding of how the caregiver diet affects child diet in Yup’ik communities can inform obesity-prevention efforts in Yup’ik communities, particularly in the context of family-based interventions.

This study examined dietary factors and BMI among dyads of Yup’ik children enrolled in Head Start preschool programs and their primary caregivers. The first goal was to characterize the concordance of three dietary factors (traditional food intake, processed food intake, and vegetable and fruit intake) and BMI between children and caregivers. The second goal was to examine the association of dietary factors with BMI among caregivers. This study contributes to obesity-prevention efforts in Alaska Native populations which are understudied despite disparities in obesity and obesity-related diseases.

Materials and Methods

Study Population and Setting

The study used baseline data from a cluster-randomized trial designed to test the effectiveness of the culturally centered multilevel intervention, “Got Neqpiaq?”, on the prevalence of overweight and obesity among children aged 3–5 years enrolled in Head Start programs in 12 remote Yup’ik and Cup’ik communities in Southwest Alaska. The 12 study communities are located in the remote Yukon-Kuskokwim region of Southwest Alaska. The Yukon-Kuskokwim region is roughly the size of the U.S. state of Oregon and has a population of only about 25,000. The region is the homeland of Yup’ik and Cup’ik people, two related Alaska Native ethnic groups living in 58 federally recognized tribal communities of which about 80% are Yup’ik ethnicity. Communities practice a mixed economy, where families are reliant on varying levels of subsistence lifestyle practices (i.e., hunting, fishing, and gathering) and cash incomes. All communities are located off the road system; communities can only be reached by small plane year-round, or by boat in the summer and snow machine in the winter [14].

All children enrolled in Head Start in the 12 study communities and their primary caregiver were invited to participate in the study via flyers sent home by Head Start staff and by word of mouth. Study communities ranged in size from less than 300 residents to just over 1000. About 240 children aged 3–5 years were enrolled in 12 community Head Start programs and were eligible to participate each year. Classroom sizes ranged from 14 to 20 students, with two of the larger communities operating two separate classes each.

Participatory Research and Ethics

This study represents a partnership between an academic institution (University of Alaska Fairbanks [UAF]), two Alaska Native tribal health organizations (Alaska Native Tribal Health Consortium [ANTHC] and Yukon-Kuskokwim Health Corporation [YKHC]), and a private, statewide, non-profit organization (Rural Alaska Community Action Program, Inc. [RurAL CAP]). The research team met with the Tribal Councils in each of the 12 study communities to share the goals of the study, answer any questions, and obtain approval to conduct the study. Local staff from Head Start helped explain the study to caregivers and facilitated data collection. Caregivers and Head Start staff in the communities also participated in naming the project “Got Neqpiaq?” which, in the Yup’ik language, loosely translates to “Got Real Food?” in reference to traditional Alaska Native foods. All study procedures were approved by the Alaska Area Institutional Review Board (AAIRB), the ANTHC Board of Directors, and the YKHC Human Studies Committee. UAF ceded IRB approval to the AAIRB.

Research Team

The research team consisted of trained researchers from ANTHC, YKHC, and UAF. Additional support was provided by the RurAL CAP dietitian and UAF students completing their final year of the UAF dietetics program. Head Start teachers and additional staff in each community program provided on-site support.

Study Measures

Height, Weight, and Body Mass Index

Heights and weights were collected for all children and caregivers by study staff trained to follow a protocol based on the National Health and Nutrition Examination Survey III (NHANES III) Anthropometric Procedures Manual [15]. Height was measured twice using a standing stadiometer and recorded to the nearest 1/8th inch. If measurements did not agree within 1/8th inch, height was measured a third time. Weight was measured on a Hopkins^® EZ Carry 440LB Digital Scale (Grand Rapids, MI) set to measure in pounds and tenths of pounds. The closest two of three measurements were averaged and recorded. Children were measured and weighed without shoes/boots or outerwear. The study biostatistician calculated BMI as weight (kg)/height (m)². Caregiver BMI was also categorized into healthy weight, overweight, and obese for adults for descriptive analyses according to the standard United States Centers for Disease Control and Prevention classifications [16].

