Body size preferences and food choice among mothers and children in Malawi

Overweight in mothers and children in sub-Saharan Africa is rapidly increasing and may be related to body size perceptions and preferences. We enrolled 268 mother–child (6–59 months) pairs in central Malawi; 71% of mothers and 56% of children were overweight/obese, and the remainder were normal weight. Interviewers used seven body silhouette drawings and a questionnaire with open- and closed-ended questions to measure mothers’ perceptions of current, preferred and healthy maternal and child body sizes and their relation to food choices. Overweight/obese and normal weight mothers’ correct identification of their current weight status (72% vs. 64%), preference for overweight/obese body size (68% both) and selection of an overweight/obese silhouette as healthy (94% vs. 96%) did not differ by weight status. Fewer overweight/obese than normal weight mothers’ preferred body silhouette was larger than their current silhouette (74% vs. 29%, p <.001). More mothers of overweight than normal weight children correctly identified the child's current weight status (55% vs. 42%, p <.05) and preferred an overweight/obese body size for the child (70% vs. 58%, p <.01), and both groups selected overweight/obese silhouettes as healthy for children. More than half of mothers in both groups wanted their child to be larger than the current size. Mothers said that increasing consumption of fruits, vegetables, meat, milk, grains, fizzy drinks and fatty foods could facilitate weight gain, but many cannot afford to purchase some of these foods. Their desired strategies for increasing weight indicate that body size preferences may drive food choice but could be limited by affordability. This analysis uses data from a study of drivers of food choice in households where the mother, child or both were overweight. The study was conducted in Lilongwe and Kasungu Districts in central Malawi. We selected these areas because they have a higher prevalence of overweight/obesity among mothers and children younger than 5 years of age than other parts of Malawi (National Statistical Office [NSO] Malawi & ICF, 2017). In each district, we chose two urban neighbourhoods and two rural communities as data collection sites. At each site, mothers with children 6 months to 5 years of age were invited for screening at a central location. Five research assistants were trained to collect anthropometric measurements (Cogill, 2003), and their measurements were standardized against those of an experienced researcher prior to starting data collection. Maternal height and standing height of children 2 years or older were measured to the nearest 0.1 cm using a portable stadiometer (Seca 213). Recumbent length of children younger than 2 years was measured to the nearest 0.1 cm using an infant measuring mat (Seca 210). Weight of mothers and children 2 years or older was measured to the nearest 0.1 kg using a digital scale (Seca 803). Weight of children younger than 2 years was measured to the nearest 0.1 kg using a digital baby scale (Seca 354). We used the anthropometric data to calculate body mass index (BMI, kg/m2) of mothers and used the standard cut‐offs for normal weight (18.5 kg/m2 ≤ BMI < 25 kg/m2) and overweight or obesity (BMI ≥ 25 kg/m2). We calculated weight‐for‐heightz‐scores(WHZ) for children using the World Health Organization (WHO) growth standard and used the WHO cut‐offs for normal weight (−2 SD < WHZ ≤ +2 SD) and overweight (WHZ > +2 SD; WHO Multicentre Growth Reference Study Group, 2006). Mother–child dyads were purposefully enrolled in three groups: overweight mother with an overweight child, overweight mother with a normal weight child and normal weight mother with an overweight child. This purposive sampling technique was used to ensure that we had sufficient representation to draw inferences about the relationship between maternal/child weight status, and body size perceptions and preferences, as well as food choice. Research assistants were trained to use a set of seven adult female and seven child body silhouette drawings (Figure 1) and a semistructured questionnaire to measure mothers’ perceptions of their current, preferred and healthy body sizes for themselves and their child. A local artist adapted mothers’ body silhouettes from a version previously used in Malawi (Bentley et al., 2005; Croffut et al., 2018) and validated in a sample of mother–daughter dyads in South Africa (McIza et al., 2005 ). The same artist adapted child body silhouettes from Hager, McGill, and Black (2010). Both the mother and child silhouettes were originally patterned on Stunkard, Sorensen, and Schulsinger (1983). For the mothers’ silhouettes, we followed a similar BMI categorization as a previous study that developed body silhouettes for an African American population (Pulvers et al., 2004), later validated in a sample of women in the Seychelles (Yepes, Viswanathan, Bovet, & Maurer, 2015). In this study, the thinnest silhouette was assigned a BMI of 17 kg/m2 with a 3‐BMI unit increment for each subsequent silhouette, making the heaviest silhouette equal to 35 kg/m2. Consequently, Silhouette 1 was classified as underweight, Silhouettes 2 and 3 as normal weight, Silhouettes 4 and 5 as overweight and Silhouettes 6 and 7 as obese. We applied the same categorization to the children’s silhouettes. Mother and child body silhouettes Each body silhouette drawing was printed separately on cardstock and laminated. The interviewer mixed the body silhouettes and laid them out in a random order before each question (i.e., current, preferred and healthy body size) separately for the mother to make selections for herself and then again for her child. The specific questions related to selection of the silhouettes for women were as follows: Open‐ended questions in the questionnaire were used to understand mothers’ body size selections and how they were related to food choices. Questions about food choice included: How does the difference between your current figure and the figure you would like to have influence what types of food and drinks you buy? How does the difference between your current figure and the figure you believe is healthy influence what types of food and drinks you buy? A similar set of questions were used to obtain information about mothers’ preferences for child body size and food choices. We performed chi‐squared tests comparing body size perceptions to actual body size by normal versus overweight status separately for mothers and children. We calculated the difference between the selected current and preferred body silhouette numbers for mothers and children to quantify how many preferred a smaller, the same or a larger body size and conducted chi‐squared tests comparing the difference by mother and child weight status. We also used chi‐squared tests to examine differences in the mother’s selections of child current, preferred and healthy body silhouettes by the mother’s weight status. We calculated descriptive statistics for participants’ socio‐economic characteristics, including age, maternal education (secondary or higher vs. primary or no education), household assets (sum of 12 household durable goods, range 0–12), household food insecurity access score (HFIAS) as a continuous variable (range 0–27; Coates, Swindale, & Bilinksy, 2007) and rural/urban residence. We sorted responses to open‐ended questions in an Excel data matrix by weight status based on BMI or WHZ. After reading through the responses, we developed codes based on the main topics that emerged from each open‐ended question. Within the sorted segments of the open‐ended data, one researcher applied the codes and summarized the findings using qualitative content analysis methods (Hsieh & Shannon, 2005). For lists of foods provided in open‐ended responses, we divided the data for mothers and children by those whose preferred silhouette was smaller than their current silhouette and those whose preferred silhouette was the same or larger than their current silhouette. We then tabulated the food items and ranked the top 10 foods from most to least frequently mentioned. Participants received an incentive of 4 US dollars. The study was approved by the College of Medicine Research Ethics Committee at the University of Malawi and by the institutional review boards at RTI International and the Harvard T.H. Chan School of Public Health.