Menu
Menu
Menu
Menu
Previous research suggests that acculturation (i.e., exposure and assimilation to local culture) is associated with changes in dietary patterns among immigrants. This study investigates this association in a refugee population using time in refugee settlement as a proxy for acculturation. A cross-sectional survey was conducted among a systematic sample to (a) identify dietary patterns in Liberian refugees and Ghanaians living in or near a refugee settlement, (b) compare adherence to these dietary patterns between groups, and (c) investigate the association between acculturation and dietary patterns in Liberian refugees. Participants were Liberian and Ghanaian women with young children living in the Buduburam refugee settlement or Awutu in Ghana (n = 480; 50% Liberian; mean age 28, SD 6.3, range 16–48 years). Time in settlement was assessed by self-report; food consumption was assessed by food frequency questionnaire. Principal component analysis was used to identify dietary patterns; a generalized linear model was used to test the association of interest. Three distinct dietary patterns emerged: Healthy, Sweets, and Fats. Ghanaians were more adherent to the Healthy pattern than Liberians (p < 0.05). Liberians were more adherent to the Sweets and Fats patterns than Ghanaians (p < 0.05). There were no significant differences in dietary pattern adherence among the Liberians based on time in settlement. Ghanaians living in Awutu were more adherent to the Healthy pattern than Ghanaians who lived in settlement (p < 0.05). Differences in dietary patterns were observed between Liberian refugees and Ghanaians. These differences were not associated with acculturation and may be related to the food environment in the settlement.
A cross‐sectional survey was administered between July and August 2008 among 480 female Liberian refugees and Ghanaians living in the Buduburam refugee settlement and the nearby urban village of Awutu (5 km from Buduburam). Women were included in the study if they were Liberian or Ghanaian, were 16 years of age or older, had a biological child between the ages of six and 59 months, were not currently pregnant, had no health problem or condition that caused a modification to their diet (e.g., diabetes and heart disease), and lived in either the Buduburam refugee settlement or Awutu. A systematic sampling approach was used to identify and recruit participants. Within Buduburam, a central location was chosen in each of 12 zones. Four teams of interviewers (one Ghanaian and one Liberian per team) were employed to collect the data. Each team began at a central location within their assigned zone. Standing in that central location, the team chose a random direction and visited the first household they encountered. The team then went to every fifth house in the same block and then moved to the next block and continued the sampling procedure until reaching the desired sample size within that zone (120 Liberians in Zones 1–10; 119 Liberians and 121 Ghanaians in Zones 11 and 12; and 120 Ghanaians in Awutu). If the mother was not present in the household at the time of the visit, the household was revisited at a more convenient time. If more than one mother living in the household met the inclusion criteria, one mother was randomly chosen to participate. This same sampling method was also employed in Awutu. Trained Liberian and Ghanaian interviewers from the target communities administered the survey. The interviewers underwent 3 days of training on conducting the interviews, interview techniques, and taking anthropometric measurements. The Liberian interviews were conducted in English (Liberian pigeon English). The Ghanaian interviews were conducted in English or the local Ghanaian dialect based on the interviewee's preference. Each interview lasted 1.5 to 2 hr. Interviews were reviewed daily for quality and standardization. Participants were revisited if data were missing or if responses were inconsistent to resolve any issues. Verbal informed consent was obtained and formally recorded for all participants prior to survey administration. Participants were assured that all information would remain confidential, would not affect access to programs within the settlement, and would not be used for determining repatriation. Institutional Review Board (IRB) approval to conduct this study was obtained from the University of Connecticut and the University of Ghana IRBs. Yale University IRB granted approval to conduct data analysis. Representatives of the Buduburam refugee settlement gave permission to defer ethical approval for this study to the collaborating universities. The administered survey was pretested among five Liberian and four Ghanaian women meeting the inclusion criteria, and the survey was modified as a result. The final survey administered at each interview assessed the following: demographic/household characteristics, degree of acculturation, household food security, infant feeding practices, infant and maternal health status, and maternal dietary intake. The following anthropometrics were also assessed in the respondent and the index child: weight, height, mid‐upper arm circumference, and head circumference (children only). Dietary intake was assessed using a culturally appropriate detailed food frequency questionnaire (FFQ). The FFQ was adapted from the Block FFQ (Block et al., 1986) to include traditional Liberian and Ghanaian foods. Traditional Liberian and Ghanaian foods were included in the FFQ after conducting key informant interviews with Liberian refugees and Ghanaians living or working within the Buduburam refugee settlement or Awutu, visiting local markets, and consulting with Liberian refugees employed by the Buduburam nutrition