Dietary Factors

The research team collected biomarkers of dietary factors at the time of weight measurement to ensure that weight data were compared to recent diet (e.g., dietary biomarkers reflect diet in the past 4 to 8 weeks). Hair stable isotope ratios were used as objective biomarkers for dietary intakes of traditional foods (the nitrogen isotope ratio, NIR) and processed foods (the carbon isotope ratio, CIR). These biomarkers have been previously validated against 24-h food recalls in Yup’ik populations [17–19] and successfully applied as dietary biomarkers in other studies with Yup’ik children and adults [20–22]. Validation studies have used isotopic measurements in both hair and blood, which are highly correlated [23, 24]. NIR and CIR values are presented in standard delta values as δX = (R_sample − R_standard)/(R_standard) · 1000‰, where X is the heavy isotope (¹⁵N or ¹³C), R is the ratio of heavy to light isotope (¹⁵N/¹⁴N or ¹³C/¹²C) and the standards are atmospheric N for nitrogen and Vienna PeeDee Belemnite for carbon. The traditional food intake biomarker NIR is elevated in marine mammals and fish, which comprise most of the energy from traditional foods in the Yup’ik diet [25]. In a validation study, hair NIR values ranged from 6.9 to 15.2‰ and corresponded to diets where the percent of energy from marine mammals and fish ranged from 0 to 57% [19].

The processed food biomarker CIR is elevated in sweets, sugar-sweetened beverages, and store-purchased meats [17, 18] which contribute to a previously described “processed food” dietary pattern [26]. The CIR has been shown to be more sensitive to these foods when adjusted for the NIR [18]. Thus, we adjust for NIR when examining bivariate associations with CIR.

For each isotopic biomarker, the hair sample was collected by cutting a small pinch of hair (~ 50 hairs) from the back of the head as close to the scalp as possible. The 1-cm section of hair proximal to the head was analyzed, which represents the last 4 to 8 weeks of intake. Samples were cleaned and prepared for stable isotope analysis as described elsewhere [23]. Stable isotope data were obtained using continuous-flow isotope ratio mass spectrometry. This method utilizes a Thermo Scientific Flash 2000 elemental analyzer and Thermo Scientific Conflo IV interfaced with a Thermo Scientific DeltaV Plus Mass Spectrometer. Typically, instrument precision is within 0.2‰. Analyses were performed by the Alaska Stable Isotope Facility at the UAF Water and Environmental Research Center.

The research team assessed relative levels of vegetable and fruit intake using the Veggie Meter (Longevity Link Corp.; Salt Lake City, UT), which measures skin carotenoid status using reflection spectroscopy and has been previously used in Yup’ik communities [22]. Following the protocol described by Ermakov et al. [27], study team members instructed participants to place their forefinger against the lens surface of the Veggie Meter with the help of a spring-loaded cover. Participants used hand sanitizer and wiped their hands prior to the reading. Three readings were collected for each participant and the glass lens of the Veggie Meter was cleaned with an optical cloth between each reading. Mean skin carotenoid status was calculated for each participant by averaging the three measurements. Skin carotenoid status is a unit-less value that ranges from 0 to 800 and represents intake from approximately the past 8 weeks [28]. A validation study found that across a year of measurements, self-reported vegetable and fruit intake was correlated with skin carotenoid status measured by reflection spectroscopy (r = 0.37, p < 0.001) [29, 30]. The same study found a high correlation between reflection spectroscopy and resonance Raman spectroscopy (r = 0.86, p < 0.001) and between reflection spectroscopy and plasma carotenoids (r = 0.70, p < 0.001).

Confounders

Potential confounders of the association of caregiver dietary factors with caregiver BMI were identified as caregiver age, caregiver sex, caregiver education, annual household income, Yup’ik cultural identity, Kass’aq (white) cultural identity, and food assistance participation (Supplemental Nutrition Assistance Program [SNAP] and Special Supplemental Nutrition Program for Women, Infants, and Children [WIC]). Confounders of the associations of caregiver dietary factors with child dietary factors and caregiver BMI with child BMI were identified as caregiver and child age, caregiver and child sex, caregiver education, annual household income, caregiver Yup’ik cultural identity, caregiver Kass’aq (white) cultural identity, food assistance participation, and caregiver dietary factors (the latter only for concordance of caregiver and child BMI). Potential confounder variables were collected via a survey administered to caregivers. Caregiver age and caregiver sex were not measured and were not adjusted for in models. In place of caregiver age and sex, an indicator for the caregiver being the child’s birth mother was used (yes/no). Caregivers that were not birth mothers were legal guardians of the child and included birth fathers, adopted parents, or grandmothers.

Statistical Analysis

Descriptive statistics for study variables were calculated using percentages, means, and standard deviations. For the first study goal, we used confounder-adjusted linear regression to estimate the concordance of caregiver dietary factors with child dietary factors and caregiver BMI with child BMI. BMI was treated as a continuous variable in all statistical analyses. For the second study goal, we used confounder-adjusted linear regression to examine the associations between caregiver dietary factors with continuous caregiver BMI. Unstandardized regression coefficients and 95% confidence intervals are reported.

We included only observations with complete data for all analytical variables in the multivariable model. All statistical analysis was completed using SPSS version 27 software (IBM SPSS Software, version 27. New York: IBM. 2022.). To account for potential clustering at the household level, an indicator for whether each caregiver had one vs. two or more children enrolled in the study was included for models that examined the concordance of child and caregiver BMI and dietary factors.

Results

Descriptive Statistics

A total of 155 caregiver-child dyads (144 unique caregivers) participated in the study, representing 61% of eligible caregiver-child dyads in the 12 partner communities. Almost all parents reported Yup’ik ethnicity (86%) (Table 1). About 44% reported following a Yup’ik way of life a lot and 54% reported following it some, while 33% reported following a Kass’aq way of life a lot and 60% reported following a Kass’aq way of life some of the time. About half (56%) of parents completed a high school education and 22% reported completing some college. Most households in the study reported an annual income < $10,000; 84% and 80% of participants participated in SNAP and WIC, respectively.

Table 1.

Child, home caregiver, and household characteristics for 144 Alaska Native (Yup’ik) preschoolers aged 3–5 years enrolled in Head Start who participated in the “Got Neqpiaq?” study in Southwest Alaska

Child, parent, and household characteristics	Caregivers of Yup’ik children aged 3–5 years enrolled in Head Start, (N = 144), mean (SD), range or n (%)a
Child age (years)	4.02 (0.61), 2.58–5.06
Child sex
Male	75 (52%)
Female	69 (48%)
Parent ethnicityb
Yup’ik	122 (86%)
Cup’ik	19 (13%)
Athabascan	4 (3%)
Inupiaq	2 (1%)
Other	3 (2%)
Missing	2
Caregiver cultural identity
Follows Yup’ik way of life
A lot	62 (45%)
Some	75 (54%)
Not at all	2 (1%)
Missing	5
Follows Kass’aq (white) way of life
A lot	45 (33%)
Some	83 (61%)
Not at all	9 (7%)
Missing	7
Caregiver education
Did not complete high school	28 (21%)
Completed high school	76 (56%)
Some college	30 (22%)
Graduated from college	2 (1%)
Missing	8
Annual household income
< $10,000	69 (50%)
$10,000-$14,999	12 (9%)
$15,000-$24,999	8 (6%)
$25,000-$34,999	7 (5%)
$35,000-$49,999	8 (6%)
$50,000-$74,999	8 (6%)
$75,000-$99,999	2 (1%)
≥ $100,000	1 (< 1.0%)
Not sure	24 (17%)
Missing	5
Food assistance participation
SNAP benefits or Quest card	118 (84%)
WIC	112 (80%)
Missing
Caregiver identity
Birth mother	106 (74%)
Other	38 (26%)
Multiple children per household
Yes	11 (8%)
No	133 (92%)

SNAP Supplemental Nutrition Assistance Program, WIC Special Supplemental Nutrition Program for Women, Infants, and Children

^aMissing data were excluded from the % calculation

^bCaregivers could select more than one ethnic category, so percentages add up to more than 100%

Dietary Factors and BMI

The mean NIR (δ¹⁵N value) was 8.8‰ for children and 9.3‰ for adults (Table 2).

Table 2.