program. Participating women were asked about their consumption (yes/no) of 132 food/beverage items over the past 6 months within the following 12 food/beverage categories: fruits; vegetables; beans and nuts; meats; fish and seafood; cereal and grains; milk and dairy products; snacks, sweets, and desserts; drinks; tubers; other foods; and traditional mixed dishes. Participants were shown pictures of various food/beverage items to ensure they clearly understood the food/beverage items in the list. They were then asked to report how many times they had consumed each food/beverage item, reported as either daily, weekly, monthly, or only occasionally. Participants were also able to provide the name and consumption frequency of any other foods/beverages they had consumed within the 12 food/beverage categories. Time in the Buduburam refugee settlement served as a proxy for acculturation. Time in the settlement was assessed in years and/or months during the interview through self‐report. Of the 132 specific food/beverage FFQ items collected, 34 were excluded in the final FFQ analyses, specifically 26 Liberian and Ghanaian traditional mixed dishes and eight food/beverage items that were added to the survey after survey administration began. The traditional mixed dishes were excluded to ensure food/beverage items were not double counted and consumption levels were not inflated because these mixed dishes consisted of mixed food groups (i.e., complex composition) and it was uncertain whether the constituents of these mixed dishes were included by participants in the reporting of individual food/beverage items. Those food/beverage items that were added to the survey after survey administration began were excluded because not all participants were asked about these foods/beverage items. Finally, any foods/beverages that participants specified in response to the prompt “other” (e.g., “other fruits, specify”) were not included in the analysis because these items were not systematically collected for every participant. This resulted in a final total of 98 food/beverage items included in the analysis. Research suggests that it is preferable to study dietary patterns and quality (i.e., whole diet) rather than individual food components (Newby & Tucker, 2004). Therefore, the FFQ data were recoded into 32 food groups using the groupings classified in the Nutrient Data System for Research software from the University of Minnesota (University of Minnesota, Nutrition Coordinating Center, Minneapolis, MN; Table 1). Average weekly frequency of consumption over a 6‐month period was calculated for each individual for each food group. For example, a response of one time per day was represented by 1 × 7 = 7 times per week. Food/beverage groups used in final analysis and descriptions Principal component analysis was used to identify distinct dietary patterns among all respondents. Principal component analysis assigned coefficients to each food group, which were used to generate dietary pattern scores (Newby & Tucker, 2004; Sofianou et al., 2011). Dietary patterns with eigenvalues greater than 1.5 were identified and were orthogonally rotated (varimax rotation; Newby & Tucker, 2004; Sofianou et al., 2011). Four dietary patterns were identified, explaining 33.3% of the variance. The food items were retained in the interpretation of their principal component if their loadings were ≥0.4, a cut‐off chosen to aid interpretation of results. Only three of the four dietary patterns were retained because it was not possible to interpret the meaning of the fourth dietary pattern (Table 2). Using all coefficients, an individual dietary pattern score was calculated for each individual for each dietary pattern. Individual dietary pattern scores were generated by multiplying the factor loading for each food group by the weekly frequency of consumption of each food group and then summing all products for each dietary pattern. The scores for each pattern were approximately normally distributed. A high positive score indicated high adherence to a dietary pattern, and a low score indicated little or no adherence to a dietary pattern (Newby & Tucker, 2004; Sofianou et al., 2011). Highest factor loadings of food/beverage groups by dietary patterna The generalized linear model was used to assess the association between time living in the Buduburam refugee settlement and dietary pattern scores. Three models were run with each dietary pattern score as an outcome. The dietary pattern score for each dietary pattern was included in the model as a continuous variable. Time living in the Buduburam refugee settlement was included in the model as the exposure proxy variable for acculturation. Five population subgroups were created based on this proxy variable: Liberian refugees who had lived in the settlement less than 8 years (the median for this group), Liberians who had lived in the settlement for 8 or more years, Ghanaians who had lived in the settlement less than 5 years (the median for this group), Ghanaians who had lived in the settlement 5 or more years, and Ghanaians who lived in Awutu (i.e., lived 0 years in the settlement). Each model was adjusted for the following covariates: age, marital status, level of education, employment status, income, household size, presence of electricity in the home, and whether or not money had been borrowed from or loaned to others in the past year. Marital status, level of education, employment status, income, presence of electricity in the home, and whether or not money had been borrowed from or loaned to others were included in the model as categorical variables. Age and household size were included in the model as continuous variables. All analyses were performed using SPSS (version 22.0).
N/A