Concordance of dietary factors and BMI^a for home caregivers of Alaska Native (Yup’ik) preschoolers aged 3–5 years enrolled in Head Start who participated in the “Got Neqpiaq?” study in Southwest Alaska

	Children	Caregivers	Concordance, estimatee (95% CI)	p
Traditional food intake biomarkera (‰)	8.78 (0.82)	9.29 (1.15)	0.39 (0.28, 0.50)	< .001
Processed food intake biomarkerb (‰)	− 17.02 (0.73)	− 17.28 (0.72)	0.43 (0.25, 0.62)	< .001
Vegetable and fruit intake biomarkerc	195.5 (93.1)	178.6 (69.7)	0.39 (0.13, 0.64)	.003
BMI (percentile or kg/m2)d	91.5 (15.7)	28.8 (6.6)	0.43 (− 0.04, 0.90)	.07

BMI body mass index, CI confidence interval, kg kilogram, m meter

^aTraditional food intake was measured using the NIR, expressed as permil (‰) relative abundance of heavy isotope (δ¹⁵N = (¹⁵N/¹⁴N sample − ¹⁵N/¹⁴Nstd)/(¹⁵N/¹⁴Nstd) • 1000‰), where the standard is atmospheric nitrogen. In a validation study, hair NIR values ranged from 6.9 to 15.2‰ and corresponded to diets where the percent of energy from marine mammals and fish ranged from 0 to 57%

^bProcessed food intake was measured using the CIR, expressed as permil (‰) relative abundance of heavy isotope (δ¹³C = (¹³C/¹²Csample − ¹³C/¹²Cstd)/(¹³C/¹²Cstd) • 1000‰) where the standard is Vienna Pee Dee Belemnite

^cVegetable and fruit intake was evaluated as skin carotenoid status which was measured using the Veggie Meter. Skin carotenoid status is a unit-less measure that ranges from 0 to 800 and represents intake from approximately the past 8 weeks

^dFor caregivers, BMI was calculated as weight (kg)/height (m)². For children, it was adjusted for the child’s sex and age using the United States Centers for Disease Control and Prevention (CDC) 2000 reference data for ages 2 to < 20 years to generate BMI-for-age percentiles

^eEstimates were generated using linear regression adjusted for all dietary factors, birth mother indicator, child age, caregiver education, annual household income, Yup’ik cultural identity, Kass’aq (white) cultural identity, food assistance participation, an indicator for the number of children for each caregiver, and caregiver dietary factors (only for concordance of caregiver and child BMI)

The mean CIR (δ¹³C value) was − 17.0‰ for children and − 17.3‰ for caregivers. Mean skin carotenoid status was 195.5 for children and 179 for caregivers.

Overall, the median BMI percentile was 91.5 (91.3 for male children, 92.4 for female children). According to United States Centers for Disease Control and Prevention (CDC) criteria, the prevalence of overweight was 34.9%, and the prevalence of obesity was 35.5%. The prevalence of a healthy weight among caregivers in the study was 30%, the prevalence of overweight was 35%, and the prevalence of obesity was 35%.

Concordance of Caregiver Dietary Factors with Child Dietary Factors

Dietary factors were highly concordant between caregivers and children. Specifically, higher levels of the traditional food biomarker in caregivers were positively linked to the traditional food biomarker (NIR) in children (beta = 0.39; 95% CI = 0.28, 0.50; p < 0.001), as were results for the processed food biomarker (CIR, beta = 0.43; 95% CI = 0.25, 0.62; p < 0.001) and the vegetable and fruit biomarker (skin carotenoid concentration, beta = 0.39; 95% CI = 0.13, 0.64; p = 0.003).

Concordance of Caregiver BMI with Child BMI

Caregiver BMI was not associated with child BMI (beta = 0.43; 95% CI = (− 0.04, 0.90; p = 0.07).

Association of Caregiver Dietary Factors with BMI

The mean level of dietary factor biomarkers according to caregiver weight status is shown in Table 3. Based on confounder-adjusted linear regression models, higher levels of the traditional food intake biomarker (NIR) were associated with a higher caregiver BMI (beta = 1.57; 95% CI = 0.45, 2.68; p = 0.006) and higher levels of the vegetable and fruit intake biomarker (skin carotenoid concentration) were associated with lower levels of caregiver BMI (beta = − 0.02; 95% CI = − 0.04, − 0.004; p = 0.02). The levels of the processed food intake biomarker (CIR) were not statistically significantly associated with caregiver BMI (beta = 0.74; 95% CI = − 1.07, 2.55; p = 0.42).

Table 3.

Association of dietary factors with BMI^a for home caregivers of Alaska Native (Yup’ik) preschoolers aged 3–5 years enrolled in Head Start who participated in the “Got Neqpiaq?” study in Southwest Alaska (N = 144)

Dietary factor	Healthy weight, n = 41 (30%)	Overweight, n = 48 (35%)	Obese, n = 49 (35%)	Continuous BMI, estimatee (95% CI)	p
Traditional food intake biomarkerb (‰), mean (SD)	8.8 (1.0)	9.4 (1.1)	9.6 (1.2)	1.57 (0.45, 2.68)	.006
Processed food intake biomarkerc (‰), mean (SD)	− 17.3 (0.8)	− 17.3 (0.7)	− 17.2 (0.7)	0.74 (− 1.07, 2.55)	.42
Vegetable and fruit intake biomarkerd, mean (SD)	189 (77)	186 (71)	161 (58)	− 0.02 (− 0.04, − 0.004)	.02

Three adults were dropped for having an underweight BMI, and three were dropped for missing data on BMI

BMI body mass index, SD standard deviation, CI confidence interval

^aTraditional food intake was measured using the NIR, expressed as permil (‰) relative abundance of heavy isotope (δ¹⁵N = (¹⁵N/¹⁴N sample − ¹⁵N/¹⁴Nstd)/(¹⁵N/¹⁴Nstd) • 1000‰), where the standard is atmospheric nitrogen. In a validation study, hair NIR values ranged from 6.9‰ to 15.2‰ and corresponded to diets where the percent of energy from marine mammals and fish ranged from 0 to 57%

^bProcessed food intake was measured using the CIR, expressed as permil (‰) relative abundance of heavy isotope (δ¹³C = (¹³C/¹²Csample − ¹³C/¹²Cstd)/(¹³C/¹²Cstd) • 1000‰) where the standard is Vienna Pee Dee Belemnite

^cVegetable and fruit intake was evaluated as skin carotenoid status which was measured using the Veggie Meter. Skin carotenoid status is a unit-less measure that ranges from 0 to 800 and represents intake from approximately the past 8 weeks

^dBMI was calculated as weight (kg)/height (m)² and healthy weight, overweight, and obese classifications were grouped according to the standard definitions according to the United States Centers for Disease Control and Prevention (CDC)

^eEstimates were generated using linear regression adjusted for all dietary factors, birth mother indicator, caregiver education, annual household income, Yup’ik cultural identity, Kass’aq (white) cultural identity, food assistance participation, and an indicator for the number of children for each caregiver

Discussion

This study evaluated the concordance of dietary factors and BMI between Yup’ik preschoolers in Head Start and their caregivers. It also examined the association of dietary factors and BMI among Yup’ik caregivers. All dietary factors were highly concordant between caregivers and children, but BMI was not. Traditional food intake was positively associated with higher BMI, vegetable and fruit intake was negatively associated with BMI, and processed food intake was not associated with BMI.

Concordant traditional food intake between children and caregivers in the study is consistent with at least one previous study among Yup’ik communities [13] and with intergenerational transmission of values and traditional ways in Yup’ik communities. To our knowledge, this is the first study to report concordance of vegetable and fruit intake and processed food intake between children and caregivers in a Yup’ik population, though similar findings have been reported from studies with other U.S. populations [31]. The high concordance of both nutrient-dense (traditional foods and vegetables and fruit) and energy-dense (processed foods) food groups highlights that obesity prevention efforts should focus on family-level interventions. Further, concordance of child and caregiver diet demonstrates that in Alaska Native communities, families play a critical role in the intergenerational transmission of traditional knowledge that includes the values, beliefs, and practices related to traditional foodways and health. Family-level interventions can foster this intergenerational exchange, thereby strengthening cultural resilience and social cohesion and supporting food sovereignty [32, 33]. Centering interventions on traditional knowledge, values, and worldviews will ensure better relevance and effectiveness.

We did not find evidence that child and caregiver BMI were associated. This conflicts with one study in a different U.S. Head Start population which suggested that caregiver BMI was positively associated with child BMI [34]. Previous work in Yup’ik communities found that caregiver intake of traditional foods was positively associated with child intake of traditional foods and negatively associated with child BMI [13]. A potential explanation for not seeing an association in this study could include that the children in the study were aged 3 to 5 years, which may be too young to detect an association between caregiver and child BMI. For example, a study among 300 2-year-olds and their caregivers, conducted in a U.S. hospital, also reported a null association between caregiver eating competence and child BMI [35]. The researchers suggested that the lack of an association could have resulted from the young age of the children in the study or from a potential mediating effect of caregiver feeding practices.

Among the Yup’ik caregivers in this study, higher levels of vegetable and fruit intake were associated with lower BMI. Vegetable and fruit intake is an important part of maintaining a healthy weight status [36]. Investigations using food frequency questionnaires, skin carotenoid concentrations, and 24-h dietary recalls have shown that the average vegetable and fruit intake in remote Yup’ik communities does not meet recommendations set by U.S. dietary guidelines and is lower than the average intake in the general U.S. population [22, 37]. Our findings suggest that because vegetable and fruit intake was an important factor for caregiver BMI, interventions focused on promoting the intake of these food groups may have relevance for obesity prevention efforts. While our earlier work did not demonstrate that vegetable and fruit intake was an important factor for child BMI in Yup’ik communities [2], recognizing the strong correlation between caregiver and child diets implies that interventions aimed at promoting vegetable and fruit consumption among caregivers could improve intake among their children, ultimately improving weight status as these children transition into adolescence and adulthood [38].

Higher levels of the traditional foods biomarker were positively associated with BMI among Yup’ik caregivers in the study. While frequently consumed traditional foods (e.g., seal oil and fatty fish) tend to be high in fat and therefore calories, the association we observed could have resulted from confounding due to age; both BMI and traditional food intake tend to increase with age [25, 39]. Previous studies have reported that traditional food intake is associated with higher diet quality [39], lower burden of cardiometabolic risk factors and disease [40, 41], and is critically important for cultural practices [42]. Previous studies have also found that traditional food intake attenuates the relationship between BMI and risk factors for chronic disease (i.e., triglycerides and c-reactive protein) [43] suggesting that BMI as a measure of health may not have the same relevance for Yup’ik populations as it does for non-Indigenous populations. In addition, vegetables and fruits are expensive and often unavailable in Yup’ik communities due to the remoteness of the region and high cost of food transportation. Promoting and supporting the intake of traditional foods should remain a priority. Not only are they a locally available choice, but their promotion supports sovereignty, which has holistic benefits [44]. Future interventions should move beyond a singular focus on obesity and consider holistic health, which aligns with a Yup’ik worldview. Given these considerations, there is strong evidence to support that interventions promoting the intake of traditional foods should continue to be prioritized.

Strengths and Limitations

The following limitations should be considered in the interpretation of the study findings. We were unable to account for some key confounders in our analysis including caregiver age and sex which may have contributed to some bias in our estimates. However, we adjusted for type of caregiver (birth mother vs. other), which attempted to minimize some of the bias due to unmeasured confounding. Another limitation is that BMI was our primary health outcome and was measured in isolation from other health determinants or risk factors. There is a documented concern that BMI may not be an appropriate marker of health, especially in Indigenous communities [45]. Nevertheless, BMI continues to be a practical measure especially in community-based studies in rural, remote regions where more invasive measures are impractical. We do urge careful consideration of the implications of using BMI as an isolated measure. Dietary biomarker data are advantageous as they are objective and simple to collect; however, they only reflect the relative intake of foods belonging to traditional dietary patterns and processed food dietary patterns and do not provide information about the intake of specific foods and calories consumed. Lastly, we used cross-sectional data, but additional research can also benefit from follow-up studies that employ longitudinal designs to evaluate whether high BMI results in morbidity.

Study strengths include the use of population-validated, objective tools to measure traditional food, processed food, and fruit and vegetable intake and associations between child and caregiver BMI. The study adds support for using these tools in future studies and for research into additional objective measures of food intake less subject to recall or measurement bias as 24-h recalls or food frequency questionnaires. Importantly, the study fills a critical gap in ascertaining elements to improve childhood and adult obesity and chronic disease prevention efforts for remote rural Yup’ik Alaska Native people.

Public Health Significance and Recommendations for Future Research

Based on the findings of this work, future studies of family-based obesity prevention interventions targeted to Yup’ik caregivers and children are needed. In addition, consideration of health metrics beyond caregiver BMI should be investigated to fully understand the determinants of child diet. Caregiver feeding practices result from the context in which the family resides, including education level, cultural background, social and economic support, and other organizational, community, policy, and media-related factors [12]. In Yup’ik communities, caregiver decision-making around feeding practices includes adult behavior modeling, convenience, seasonality of foods, and the importance of introducing traditional foods and subsistence activities to children at an early age [46], all of which should be taken into account for the development of contextually relevant obesity prevention efforts.